Alpha-Melanocyte–Stimulating Hormone Suppresses Oxidative Stress through a p53-Mediated Signaling Pathway in Human Melanocytes

The melanocortin 1 receptor (MC1R) is best known for its role in stimulating eumelanin synthesis in melanocytes (MC). In the skin, presence of eumelanin is considered the major photoprotective factor against UV radiation (UVR)-induced DNA damage and carcinogenesis. Eumelanin creates not only a physical barrier against the deep penetration of UVR into the epidermal and dermal layers (1) but also reduces oxidative stress by scavenging free radicals (2–4). MC1R signaling also plays important roles in reducing the burden of UVR-induced DNA damage. This is affected by enhancing the repair of DNA photoproducts (5, 6) and by inhibiting the generation of reactive oxygen species (ROS), as well as by upregulating the activity and expression of antioxidant enzymes (7, 8).

We have reported that in vitro exposure of cultured melanocytes to UVR results in increased ROS generation and oxidative DNA damage (7). Reactive oxygen harms DNA by producing lesions such as damaged bases, apurinic-apyrimidinic sites, and strand breaks (9–11). These lesions, if not repaired, induce mutations and chromosomal instability, events associated with the process of carcinogenesis (12). Several studies relate ROS and oxidative stress to increased susceptibility to melanoma (13–16). Loss of CDKN2A/p16, a high-penetrance melanoma susceptibility gene, diminishes the capability of melanocytes to counteract ROS (17). The findings that loss of function of MC1R and loss of p16 are associated with increased susceptibility to melanoma suggest that oxidative stress is an underlying mechanism contributing to melanoma formation (18, 19).

This study aimed at elucidating the signaling pathways activated by the MC1R that mediate the antioxidant effect of α-melanocyte–stimulating hormone (α-MSH) in melanocytes. We show that the tumor suppressor p53, the universal sensor of cellular stress and regulator of the DNA damage response, is one target of the MC1R signaling pathway and a mediator of the antioxidant effects of α-MSH. Our results unraveled a mechanism involving α-MSH and p53 in the adaptive response of melanocytes to UVR. Others have shown that p53 increases the expression of the melanogenic enzymes tyrosinase and tyrosine-related protein 1 (TYRP-1), suggesting its role in regulating melanogenesis (20, 21). In mouse skin, p53 stimulates the expression of proopiomelanocortin (POMC), the precursor of α-MSH and adrenocorticotropic hormone (ACTH; refs. 22, 23). Collectively, findings by others and the results that we herein report underscore the significance of p53 in the DNA damage response of epidermal cells.

Primary cultures of melanocytes were established from discarded neonatal foreskins as described previously (24). The Institutional Review Board at the University of Cincinnati (Cincinnati, OH) has deemed the protocol for obtaining the skin samples as exempt. The primary culture of melanocytes derived from a patient with ocular-cutaneous albinism type 2 (OCA2) that expressed inactivating mutation of the P gene was a generous gift from Dr. Raymond Boissy (Department Of Dermatology, University of Cincinnati). These cells express functional MC1R but are deficient in melanin production due to impairment of P gene. All the melanocyte cultures were maintained in medium containing growth factors, except for the experiments carried out with α-MSH, in which the cells were kept in medium devoid of bovine pituitary extract (BPE) known to contain significant levels of α-MSH.

Melanocytes naturally expressing loss-of-function MC1R gene (genotype R160W/D294H) were transfected with pcDNA3 containing the wild-type MC1R gene (pMC1R-WT, a gift from Dr. Garcia-Borron, University of Murcia, Murcia, Spain), as described earlier (8).

Melanocytes were irradiated with a bank of FS-20 lamps, with 75% UVB and 25% UVA emissions. UVC rays were blocked by Kodacel filter (Eastman Kodak). Cells were irradiated with a dose of 105 mJ/cm² UVR unless otherwise specified, as described before (8). For the experiments using H₂O₂, cells were washed in PBS and incubated for 30 minutes at 37°C in the presence of 100 μmol/L of H₂O₂, or another concentration of H₂O₂ as indicated, and then replaced by fresh medium.

Single-cell electrophoresis was conducted to determine the extent of DNA damage induced by UVR or H₂O₂, as previously described by Song and colleagues (7). Briefly, melanocytes kept in medium deprived of BPE (−BPE) were pretreated with 1 nmol/L α-MSH for 4 days before 105 mJ/cm² UVR or pulse treatment with 100 μmol/L H₂O₂ (unless otherwise indicated). Two hours after UVR or H₂O₂, cells embedded in 1% low–melting point agarose gel were subjected to electrophoresis. Under these conditions, the undamaged DNA remains in the nucleus and the DNA containing strand breaks (alkali-labile sites) streams toward the anode. The extension of each tail moment, defined as the product of DNA in the tail and the mean distance of its migration, was analyzed using a computerized image analysis system (TriTek CometScore Freeware). The tail moment values obtained from a minimum of 50 randomly selected cells from each slide were expressed as the mean value ± SEM, and data were analyzed by ANOVA followed by Student–Newman–Keuls (SNK; P ≤ 0.05).

Melanocytes were plated at a density of 1 × 10⁶ cells, irradiated 4 days later with 105 mJ/cm² UVR, and treated with 1 nmol/L α-MSH immediately thereafter. The release of H₂O₂ was measured at different time points after UVR, as described (7). Data were expressed as mean of H₂O₂ in pmol/mL/10⁶ cells ± SEM, and data were analyzed by ANOVA followed by SNK (P ≤ 0.05).

Melanocytes were treated as indicated for Comet assay, and cell extracts were obtained at the indicated time points after UVR or H₂O₂ exposure. Western blotting was carried out using the following antibodies: p53 (DO-1) anti-mouse (Santa Cruz); phospho-p53 serine 15 anti-rabbit (Cell Signaling); phospho-p38 Thr180/Tyr182 (Cell Signaling); p21 (Waf1/Cip1) anti-mouse (Cell Signaling); GADD45α anti-mouse (Santa Cruz); phospho-ATR Ser428 anti-rabbit (Cell Signaling); DNA-PK anti-mouse (Santa Cruz); phospho-DNA-PK anti-rabbit (Santa Cruz); OGG1anti-rabbit (Novus Biologicals); APE-1/Ref-1 anti-mouse (Novus Biologicals); and actin horseradish-conjugated (Santa Cruz); pan-p63 antibody (4A4) against all p63 isoforms and specific antibody anti-alpha-p63 (H-129). Aliquots of 50 μg from total cell lysates obtained with radioimmunoassay buffer supplemented with protease inhibitors (RIPA buffer) were used for Western blotting, as described (25).

To determine the effect of α-MSH on p53 phosphorylation and activation, melanocytes kept in −BPE medium were pretreated with 0 and 2 μmol/L H-89, a PKA inhibitor (Sigma Chemical Co.), before UVR exposure. Western blotting of p53 phosphorylated on serine 15 (p53-Ser15) was conducted 1 hour after UVR.

Melanocytes were kept in medium without any growth factors for 18 hours before exposure to UVR. Expression of phospho-p38 was determined by Western blotting 1 hour after UVR. The role of p38 in the phosphorylation of p53 was determined by pretreating melanocytes with 10 μmol/L SB203580, a p38 inhibitor (Sigma Chemical Co.), for 18 hours before UVR and 1.5 hours after irradiation.

Melanocytes rescued for wild-type MC1R expression were plated at the density of 0.15 × 10⁵ cells per coverslip and treated for 4 days with 1 nmol/L α-MSH before irradiation with 105 mJ/cm² UVR. Cells were kept in fresh medium supplemented with 0 or 1 nmol/L α-MSH for 2 hours after UVR and then fixed with acetone:methanol (1:1) for 10 minutes. Coverslips were incubated with 10% bovine serum albumin (BSA) in PBS for 1 hour at room temperature, followed by 1-hour incubation at 37°C with antibody against p53 (1:200) and anti-mouse fluorescein isothiocyanate (FITC)-conjugated antibody (1:200; Jackson Laboratories). Immunofluorescence was visualized using a fluorescence microscope and images acquired with a computer-assisted program.

To determine the effect of p53 inhibition on oxidative DNA damage, melanocytes were preincubated with 10 μmol/L pifithrin-α (PFT; Sigma Chemical Co.), an inhibitor of p53, for 18 hours before and immediately after UVR or 30-minute pulse-treated with 50 and 75 μmol/L H₂O₂, and the effect on DNA damage determined 2 hours later by Comet assay. The effect of p53 accumulation on oxidative DNA damage was determined by pretreating melanocytes expressing loss-of-function MC1R with 10 μmol/L nutlin-3 (Sigma Chemical Co.), for 18 hours before exposure to 105 mJ/cm² UVR or 100 μmol/L H₂O₂. The tail moment values obtained from a minimum of 50 randomly selected cells from each slide were expressed as the mean value ± SEM, and data were analyzed by ANOVA followed by SNK (P < 0.05).

Neonatal melanocytes (hMC1, hMC2, and hMC3) were used for silencing of p53 with 1.25 × 10⁵ (Infectious Units) short hairpin RNA (shRNA) lentiviral particles (Santa Cruz Biotechnology), according to manufacturer's instructions. Nontargeted scrambled sequence shRNA was used as transduction control and stable infections were obtained with puromycin selection. Confirmation of p53 silencing was verified by Western blotting using p53 antibody.

Melanocytes were plated at a density of 1 × 10⁶ cells per 100-mm dish and kept in medium with growth factors. Eighteen hours before and immediately after irradiation with 50, 75, and 105 mJ/cm² UVR, cells were treated with 10 μmol/L nutlin-3. Two and 4 hours after irradiation, cells were fixed with ice-cold 1% paraformaldehyde for 15 minutes and 70% ethanol for 1 hour. After permeabilization with 0.2% Triton X-100 in 3% BSA/PBS for 10 minutes, nonspecific binding was blocked with 20% normal goat serum for 30 minutes at room temperature. Cells were incubated with 1 μg/mL anti-γ-H2AX antibody [anti-phospho histone H2AX (Ser 139; Millipore)] overnight at 4°C, followed by goat anti-mouse IgG Alexa Fluor 488 (1:200; Invitrogen) for 1 hour at room temperature. After 2 washes with 0.1% Triton X-100 in 1% BSA/PBS and one with PBS, cells were resuspended in PBS containing 0.11 mg/mL RNase A and 6.85 μg/mL propidium iodide and 30 minutes later analyzed by flow cytometer.

We have shown that α-MSH decreases UVR- or H₂O₂-induced oxidative DNA damage in melanocytes (7). To further show that this inhibitory effect is mediated by activation of the MC1R, we compared the responses of melanocytes naturally expressing nonfunctional receptor with that of the same culture stable transfected with the wild-type MC1R. Treatment with α-MSH immediately after UVR had no significant effect on the melanocytes lacking functional MC1R (Fig. 1A and B) whereas forskolin, a direct activator of adenylyl cyclase, markedly reduced the UVR-induced H₂O₂ production by 30% and 55% at 0 and 30 minutes, respectively. Similar effects were observed in cells expressing functional MC1R treated with forskolin, with reductions of 42% and 46% at 0 and 30 minutes, respectively (Fig. 1C and D). Transfection with the wild-type MC1R (Fig. 1C) not only reduced 6-fold the levels of H₂O₂ generated after UVR compared with cells expressing nonfunctional MC1R (Fig. 1A) but also rescued the responsiveness to α-MSH, as evidenced by the 35%, 51%, and 45% reductions of the UVR-induced H₂O₂ generation at 15, 30, and 45 minutes, respectively.

To exclude the possibility that the observed protective effect of α-MSH was due to an increase in melanin synthesis, we used melanocytes derived from a patient with OCA2, characterized by a mutation in the P gene, but functional MC1R. Mutation in the P gene results in lack of synthesis of melanin and albino phenotype (23). In the OCA2-MCs α-MSH reduced the UVR-induced H₂O₂ generation by 30%, 33%, 44%, and 62% at 0, 15, 30, and 45 minutes, respectively (Fig. 1E and F). Using the Comet assay, OCA2-MCs treated with 1 nmol/L α-MSH for 4 days before and 2 hours after UVR showed a 33% reduction in the total amount of DNA damage induced by UVR and 85% decrease in the amount of oxidative DNA damage induced 2 hours after acute exposure to H₂O₂ (Fig. 1G).

Irradiation of melanocytes with 105 mJ/cm² UVR resulted in increased levels of p53 and p53-Ser15, which were further augmented by 1 nmol/L α-MSH in melanocytes expressing functional, but not in melanocytes expressing loss-of-function MC1R (Fig. 2A). The cellular distribution of p53, determined by immunofluorescence, showed that treatment with α-MSH increased the accumulation and translocation of p53 to the nuclei of UVR-irradiated melanocytes (Fig. 2B). The translocation of p53 was consistent with an increase in its transcriptional activity, as evidenced by the greater induction of the p53 targets p21 and GADD45 (Fig. 2C). Because activation of the MC1R increases cyclic AMP (cAMP) levels, we tested the effect of H-89, an inhibitor of the PKA, on p53 phosphorylation. Cells treated with 1 nmol/L α-MSH showed a 69% increase in p53-Ser15 as compared with control. A modest 22% increase was observed in cells exposed to UVR and treated with α-MSH, compared with untreated irradiated cells. Pretreatment of melanocytes with 2 μmol/L H-89 for 18 hours before and 1.5 hours after UVR caused a significant reduction of p53-Ser15 in the UV-irradiated group treated with α-MSH compared with the group treated similarly in the absence of H-89 (Fig. 2D). To rule out the possibility that the effects of α-MSH occurred through activation of other p53 family members, the expression of p63 isoforms was investigated by Western blotting. The results indicate that melanocytes do not express any of the p63 proteins (Supplementary Fig. S1) or p73 (data not shown) because no detection was observed even at high protein concentrations (≥70 μg).

The mitogen-activated protein kinase (MAPK) p38 induced by stress phosphorylates p53 on serine 15. Our results showed that p38 was phosphorylated by UVR as shown by the 2-fold increase above control. This phosphorylation was further enhanced to 3-fold above control in the presence of 1 nmol/L α-MSH (Fig. 3A). Pretreatment with SB203580, a p38 inhibitor, abrogated the increase in p53 phosphorylation that was induced by α-MSH and UVR (Fig. 3B). We found that ATR, another upstream activator of p53, was phosphorylated by pulse treatment with H₂O₂ in a time-dependent manner, and a transient enhancement of phosphorylation (30 minutes) was observed in the presence of α-MSH (Fig. 3C). Check 2 (Chk2), the kinase downstream of ATR, was phosphorylated rapidly and transiently (30 minutes) after treatment with H₂O₂. However, in the presence of α-MSH, the phosphorylation of Chk2 was maintained for at least for 1 hour after H₂O₂ treatment (Fig. 3C). Similarly, α-MSH enhanced the phosphorylation of DNA-PK by more than 3-fold above control in melanocytes exposed to UVR (Fig. 3D).

Treatment of melanocytes with increasing doses of nutlin-3, a small molecule that blocks the HDM2-binding site to p53, induced a dose-dependent accumulation of p53 (Supplementary Fig. S2). Treatment of melanocytes expressing loss-of-function MC1R with nutlin-3 resulted in similar effects as of α-MSH on melanocytes expressing functional MC1R (Fig. 1G), namely, significant reduction of UV- and H₂O₂-induced DNA damage, as determined by Comet assay (Fig. 4A). Moreover, 10 μmol/L nutlin-3 markedly reduced γ-H2AX formation in melanocytes irradiated with increasing doses of UVR (Supplementary Fig. S3). To further show the significance of p53 in reducing oxidative stress in melanocytes, we investigated the effect of PFT, an inhibitor of p53, on the generation of H₂O₂ following UVR and DNA damage induced by UVR and H₂O₂. Suppression of p53 by 10 μmol/L PFT increased H₂O₂ generation by UVR-exposed melanocytes (Supplementary Fig. S4) and produced more DNA damage in these melanocytes (1.5-fold) and in melanocytes exposed to 50 and 75 μmol/L H₂O₂ (4-fold), as shown by the Comet assay (Fig. 4B). In addition, p53 was silenced using shRNA in 3 different melanocyte primary cultures, downregulation of p53 was verified at the functional level by reduced induction of p21 (Supplementary Fig. S5). In melanocytes silenced for p53, α-MSH failed to reduce the levels of UVR-induced H₂O₂ (Fig. 4E and F). Moreover, the initial levels of H₂O₂ generated immediately after UVR exposure (Fig. 4E) were higher than in the shRNA control cells (Fig. 4C). As expected, melanocytes transduced with the nontargeted sequence responded to α-MSH with a significant decrease in UVR-induced H₂O₂ (Fig. 4C and D). In p53-silenced melanocytes, as well as in melanocytes treated with PFT, the H₂O₂- or UVR-induced oxidative DNA damage was not reduced by α-MSH (Fig. 4B and G).

To determine the mechanism by which p53 regulates the repair of oxidative DNA damage, we investigated the expression of OGG1 and APE-1/Ref-1, enzymes essential for base excision repair (BER). The levels of both enzymes increased upon treatment with 100 μmol/L H₂O₂ as well as with 1 n mol/L α-MSH (Fig. 5A and B). In contrast to melanocytes expressing p53, in melanocytes silenced for p53, α-MSH failed to increase the levels of OGG1 or its translocation to the nucleus, as well as prevented the nuclear increase in APE-1, as indicated by the nuclear:cytoplasmic ratios of both proteins (Fig. 5C).

The results of our studies unraveled a novel mechanism by which α-MSH, through activation of the MC1R, protects melanocytes from oxidative DNA damage. We show for the first time that activation of the MC1R by α-MSH resulted in increased phosphorylation of p53 on serine 15 in UVR-irradiated melanocytes that leads to p53 accumulation and activation. This increase in p53 was significant in reducing ROS generation, and hence oxidative DNA damage, as well as in regulating the expression of enzymes that have a critical role in BER, the main repair pathway for oxidative DNA damage.

Our results clearly show that reduction of UV-induced H₂O₂ generation and increase in p53 levels by α-MSH treatment required functional MC1R, as these responses were absent in melanocytes expressing loss-of-function receptor. These findings strongly suggest that loss-of-function MC1R, which was associated with diminished accumulation of p53 after UVR exposure renders melanocytes more vulnerable to oxidative stress, increasing the risk for malignant transformation to melanoma. Interestingly, the reduction in oxidative DNA damage by α-MSH was independent of pigmentation because it was present in OCA-MCs that lack the ability to synthesize melanin.

Accumulation of p53 was accompanied by increased phosphorylation of p53 on serine 15 (Fig. 2A), a requirement for its nuclear translocation and transcriptional activity (ref. 26; Fig. 2B and C). Activation of the cAMP signaling pathway was necessary for further phosphorylation of p53 on Ser15, as inhibition of the cAMP-dependent PKA abrogated this effect (Fig. 2D). Similar to findings in melanoma cells and melanocytes in which α-MSH activates the MAPK p38 (27–29), we found that α-MSH increased the UV-induced phosphorylation of the p38 (Fig. 3A), and inhibition of p38 phosphorylation abrogated the increased accumulation of p53 (Fig. 3B). Treatment of melanocytes with α-MSH also increased the phosphorylation of the phosphoinositide 3-kinase (PI3K) ATR and DNA protein kinase (DNA-PK; Fig. 3C and D), the early sensors of DNA damage and activators of p53 (26, 30, 31).

To verify whether p53 is the only player in the antioxidant responses, we examined whether melanocytes express other isoforms of the p53 family members, that is, p63 proteins. However, unlike keratinocytes, melanocytes do not express any of the p63 isoforms (Supplementary Fig. S1).

The significance of p53 in reducing oxidative DNA damage was further confirmed by using 2 strategies to manipulate p53 accumulation: (i) inhibiting its degradation with nutlin-3, an inhibitor of the ubiquitin ligase HDM2 and (ii) inhibiting p53 pharmacologically using PFT, or downregulating p53 expression by shRNA. Treatment of melanocytes expressing loss-of-function MC1R with nutlin-3 had similar effects as treatment of melanocytes expressing functional MC1R with α-MSH, namely, significant reduction in UVR- and H₂O₂-induced oxidative DNA damage (Fig. 4A). The nutlin-3–reduced DNA damage was consistent with the decreased expression of γ-H2AX, an established marker of DNA damage (ref. 32; Supplementary Fig. S3). On the other hand, treatment with PFT (Fig. 4B) or downregulation of p53 by shRNA (Fig. 4E–G) led to a sustained UVR-induced increase in H₂O₂ levels and DNA damage and completely prevented the effect of α-MSH on DNA repair. In addition, silencing of p53 abolished the effects of α-MSH on the increase in OGG1 and APE-1 (Fig. 5C), the major enzymes involved in BER (33, 34). Our results are in agreement with other reports that accumulation of p53 in melanocytes or melanoma cells induces genes that regulate cell-cycle arrest rather than apoptosis and increases the expression of OGG1 in human fibroblasts (35, 36).

It is possible that the regulation of p53 activity and function in melanocytes differ from other cells, in as much as p53 also has important roles that are specific to melanocytes. It has been shown that p53 regulates the expression of genes that play important roles in melanogenesis, such as tyrosinase and TRP-1 (20, 37). Recently, it has been reported that in UV-irradiated mouse skin, p53 increases the expression of the proopiomelanocortin gene that encodes for the precursor of α-MSH and ACTH, the agonists for the human MC1R (22, 23, 38). We hereby present evidence for a new role of p53 in counteracting oxidative damage and in maintaining homeostasis of melanocytes. These findings provide more understanding of the role of the MC1R as a melanoma susceptibility gene and how loss of function of the MC1R increases the risk for melanoma. We propose that activation of the p53 pathway could potentially be used as a melanoma preventative strategy. The fact that melanoma tumors express wild-type p53 (39) offers the unique opportunity of using alternative strategies based on targeting p53 for the treatment of advanced stages of the disease.

No potential conflicts of interests were disclosed.

Conception and design: A.L. Kadekaro, Z. Abdel-Malek

Development of methodology: A.L. Kadekaro

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.L. Kadekaro, J. Chen, J. Yang, S. Chen, J. Jameson, V.B. Swope, T. Cheng, M. Kadakia, Z. Abdel-Malek

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.L. Kadekaro, J. Chen, S. Chen, V.B. Swope

Writing, review, and/or revision of the manuscript: A.L. Kadekaro, J. Chen, V.B. Swope, M. Kadakia, Z. Abdel-Malek

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.L. Kadekaro, J. Chen, Z. Abdel-Malek

Study supervision: A.L. Kadekaro

The authors thank Dr. George Babcock, Department of Surgery and Department of Pathology, University of Cincinnati and Shriners Hospital for Children (Cincinnati, OH) for the use of the flow cytometric facility.

This work was supported by the NIEHS (R21ES016900 to A.L. Kadekaro; RO1ES009110 to Z. Abdel-Malek), Dermatology Foundation Research Grant (A.L. Kadekaro) and NGBI-CEG (P30-ES 006096, Center for Environmental Genetics, University of Cincinnati to A. L. Kadekaro).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.