Chemico-Biological Interactions 205 (2013) 157–164

Contents lists available at ScienceDirect

Chemico-Biological Interactions

journal homepage: www.elsevier .com/locate /chembioint

Fenofibrate suppresses melanogenesis in B16-F10 melanoma cells viaactivation of the p38 mitogen-activated protein kinase pathway

0009-2797/$ - see front matter � 2013 Elsevier Ireland Ltd. All rights reserved.http://dx.doi.org/10.1016/j.cbi.2013.07.008

⇑ Corresponding author. Tel./fax: + 886 4 2631 1167.E-mail address: ychuang@pu.edu.tw (Y.-C. Huang).

Yu-Chun Huang a,⇑, Kao-Chih Liu a, Yi-Ling Chiou a, Chao-Hsun Yang a, Tien-Hui Chen b,Ting-Ting Li a, Li-Ling Liu a

a Department of Cosmetic Science, Providence University, No. 200, Sec. 7, Taiwan Boulevard, Shalu Dist., Taichung City 43301, Taiwan, ROCb Department of Obstetrics & Gynecology, Chang Bing Show Chwan Memorial Hospital, No. 6 Lugong Rd., Lukang Zhen, Changhua County, Taiwan, ROC

a r t i c l e i n f o

Article history:Received 2 March 2013Received in revised form 17 June 2013Accepted 11 July 2013Available online 18 July 2013

Keywords:FenofibrateTyrosinasep38 Mitogen-activated protein kinaseLiver X receptorPeroxisome proliferator-activated receptorMelanocortin 1 receptor

a b s t r a c t

Fenofibrate and ciglitazone belong to the classes of fibrates and thiazolidinediones, respectively. Theirpharmacological actions on peroxisome proliferator-activated receptors (PPARs) present a potentialtherapy for hyperlipidemia and hyperglycemia. However, the melanogenesis affected by PPAR ligandsin melanocytes has not been well investigated. By determining the melanin content of cells treated withPPAR agonists, we showed that fenofibrate significantly reduced melanin synthesis, but its major activemetabolite, fenofibric acid, did not. Notably, the suppression of melanogenesis by fenofibrate could not beprevented by the PPARa specific antagonist GW6471. In addition, T0901317, a liver X receptor (LXR)agonist, restored the antimelanogenic activity of fenofibrate. Accordingly, fenofibrate may suppressmelanogenesis through a PPARa-independent pathway. Treatment of cells with fenofibrate led to thedown-regulated gene expression of melanocortin 1 receptor (MC1R). Fenofibrate also attenuated thedihydroxyphenylalanine (DOPA)-staining activity and expression of tyrosinase as well as the expressionof microphthalmia-associated transcription factor (MITF). The phosphorylation of p38 mitogen-activatedprotein kinase (MAPK) was stimulated by fenofibrate. Furthermore, the p38 MAPK inhibitor SB203580prevented the repressive effects of fenofibrate on the melanin production. Taken together, the resultsof the present study suggest that fenofibrate inhibits melanin synthesis via the down-regulation ofMC1R, the up-regulation of p38 MAPK, and interference with LXR signaling pathways to decrease theexpression of tyrosinase in B16-F10 melanoma cells.

� 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

The color of mammalian skin and hair is determined by a num-ber of factors, including the distribution of the pigment melanin.Melanocytes, which synthesize melanin, act as pivotal regulatorsof skin pigmentation, or melanogenesis. Of the five members ofmelanocortin receptors, the melanocortin 1 receptor (MC1R),expressed primarily in melanocytes, plays a critical role in regula-tion of pigmentation [1]. Microphthalmia-associated transcriptionfactor (MITF) is the most critical transcription factor in theregulation of the development and differentiation of melanocytes.Transcription dependent on MITF regulates the expression oftyrosinase and tyrosinase-related proteins that control the conver-sion of tyrosine to melanin [2]. The MITF promoter is regulated inpart by various other transcription factors, such as paired box pro-tein PAX3, SRY-related HMG-box 10, Lymphoid enhancer-bindingfactor 1/T cell-specific factor, and cAMP response element-binding

protein [1]. Accordingly, multiple signaling pathways are inte-grated into the modulation of melanocyte functions. Endogenousand exogenous factors up-regulate or down-regulate melanogenicprocesses, and dysregulation of these factors causes pigmentationdisorders [3].

Biochemical reaction and exposure to environmental stresssuch as ultraviolet radiation and pollutants can generate reactiveoxygen species (ROS), including oxygen radicals, and certainnonradicals that are either oxidizing agents and/or are easily con-verted into radicals [4,5]. Accumulation of ROS has been shown inthe epidermis of vitiligo patients [6]. Gene expression pattern isregulated by ROS via modulation of transcription factor activityparticularly nuclear factor kappa B (NF-jB), activator protein-1(AP-1), and peroxisome proliferator-activated receptor (PPAR)family of transcriptional activators. Oxidative stress is defined asan imbalance between cellular production of oxidant moleculesand the availability of appropriate antioxidants that defend againstthem [7]. The reduction–oxidation (redox) balance plays an impor-tant role in physiological modulation. Continued oxidative stresscan lead to the development of chronic diseases, including cancer,neurodegeneration, cardiovascular and metabolic diseases.

158 Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164

Obesity is associated with insulin resistance and type 2 diabetesmellitus (DM). Obese and DM patients have many cutaneous prob-lems, which may be associated with prolonged wound healing,inflammatory dermatoses, and melanoma. A link between excessadiposity and malignant melanoma has been found by variousepidemiological studies [8,9]. Therapeutic agents, such as fibratesand thiazolidinediones, have been evaluated for the treatment ofpatients with type 2 DM with multiple cardiometabolic risk factors[10]. Moreover, a meta-analysis of trials involving statins, fibrates,and melanoma risk was performed [11]. Fibrates (fenofibrate, gem-fibrozil) activate PPARa. In clinical trials, significantly fewer pa-tients treated with gemfibrozil were diagnosed with melanomacompared with the control group [12]. Thiazolidinediones, alsoknown as glitazones, are potent and highly selective agonists forPPARc.

Three PPAR isotypes have been isolated – PPARa, PPARb/d, andPPARc – with distinct tissue distributions and biological activities.Each PPAR heterodimerizes with the retinoid X receptor to form acomplex that translocates to the nucleus and regulates geneexpression. PPARs are involved in various skin diseases, particu-larly those involving inflammation, epidermal hyperproliferation,and differentiation [13–15]. PPARa and PPARc agonists exhibitthe potential therapeutic application for the treatment of pigmen-tary disorders such as melasma, vitiligo, and postinflammatoryhyperpigmentation [16]. For instance, dioic acid, through bindingof PPARc, exhibited lightening effects in the treatment of melasma[17]. The mRNA of all three PPAR subtypes is expressed in melano-cytes [16]. Cultured mouse and human melanoma cells producemore PPARa and PPARc protein compared to melanocytes [18].PPARa and PPARc are expressed in B16-F10 and B16 cells, respec-tively [18,19]. Nevertheless, the molecular events following PPARligand binding have not been well investigated in melanocytes.Therefore, the melanogenic activity of murine B16-F10 melanomacells treated with various PPAR agonists was assessed in thepresent study.

2. Materials and methods

2.1. Chemicals and reagents

Fenofibrate, GW0742, 3-[4,5-dimethylthiazol-2-yl]-2,5-diphe-nyltetrazolium bromide (MTT), L-3,4-dihydroxyphenyalanine(L-DOPA), N-acetyl-L-cysteine (NAC), and melanin were obtainedfrom Sigma–Aldrich (St. Louis, MO, USA). Ciglitazone, MG132,SB203580, and T0901317 were purchased from Enzo Life Sciences(Farmingdale, NY, USA). Fenofibric acid was purchased from Toron-to Research Chemicals (Toronto, Ontario, Canada). GW6471 waspurchased from Tocris Bioscience (Bristol, UK). GW7647 was pur-chased from Cayman Chemical (Ann Arbor, MI, USA). An antibodyagainst tyrosinase was obtained from Epitomics (Burlingame, CA,USA). The anti-MITF antibody was obtained from Santa Cruz Bio-technology (Santa Cruz, CA, USA). Antibodies against p38 MAPKand phosphorylated p38 MAPK (Thr 180/Tyr 182) were purchasedfrom Cell Signaling Technology (Danvers, MA, USA). Antibodiesagainst a-tubulin and actin were supplied by Millipore (Temecula,CA, USA).

Fenofibrate, fenofibric acid, GW0742, GW 6471, GW 7647,ciglitazone, MG132, T0901317, and SB203580 were dissolved indimethylsulfoxide (DMSO) and further diluted in culture medium.To characterize the action mechanism of fenofibrate, B16-F10cells were pretreated with MG132, T0901317, SB203580 or NACfor 30 min and then treated with fenofibrate for further times.The final DMSO concentration in the medium was 0.1% and didnot affect cellular function or the assay systems used in thisstudy.

2.2. Cell culture

The murine B16-F10 melanoma cell line was purchased fromthe Bioresource Collection and Research Center (Hsinchu, Taiwan),and cells were maintained in Dulbecco’s modified Eagle’s medium(Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine ser-um (Biological Industries, Kibbutz Beit Haemek, Israel) in a humid-ified incubator at 37 �C with 5% CO2.

2.3. Cell viability

Cell viability was determined using the MTT assay. Briefly, theB16-F10 cells were seeded at a density of 5 � 103 cells/well on96-well plates and cultured overnight as described above. Themedium was replaced with fresh medium containing test com-pounds at various concentrations. After incubation for 48 h at37 �C with 5% CO2, MTT was added to a final concentration of0.5 mg/mL, and the cells were incubated at 37 �C for 2 h. Finally,the cells were lysed, and absorbance was detected at 550 nm.The number of cells was determined by interpolation of thedetected absorbance density values using a standard correlationbetween known cell numbers and their absorbance density values.

2.4. Melanin content determination

Melanin content was measured using a previously describedmethod [20]. Briefly, cells were seeded at a density of 2 � 105 -cells/60-mm culture dishes and cultured as described above. Afterovernight incubation, the cells were cultured with or without testcompounds for 48 h. The medium was removed, and the cells werewashed twice with phosphate-buffered saline (PBS) and harvestedby trypsinization using 0.05% trypsin/0.02% EDTA. The harvestedcells were centrifuged, the pellet was dissolved by adding 1 NNaOH, and the mixture was incubated at 60 �C for 1 h. The amountof melanin in the solution was determined by measuring theabsorbance at 470 nm using the microplate reader (BioTek Instru-ments, Winooski, VT, USA). The total melanin content was esti-mated using a standard curve of synthetic melanin. The melanincontent was calculated by normalizing the melanin contents to to-tal cellular protein (lg of melanin/mg of protein) and reported as apercentage of control. The protein content was determined usingthe Micro BCA protein assay kit (Pierce Biotechnology, Rockford,IL, USA).

2.5. DOPA-staining assay (tyrosinase zymography)

A DOPA-staining assay was performed as described by Sato andco-workers [21] with slight modifications. Briefly, the cells werelysed in 0.1 M sodium phosphate buffer (pH 6.8) containing 1% Tri-ton X-100 and 1 mM phenylmethylsulfonyl fluoride and aliquots of30 lg of normalized proteins were resolved by 8% sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS–PAGE) under non-denaturing conditions. Afterward, the gels were transferred into astaining solution containing 1 mM L-DOPA and incubated in thedark for 1 h at 37 �C. Tyrosinase activity was visualized in the gelsas dark, melanin-containing bands. To directly assess the activity oftyrosinase, fenofibrate was added to the untreated cell lysate andincubated for 5 min at room temperature. The cell lysates weremixed with L-DOPA solution and incubated at 37 �C for 2 h, asdescribed above.

2.6. Western blots

Whole cell lysates were prepared using RIPA buffer. Aliquots ofthe cell lysates (20 lg each lane) were separated by SDS–PAGE,transferred onto polyvinylidene difluoride membranes, and blotted

Fig. 1. Effects of PPAR agonists on cell viability and melanin synthesis in B16-F10melanoma cells. The cells were treated with 0.1–10 lM (A) and 10 lM (B) of thePPAR agonists, fenofibrate, GW0742, or ciglitazone, for 24–48 h. Cell viability (A)and melanin content (B) were determined. Each value is expressed as a percentagerelative to the DMSO-treated control. The data are reported as the means ± S.E.M. ofthree independent experiments. **P < 0.01 compared with the control.

Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164 159

with the indicated antibodies: anti-tyrosinase 1:100, anti-MITF1:500, anti-P-p38 1:5000, anti-p38 1:6000, anti-Bcl-2 1:100, anti-Bax 1:100, anti-actin 1:20,000, anti-a-tubulin 1:20,000, and anti-GAPDH 1:5000. Horseradish peroxide-conjugated goat anti-rabbitand anti-mouse immunoglobulin (Jackson ImmunoResearchLaboratories Inc., West Grove, PA, USA) were used at 1:12,000and 1:5000, respectively. The proteins were detected using anenhanced chemiluminescence kit (Amersham Biosciences, Buck-inghamshire, UK). Western blots were representative of at leastthree independent experiments. Quantitative analysis of specificbands was performed using Image Quant analysis software (GEHealthcare).

2.7. Real-time reverse transcriptase-polymerase chain reaction (RT-PCR)

Total RNA was extracted from cells using Trizol reagent(Invitrogen, Carlsbad, CA, USA), and 1 lg of total RNA was reversetranscribed to cDNA using the iScript™ cDNA Synthesis Kit(Bio-Rad, Hercules, CA, USA). The real-time PCR amplificationswere performed in a MiniOpticon™ real-time PCR system usingiQ™ SYBR

�Green Supermix (Bio-Rad, Hercules, CA, USA) and prede-

signed gene specific primer sets. Primers for mouse genes were asfollows: Tyrosinase, forward 50-AACAGCGATGGGAAACTA-30, re-verse 50-AACAGGTGTGAAGGTCTC-30; MITF, forward 50-TCCGTTTAGAGTGCTCAA-30, reverse 50-AAAGTCAGAAGGTTTGGTATT-30; Mela-nocortin 1 receptor (MC1R), forward 50-TGCCTCTATGGTGACAAGAA-30, reverse 50-ATGCTACGGTCCTGCTAC-30; Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), forward 50-TCCAAGGAGTAAGAAACC-30, reverse 50-GAAATTGTGAGGGAGATG-30. The expressionof GAPDH was used as the internal standard. PCRs were done intriplicates. Genes of interest from three separateexperiments wereanalyzed by CFX manager software (version 2.1), and expressed asfold changes over that in control group.

2.8. Statistics

The data were expressed as the means ± standard error of themean (S.E.M.) of the indicated number of separate experiments.A one-way analysis of variance was performed for multiple com-parisons. In the event of significant variation among treatmentgroups, the mean values were compared to the respective controlusing Dunnett’s test. P-values less than 0.05 were considered sta-tistically significant.

3. Results

3.1. The PPARa agonist fenofibrate decreases melanin synthesis

Briefly, B16-F10 melanoma cells were treated with PPAR ago-nists for 48 h to assess the effect of the agonists on cell viability.Fenofibrate, GW0742, and ciglitazone are agonists of PPARa, b/d,and c, respectively. The MTT assay revealed that the compoundsdid not alter cell survival even at 10 lM (Fig. 1A). The cellular mel-anin contents were evaluated simultaneously. As shown in Fig. 1B,fenofibrate, but not the other PPAR agonists, decreased the melanincontent after 24 or 48 h of incubation. The inhibitory effect offenofibrate on melanin synthesis was time- and concentration-dependent (Fig. 1B and Fig. 2A). Melanin production was inhibitedby approximately 40% in cells treated with 10 lM fenofibrate for48 h. Although the inhibitory effect of 20 lM fenofibrate was morepotent than that of 10 lM (Fig. 2A), an anti-proliferative effect, asnoted in a previous study [19], was also observed at the higherconcentration. We also evaluated the effects of fenofibrate on cellproliferation and membrane integrity using a trypan blue

exclusion assay, and no statistically significant differences wereobserved between 10 lM fenofibrate-treated cells and DMSO-trea-ted control cells (data not shown). The cell number of 20 lM fenof-ibrate-treated cells was reduced to 54.6 ± 9.4% of the control cellcounts (P < 0.001). Therefore, 10 lM fenofibrate was used for sub-sequent experiments.

3.2. PPARa does not mediate the effect of fenofibrate on melanogenicactivity

To further elucidate the mechanism of fenofibrate on melaninproduction, we sought to determine if the effects of fenofibrateare mediated through the activation of the PPARa receptor. Wetherefore examined the biological effects of other potent PPARaagonists, including a fibric acid derivative and GW7647. Fenofi-brate is metabolized to the active metabolite fenofibric acid byesterase following oral administration. Surprisingly, fenofibric acidfailed to diminish the melanin content at every concentrationtested (Fig. 2A), including 100 lM, the highest concentration thatwas studied (data not shown). The results shown in Fig. 2B indicatethat the addition of 10 lM GW7647 alone did not significantlydecrease melanogenesis. We also examined the effect ofGW6471, a specific inhibitor of the DNA binding of PPARa [22],on melanin levels. The suppression of melanin levels by fenofibratewas not altered by the GW6471 pretreatment. These results implythat the PPARa antagonist did not affect the fenofibrate-repressedmelanogenic activity. Interestingly, the liver X receptor (LXR)

Fig. 2. Effects of fenofibrate on melanin synthesis in B16-F10 melanoma cells. (A)The cells were treated with 1–20 lM fenofibrate or fenofibric acid for 48 h. (B) Thecells were treated with 10 lM fenofibrate or GW7647 in the presence or absence of0.1 lM GW6471 for 48 h. (C) The cells were treated with 10 lM fenofibrate in thepresence or absence of 0.1 lM T0901317 for 48 h. The cellular melanin content wasdetermined. Each value is expressed as a percentage relative to the DMSO-treatedcontrol. The data are reported as the means ± S.E.M. of at least three independentexperiments. *P < 0.05, ***P < 0.001 compared with the control. #P < 0.05 comparedwith fenofibrate-treated cells. NS denotes no significant difference.

Fig. 3. Effects of fenofibrate on intracellular tyrosinase activity in B16-F10melanoma cells. The cells were treated with 0.1–10 lM fenofibrate for 48 h. Theintracellular tyrosinase activity was determined by L-DOPA zymography. One ofthree similar zymogram results is shown, and the relative band intensity wasquantified. Each value is expressed as a percentage relative to the DMSO-treatedcontrol. *P < 0.05 compared with the control.

160 Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164

agonist, T0901317, could override the fenofibrate-induced repres-sion of melanogenesis (Fig. 2C). Thus, the biological effects offenofibrate do not seem to be dependent on the transcriptionalactivity of PPARa. We speculate that intrinsic LXR signaling mightcontribute to melanogenesis in B16-F10 cells.

3.3. Fenofibrate inhibits tyrosinase activity

Intracellular tyrosinase activity was confirmed by the DOPA-staining assay (Fig. 3). After densitometric analysis, we observed

that 10 lM fenofibrate significantly reduced tyrosinase activity toapproximately 54.6 ± 0.06% that of the control. To exclude the pos-sibility of a direct influence of fenofibrate on tyrosinase activity, weadded fenofibrate directly to an untreated cell lysate and measuredthe in vitro tyrosinase activity (data not shown). Fenofibrate didnot directly regulate tyrosinase activity in the in vitro system, butdid inhibit intracellular tyrosinase activity in B16-F10 cells.

3.4. Fenofibrate influences signaling pathways that mediatemelanogenesis

The expression of melanogenesis-related proteins is crucial tothe synthesis of pigments and is modulated by transcription fac-tors. Therefore, the cell lysates that were obtained after fenofibratetreatment were subjected to western blot analysis. We found thatthe protein expression of tyrosinase and MITF was repressed withincreasing concentrations of fenofibrate (Fig. 4). Quantification ofthe detected signal revealed that the protein levels of tyrosinaseand MITF following treatment with 10 lM fenofibrate were re-duced to 0.52 ± 0.08- and 0.73 ± 0.05-fold (P < 0.05, n = 3) of theDMSO-treated control, respectively.

Among the signal transduction pathways related to pigmenta-tion, the p38 mitogen-activated protein kinase (MAPK) pathwayis involved in the regulation of melanogenesis [23,24]. To furthercharacterize fenofibrate’s mechanism of action, we performed atime-course study using 10 lM fenofibrate. Fig. 5A shows thatfenofibrate induced early phosphorylation of p38 MAPK within1 h, and this phenomenon was sustained until 3 h. The phosphor-ylation of p38 MAPK persistent elevated even at 12- and 24-h offenofibrate treatment.

p38 MAPK might control tyrosinase stability by regulating itsproteasome-dependent degradation [24]. Therefore, we usedMG132, a membrane-permeable proteasome inhibitor, to assesswhether fenofibrate-dependent suppression of tyrosinase expres-sion is involved in the regulation of the ubiquitin–proteasomepathway. Accordingly, we observed that the fenofibrate-induceddecrease of tyrosinase expression was prevented by MG132 pre-treatment (Fig. 5B). To confirm the involvement of p38 MAPK sig-naling, the inhibitor SB203580 was used to assess how geneexpression was affected by fenofibrate. We examined the mRNA

Fig. 4. Effects of fenofibrate on the expression of melanogenesis-related proteins inB16-F10 melanoma cells. The cells were treated with the indicated concentrationsof fenofibrate for 48 h. Whole cell lysates were subjected to western blot analysisusing specific antibodies against tyrosinase and MITF. Equal protein loading wasconfirmed using an antibody against actin or GAPDH. Densitometric scanning ofband intensities obtained from three independent experiments was used toquantify the changes of protein expression. *P < 0.05, **P < 0.01 compared with thecontrol.

Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164 161

expression of tyrosinase and MITF. Quantiative RT-PCR analysis re-vealed that fenofibrate treatment slightly affected gene expressionof tyrosinase and MITF, about 0.70 ± 0.11- and 0.77 ± 0.15-folddecrease, respectively. We further evaluated the expression levelof MC1R. The results revealed that expression of MC1R mRNAwas significantly inhibited by fenofibrate after a 3-h treatment(Fig. 5C). However, the gene expression influenced by fenofibratewas not significantly changed in the presence of SB203580. Wealso determined the effect of SB203580 on fenofibrate-decreasedmelanin synthesis. According to Fig. 5D, the melanin count sup-pressed by fenofibrate was approximately reverted to the controllevel in the presence of SB203580. These results demonstrate thatthe activation of p38 MAPK signaling contributes to fenofibrate-mediated suppression of melanogenesis.

To examine the possible roles of oxidative stress on fenofibrate-mediated inhibitory effects of melanogenesis, the ROS scavenger,N-acetyl-L-cysteine (NAC), was accompanied with fenofibratetreatment. The results of melanin synthesis assay (Fig. 5D) dis-played that fenofibrate-decreased melanin contents was not coun-teracted by the NAC pretreatment. It seems that the intracellularROS generation is not involved in the inhibitory effect of fenofi-brate on melanogenesis.

Since the early phase of apoptosis may obscure the effects offenofibrate on melanogenesis, the expression of Bcl-2 family pro-teins were examined by the Western blots. Fig. 5E showed thatthe protein levels of Bax and Bcl-2 remained unchanged at anytime points of fenofibrate treatment. These results clarify thatthe effects of fenofibrate on the regulation of melanogenesis isnot related to the apoptotic events.

4. Discussion

PPARs and their corresponding ligands have been shown toregulate important physiological functions in the body. In human

melanocytes, activators of PPARa (WY-14643) and PPARc (ciglitaz-one) appear to stimulate melanogenesis by increasing tyrosinaseactivity rather than tyrosinase expression [25]. Grabacka et al.[19] showed that mouse B16-F10 cells express PPARa and thatthe transcriptional activation of PPARa by fenofibrate inhibitsmigration and anchorage-independent growth. Further work inmouse melanoma S91 cells indicated that activation of PPARcinduces events resembling differentiation. Exposure to ciglitazoneregulates MITF in a biphasic manner and induces tyrosinaseactivity [26]. In contrast with other PPARb/d or PPARc agonists,we showed that fenofibrate alone inhibited melanogenesis inB16-F10 cells. The decreased levels of melanin were accompaniedby a parallel decline in tyrosinase activity. Comparison of thephenomena observed in the present study with those observed inprevious research indicates that PPAR agonists affect differentregulatory pathways.

Fenofibrate has long been thought to act by binding to PPARa toachieve the therapeutic lowering of lipid levels. At a higher concen-tration (25 lM), fenofibrate showed a twofold stimulation of thePPAR responsive elements in B16-F10 cells [19]. According to thedata shown here, fenofibrate diminishes melanin levels, and thisphenomenon was not reversed by the selective PPARa antagonistGW6471. A high affinity compound for PPARa, GW7647, also didnot alter melanin synthesis. In addition, we observed that the ma-jor active metabolite of fenofibrate, fenofibric acid, failed to inhibitmelanin production. Fenofibric acid contains a carboxylic acidmoiety instead of an isopropyl ester moiety in fenofibrate. Thestructural analysis reveals that the carboxyl group aligns on thesame side as the ketone carbonyl group in fenofibric acid. However,the carboxyl group of fenofibrate is located away from the ketonecarbonyl group [27]. It can be inferred that the different orientationof functional groups in fenofibrate and fenofibric acid may be re-lated to the distinct effects of such chemicals interact with theirtarget molecules in melanoma cells. Nevertheless, the affinity ofthe ligands for each PPAR subtypes varies [16]. Recently, evidencehas shown that unmetabolized fenofibrate results in differentenzymatic activity than fenofibric acid [28]. A single PPAR ligandmay interact with more than one receptor subtype. It has also beenproposed that fenofibric acid is a PPARa agonist, while fenofibrateis a LXR antagonist [29]. The fibrates display greater affinity forLXRs than the corresponding fibric acids do for PPARa [30]. TheLXR agonist T0901317 restored the fenofibrate-induced suppres-sion of melanogenesis. The direct effects of fenofibrate on PPARacan be excluded because fenofibric acid and GW7647, which havehigher affinities for PPARa than fenofibrate, failed to decreasemelanin synthesis in B16-F10 cells. Therefore, we assume thatfenofibrate suppressed melanin synthesis via a PPAR-independentmechanism. Similarly, fenofibrate suppresses the growth of humanhepatocellular carcinoma cells and induces apoptosis in mantlelymphoma by a PPARa-independent mechanism [31,32]. A previ-ous report noted that the expression of the nuclear receptor LXRain perilesional human melanocytes is higher relative to itsexpression in the normal skin of a vitiligo patient [33]. It can bespeculated that LXRs are involved in the fenofibrate-mediated reg-ulation of melanin production. The B16-F10 melanoma cell linewith altered pigmentation signaling pathways is a well-establishedmodel for the study of melanocyte biology [34]. Further work to as-sess the effects of fenofibrate on normal human melanocytes orreconstituted human epidermis will clarify the physiological rele-vancy of fenofibrate.

Tyrosinase is the rate-limiting enzyme for melanogenesis, andits stability is regulated by endoplasmic reticulum-associatedprotein degradation or the endosomal/lysosomal degradation sys-tem [35]. Anisomycin, a well-characterized and relatively specificactivator of p38 MAPK, causes tyrosinase degradation in B16-F0melanoma cells and in normal human melanocytes. In addition,

Fig. 5. Effects of fenofibrate on the activation of the p38 MAPK pathway in B16-F10 melanoma cells. (A and E) The cells were treated with 10 lM fenofibrate for the indicatedperiods to assay for changes in the expression profile of phosphorylated p38 MAPK (P-p38), Bcl-2, and Bax. (B and C) Cells were treated with 10 lM fenofibrate in the presenceor absence of the proteasome inhibitor MG132 (150 nM) or the p38 MAPK inhibitor SB203580 (5 lM) for 3 h. Whole cell lysates were subjected to western blot analysis andequal protein loading was confirmed using an antibody against a-tubulin. The mRNA expression levels of tyrosinase, MITF, and MC1R were examined by real-time RT-PCR.Quantitative analysis of mRNA is displayed by relative values normalized to the GAPDH mRNA levels. (D) The cells were treated with fenofibrate in the presence or absence ofSB203580 or NAC (0.1 mM) for 48 h, and the cellular melanin contents were determined. *P < 0.05 compared with the control. #P < 0.05 compared with fenofibrate-treatedcells.

162 Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164

that degradation is attenuated by MG132 [24]. In the presentstudy, fenofibrate exhibited effects similar to those of anisomycin.The results suggest that fenofibrate regulates the melanogenesisthrough the post-translation modulation. Since fenofibrate alsoaffected the mRNA levels of tyrosinase, MITF, and MC1R, these datado not obviate involvement of transcriptional regulation in thefenofibrate-decreased tyrosinase expression. However, SB203580could not significantly restore the fenofibrate-induced decreaseof gene expression. It has been demonstrated that p38 silencingshowed no significant change in MITF mRNA transcription [24].Thus, our data on the effects of fenofibrate on the melanogenic en-zymes is partially consistent with the previous study. Recently,SB203580, a pyridinyl imidazole compound that specifically inhib-its p38a and p38b [36], has been shown to inhibit melanogenesisin a p38-independent manner [24,37]. It seems that the pyridinylimidazole class of compounds employed to study p38 MAPK-dependent melanogenesis is inadequate. Further studies usingsmall interfering RNA experiments will address the targets offenofibrate and resolve the inadequacies of chemical agents.

Western blots illustrated that fenofibrate down-regulated theprotein expression of tyrosinase and MITF while up-regulatingthe phosphorylation of p38 MAPK. Fenofibrate-induced inhibition

of melanin synthesis was restored by SB203580. B16-F10 cellsexpressing MC1R inhibited activation of p38 MAPK, but not ERKand JNK/SAPK, whereas B16-F10 cells expressing siRNA targetingMC1R showed enhanced phosphorylation of p38 MAPK [38]. It ispossible that fenofibrate through the perturbation of MC1R expres-sion to regulate p38 MAPK signaling in B16-F10 cells. Takentogether, these data suggest that fenofibrate suppresses melano-genesis by the down-regulation of MC1R, activation of the p38MAPK signaling pathway, and the repression of the MITF signalingpathway. Curcumin, through activation of phosphatidylinositol3-kinase (PI3K/Akt), extracellular signal-regulated kinase (ERK),or p38 MAPK signaling, inhibits melanogenesis in human melano-cytes [39]. During a 48 h incubation, 25 lM fenofibratesignificantly interferes with Akt and ERK1/2 signaling pathwaysin B16-F10 cells [19]. In the present study, we did not explorethe interactions of 10 lM fenofibrate with these signaling mole-cules. Further studies are needed to verify the relationships offenofibrate with these kinases.

Cellular redox may be key to oxidant-induced activation of thestress-activated protein kinase (SAPK) pathways [40]. p38 MAPK iswell accepted as a stress-stimulated MAPK and activated bydiverse stressors, such as ROS generation. Hydrogen peroxide

Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164 163

(H2O2) concentration in the mM range has been shown to increasep38 MAPK activity in B16-F10 cells [38]. Based on the results ofFig. 5D, fenofibrate-decreased melanin synthesis was not effi-ciently offset by NAC, a H2O2 scavenger, whereas SB203580 didit. Besides, treatment of fenofibrate did not show stimulatory effecton ROS generation as detected by 20,70-dichlorofluorescein diace-tate with flow cytometry analysis (data not shown). It is likely thatthe inhibitory effect of fenofibrate on melanogenesis is not relevantto the intracellular ROS generation. Animal studies demonstratedthat fenofibrate through PPARa activation significantly attenuatedoxidative damage in brain by modulating antioxidant defensemechanisms, such as level of reduced glutathione and enzymeactivity of catalase and superoxide dismutase [41]. Further eluci-dating mode of antioxidant activity will help to refine the redoxeffects of fenofibrate. A recent study showed that fenofibrate elicitsdual beneficial effects in retinal pigment epithelium by down-reg-ulation of stress-mediating signaling and induction of autophageand survival pathways [42]. Fenofibrate triggers Bim-mediatedapoptosis of glioblastoma cells in vitro [43]. We demonstrated thatprotein expression of Bcl-2 and Bax was not altered by the fenofi-brate. The effect of fenofibrate on melanogenesis is independent ofapoptosis.

Thiazolidinediones are a class of molecules that are chemicallyrelated to fibrates [44]. Mitochondrial dysfunction induced byfibrates and thiazolidinediones was response to the insulin-sensi-tizing effects of certain drugs [44,45]. Fenofibrate stimulated thephosphorylation of p38 MAPK and then inhibited the melanin syn-thesis. However, ciglitazone did not reveal the negative regulationof melanogenesis. Sphingosylphosphorylcholine-induced hypogig-mentation is involved in the inhibition of protein phosphatase 2A[46]. It is valuable to consider the role of protein phosphatase onthe modulation of melanogenesis. Due to the limited information,whether fenofibrate or ciglitazone regulates some protein phos-phatases to influence the activation of p38 MAPK pathways isunclear.

5. Conclusions

In summary, we have demonstrated that fenofibrate suppressesmelanogenesis through the down-regulation of the DOPA-stainingactivity and expression of tyrosinase as well as the down-regula-tion of MITF expression. To the best of our knowledge, this is thefirst time that fenofibrate has been shown to inhibit melanogenesisby down-regulating MC1R gene expression and activating the p38MAPK signaling pathway in a manner independent of PPAR activa-tion. The potential to influence skin cancer risks is relevant to thetherapeutic use of fenofibrate in conventional disease treatment.

Acknowledgements

This study was supported by Grants from the National ScienceCouncil of Republic of China (NSC98-2320-B-126-001-MY3) andChang Bing Show Chwan Memorial Hospital (CBSH10010010).

References

[1] J.Y. Lin, D.E. Fisher, Melanocyte biology and skin pigmentation, Nature 445(2007) 843–850.

[2] J. Vachtenheim, J. Borovansky, ‘‘Transcription physiology’’ of pigmentformation in melanocytes: central role of MITF, Exp. Dermatol. 19 (2010)617–627.

[3] P. Grimes, J.J. Nordlund, A.G. Pandya, S. Taylor, M. Rendon, J.P. Ortonne,Increasing our understanding of pigmentary disorders, J. Am. Acad. Dermatol.54 (2006) S255–S261.

[4] Q. Ma, Transcriptional responses to oxidative stress: pathological andtoxicological implications, Pharmacol. Ther. 125 (2010) 376–393.

[5] K.C. Wen, H.H. Chiu, P.C. Fan, C.W. Chen, S.M. Wu, J.H. Chang, H.M. Chiang,Antioxidant activity of Ixora parviflora in a cell/cell-free system and in UV-exposed human fibroblasts, Molecules 16 (2011) 5735–5752.

[6] K.U. Schallreuter, M.M. Salem, S. Hasse, H. Rokos, The redox–biochemistry ofhuman hair pigmentation, Pigment Cell Melanoma Res. 24 (2011) 51–62.

[7] M.M. Elahi, Y.X. Kong, B.M. Matata, Oxidative stress as a mediator ofcardiovascular disease, Oxid. Med. Cell. Longev. 2 (2009) 259–269.

[8] A.R. Shipman, G.W. Millington, Obesity and the skin, Br. J. Dermatol. 165(2011) 743–750.

[9] K.Y. Wolin, K. Carson, G.A. Colditz, Obesity and cancer, Oncologist 15 (2010)556–565.

[10] J.E. Gerich, Type 2 diabetes mellitus is associated with multiplecardiometabolic risk factors, Clin. Cornerstone 8 (2007) 53–68.

[11] S.R. Freeman, A.L. Drake, L.F. Heilig, M. Graber, K. McNealy, L.M. Schilling, R.P.Dellavalle, Statins, fibrates, and melanoma risk: a systematic review and meta-analysis, J. Natl. Cancer Inst. 98 (2006) 1538–1546.

[12] H.B. Rubins, S.J. Robins, D. Collins, C.L. Fye, J.W. Anderson, M.B. Elam, F.H. Faas,E. Linares, E.J. Schaefer, G. Schectman, T.J. Wilt, J. Wittes, Gemfibrozil for thesecondary prevention of coronary heart disease in men with low levels ofhigh-density lipoprotein cholesterol. Veterans Affairs High-DensityLipoprotein Cholesterol Intervention Trial Study Group, N. Engl. J. Med. 341(1999) 410–418.

[13] A.S. Boyd, Thiazolidinediones in dermatology, Int. J. Dermatol. 46 (2007) 557–563.

[14] M. Schmuth, Y.J. Jiang, S. Dubrac, P.M. Elias, K.R. Feingold, Thematic reviewseries: skin lipids. Peroxisome proliferator-activated receptors and liver Xreceptors in epidermal biology, J. Lipid Res. 49 (2008) 499–509.

[15] P. Sertznig, J. Reichrath, Peroxisome proliferator-activated receptors (PPARs) indermatology: challenge and promise, Dermatoendocrinology 3 (2011) 130–135.

[16] P. Sertznig, M. Seifert, W. Tilgen, J. Reichrath, Peroxisome proliferator-activated receptors (PPARs) and the human skin: importance of PPARs inskin physiology and dermatologic diseases, Am. J. Clin. Dermatol. 9 (2008) 15–31.

[17] S. Konda, A.N. Geria, R.M. Halder, New horizons in treating disorders ofhyperpigmentation in skin of color, Semin. Cutan. Med. Surg. 31 (2012) 133–139.

[18] L.L. Eastham, C.N. Mills, R.M. Niles, PPARalpha/gamma expression and activityin mouse and human melanocytes and melanoma cells, Pharm. Res. 25 (2008)1327–1333.

[19] M. Grabacka, P.M. Plonka, K. Urbanska, K. Reiss, Peroxisome proliferator-activated receptor alpha activation decreases metastatic potential ofmelanoma cells in vitro via down-regulation of Akt, Clin. Cancer Res. 12(2006) 3028–3036.

[20] Y.C. Huang, K.C. Liu, Y.L. Chiou, Melanogenesis of murine melanoma cellsinduced by hesperetin, a Citrus hydrolysate-derived flavonoid, Food Chem.Toxicol. 50 (2012) 653–659.

[21] K. Sato, H. Takahashi, R. Iraha, M. Toriyama, Down-regulation of tyrosinaseexpression by acetylsalicylic acid in murine B16 melanoma, Biol. Pharm. Bull.31 (2008) 33–37.

[22] H.E. Xu, T.B. Stanley, V.G. Montana, M.H. Lambert, B.G. Shearer, J.E. Cobb, D.D.McKee, C.M. Galardi, K.D. Plunket, R.T. Nolte, D.J. Parks, J.T. Moore, S.A. Kliewer,T.M. Willson, J.B. Stimmel, Structural basis for antagonist-mediatedrecruitment of nuclear co-repressors by PPARalpha, Nature 415 (2002) 813–817.

[23] S. Corre, A. Primot, E. Sviderskaya, D.C. Bennett, S. Vaulont, C.R. Goding, M.D.Galibert, UV-induced expression of key component of the tanning process, thePOMC and MC1R genes, is dependent on the p-38-activated upstreamstimulating factor-1 (USF-1), J. Biol. Chem. 279 (2004) 51226–51233.

[24] B. Bellei, V. Maresca, E. Flori, A. Pitisci, L. Larue, M. Picardo, P38 regulatespigmentation via proteasomal degradation of tyrosinase, J. Biol. Chem. 285(2010) 7288–7299.

[25] H.Y. Kang, E. Chung, M. Lee, Y. Cho, W.H. Kang, Expression and function ofperoxisome proliferator-activated receptors in human melanocytes, Br. J.Dermatol. 150 (2004) 462–468.

[26] M. Grabacka, W. Placha, K. Urbanska, P. Laidler, P.M. Płonka, K. Reiss, PPARgamma regulates MITF and beta-catenin expression and promotes adifferentiated phenotype in mouse melanoma S91, Pigment Cell MelanomaRes. 21 (2008) 388–396.

[27] N.P. Rath, W. Haq, G.K. Balendiran, Fenofibric acid, Acta Crystallogr. C 61(2005) o81–o84.

[28] R. Murakami, H. Murakami, H. Kataoka, X.W. Cheng, R. Takahashi, Y.Numaguchi, T. Murohara, K. Okumura, Unmetabolized fenofibrate, but notfenofibric acid, activates AMPK and inhibits the expression ofphosphoenolpyruvate carboxykinase in hepatocytes, Life Sci. 87 (2010) 495–500.

[29] C.H. Kim, R. Ramu, J.H. Ahn, M.A. Bae, Y.S. Cho, Fenofibrate but not fenofibricacid inhibits 11beta-hydroxysteroid dehydrogenase 1 in C2C12 myotubes,Mol. Cell Biochem. 344 (2010) 91–98.

[30] J. Thomas, K.S. Bramlett, C. Montrose, P. Foxworthy, P.I. Eacho, D. McCann, G.Cao, A. Kiefer, J. McCowan, K.L. Yu, T. Grese, W.W. Chin, T.P. Burris, L.F. Michael,A chemical switch regulates fibrate specificity for peroxisome proliferator-activated receptor a (PPARa) versus liver X receptor, J. Biol. Chem. 278 (2003)2403–2410.

[31] D. Yamasaki, N. Kawabe, H. Nakamura, K. Tachibana, K. Ishimoto, T. Tanaka, H.Aburatani, J. Sakai, T. Hamakubo, T. Kodama, T. Doi, Fenofibrate suppresses

164 Y.-C. Huang et al. / Chemico-Biological Interactions 205 (2013) 157–164

growth of the human hepatocellular carcinoma cell via PPARa-independentmechanisms, Eur. J. Cell Biol. 90 (2011) 657–664.

[32] Z. Zak, P. Gelebart, R. Lai, Fenofibrate induces effective apoptosis in mantle celllymphoma by inhibiting the TNFa/NF-jB signaling axis, Leukemia 24 (2010)1476–1486.

[33] R. Kumar, D. Parsad, D. Kaul, A. Kanwar, Liver X receptor expression in humanmelanocytes, does it have a role in the pathogenesis of vitiligo?, Exp Dermatol.19 (2010) 62–64.

[34] V. Maresca, E. Flori, E. Camera, B. Bellei, N. Aspite, M. Ludovici, C. Catricalà, G.Cardinali, M. Picardo, Linking aMSH with PPARc in B16-F10 melanoma,Pigment Cell Melanoma Res. 26 (2013) 113–127.

[35] H. Ando, H. Kondoh, M. Ichihashi, V.J. Hearing, Approaches to identifyinhibitors of melanin biosynthesis via the quality control of tyrosinase, J.Invest. Dermatol. 127 (2007) 751–761.

[36] Y. Kuma, G. Sabio, J. Bain, N. Shpiro, R. Márquez, A. Cuenda, BIRB796 inhibits allp38 MAPK isoforms in vitro and in vivo, J. Biol. Chem. 280 (2005) 19472–19479.

[37] B. Bellei, A. Pitisci, E. Izzo, M. Picardo, Inhibition of melanogenesis by thepyridinyl imidazole class of compounds: possible involvement of the Wnt/b-catenin signaling pathway, PLoS One 7 (2012) e33021.

[38] H. Chung, J.H. Lee, D. Jeong, I.O. Han, E.S. Oh, Melanocortin 1 receptor regulatesmelanoma cell migration by controlling syndecan-2 expression, J. Biol. Chem.287 (2012) 19326–19335.

[39] C.X. Tu, M. Lin, S.S. Lu, X.Y. Qi, R.X. Zhang, Y.Y. Zhang, Curcumin inhibitsmelanogenesis in human melanocytes, Phytother. Res. 26 (2012) 174–179.

[40] V. Adler, Z. Yin, K.D. Tew, Z. Ronai, Role of redox potential and reactive oxygenspecies in stress signaling, Oncogene 18 (1999) 6104–6111.

[41] D.K. Bhateja, D.K. Dhull, A. Gill, A. Sidhu, S. Sharma, B.V. Reddy, S.S. Padi,Peroxisome proliferator-activated receptor-a activation attenuates 3-nitropropionic acid induced behavioral and biochemical alterations in rats:possible neuroprotective mechanisms, Eur. J. Pharmacol. 674 (2012) 33–43.

[42] S. Miranda, Á. González-Rodríguez, M. García-Ramírez, J. Revuelta-Cervantes,C. Hernández, R. Simó, Á.M. Valverde, Beneficial effects of fenofibrate in retinalpigment epithelium by the modulation of stress and survival signaling underdiabetic conditions, J Cell Physiol. 227 (2012) 2352–2362.

[43] A. Wilk, K. Urbanska, M. Grabacka, J. Mullinax, C. Marcinkiewicz, D. Impastato,J.J. Estrada, K. Reiss, Fenofibrate-induced nuclear translocation of FoxO3Atriggers Bim-mediated apoptosis in glioblastoma cells in vitro, Cell Cycle 11(2012) 2660–2671.

[44] R. Scatena, P. Bottoni, B. Giardina, Mitochondria, PPARs, and cancer: isreceptor-independent action of PPAR agonists a key?, PPAR Res (2008),http://dx.doi.org/10.1155/2008/256251.

[45] A.S. Divakaruni, S.E. Wiley, G.W. Rogers, A.Y. Andreyev, S. Petrosyan, M.Loviscach, E.A. Wall, N. Yadava, A.P. Heuck, D.A. Ferrick, R.R. Henry, W.G.McDonald, J.R. Colca, M.I. Simon, T.P. Ciaraldi, A.N. Murphy, Thiazolidinedionesare acute, specific inhibitors of the mitochondrial pyruvate carrier, Proc. Natl.Acad. Sci. USA 110 (2013) 5422–5427.

[46] H.S. Jeong, K.C. Park, D.S. Kim, PP2A and DUSP6 are involved insphingosylphosphorylcholine-induced hypopigmentation, in: Mol. Cell.Biochem. 367 (2012) 43–49.