MITF and UV responses in skin: from pigmentation to addiction

Abstract

Ultraviolet radiation (UVR) has numerous effects on skin, including DNA damage, tanning, vitamin D synthesis, carcinogenesis, and immunomodulation. Keratinocytes containing damaged DNA secrete both α-melanocyte-stimulating hormone (α-MSH), which stimulates pigment production by melanocytes, and the opioid β-endorphin, which can trigger addiction-like responses to UVR. The pigmentation (tanning) response is an adaptation that provides some delayed protection against further DNA damage and carcinogenesis, while the opioid response may be an evolutionary adaptation for promoting sun-seeking behavior to prevent vitamin D deficiency. Here we review the pigmentation response to UVR, driven by melanocytic microphthalmia-associated transcription factor (MITF), and evidence for UVR-induced melanomagenesis and addiction. We also discuss potential applications of a novel approach to generate protective pigmentation in the absence of UVR (sunless tanning) using a topical small molecule inhibitor of the salt inducible kinase (SIK) family.

1. INTRODUCTION

Ultraviolet radiation (UVR), from either sunlight or indoor tanning devices, is considered a predominant environmental risk factor for melanoma as well as non-melanoma skin cancers. It can contribute to carcinogenesis through induction of DNA damage and suppression of immunological responses. The tanning pathway promoting pigmentation synthesis has been well established as the skin’s major photo-defensive mechanism against acute and chronic UVR exposure. Frequent and excessive skin contact with UV potentially leads to sun-seeking addictive behavior. In this review, we focus primarily on the effects of UV on the tanning response, melanomagenesis, and addiction. A novel application of topical strategy to trigger sunless tanning and its clinical relevance are also discussed.

2. MELANOGENESIS: SYNTHESIS OF PHEOMELANIN AND EUMELANIN

An epidermal melanin unit (EMU) consists of the melanocyte and approximately 36 to 40 associated keratinocytes that contact the melanocyte’s multiple dendrites (Chen, Weng, & Fisher, 2014; Jimbow, Quevedo, Fitzpatrick, & Szabo, 1976). Melanocytes contain subcellular lysosome-like organelles termed melanosomes, which are responsible for melanogenesis (Seiji, Fitzpatrick, & Birbeck, 1961; Park, Kosmadaki, Yaar, & Gilchrest, 2009). After melanin is synthesized, mature melanosomes are transported along melanocyte microtubules and transferred to neighboring basal keratinocytes (Wolff, 1973; Tadokoro & Takahashi, 2017). These melanosomes form a layer to protect keratinocytes’ genetic material from being damaged by UVR (Boissy, 2003).

In mammalians, melanogenesis is a synthetically complex pathway involving a series of enzymatic and non-enzymatic chemical reactions to produce two distinct types of melanin: reddish-yellow pheomelanin and brownish-black eumelanin (Simon, Peles, Wakamatsu, & Ito, 2009) (Figure 1). The first step in the melanin synthesis pathway is the rate-limiting oxidation of L-tyrosine, first to L-dihydroxyphenylalanine (L-DOPA), and then rapidly to dopaquinone, with both steps catalyzed by tyrosinase (TYR). Dopaquinone, serving as a common precursor to both pheomelanin and eumelanin, is a highly reactive chemical intermediate (Land, Ramsden, & Riley, 2001). In the presence of cysteine, dopaquinone reacts with cysteine to form 2, 5- or 5, 5-cysteinyldopa, which oxidizes to give rise to cysteinyldopa-quinones, then benzothiazine intermediates and subsequently produce soluble reddish-yellow pigment, pheomelanin (Land & Riley, 2000; Land et al., 2001). In addition to cysteine, glutathione, a tripeptide component of cysteine, glutamate, and glycine, can also combine with dopaquinone to form pheomelanin (Jara, Aroca, Solano, Martinez, & Lozano, 1988; Rorsman et al., 1988). When intramelanosomal cysteine is depleted, dopaquinone spontaneously undergoes cyclization to form leukodopachrome (cyclodopa). The redox reaction between leukodopachrome and unchanged dopaquinone subsequently gives rise to orange dopachrome. Dopachrome spontaneously decarboxylates into 5, 6-dihydroxyindole (DHI), which rapidly oxidizes and polymerizes to form insoluble, brownish-black eumelanin. However, if dopachrome tautomerase (TYRP2, DCT) is available, dopachrome will tautamerize without losing its carboxylic acid to form DHI-2-carboxylic acid (DHICA). Tyrosinase-related protein 1 (TYRP1, GP75) catalyzes the conversion of DHICA to moderately soluble, lighter brown eumelanin (Ito, 2003).

Figure 1. Synthetic pathways of pheomelanin and eumelanin.

Oxidation of L-tyrosine, first to L-dihydroxyphenylalanine (L-DOPA) and then to dopaquinone, catalyzed by tyrosinase (TYR). In the presence of cysteine, dopaquinone reacts with cysteine to form 2, 5- or 5, 5-cysteinyldopa, which gives rise first to cysteinyldopa-quinones, then benzothiazine intermediates to produce soluble reddish-yellow pheomelanin. When cysteine is depleted, dopaquinone spontaneously undergoes cyclization to form leukodopachrome. Leukodopachrome then reacts with unchanged dopaquinone to give rise to orange dopachrome, which spontaneously decarboxylates into 5, 6-dihydroxyindole (DHI) to produce brownish-black eumelanin. However, if dopachrome tautomerase (TYRP2) is available, dopachrome will tautamerize to form DHI-2-carboxylic acid (DHICA). Tyrosinase-related protein 1 (TYRP1) catalyzes the conversion of DHICA to lighter brown eumelanin.

3. CENTRAL DETERMINANTS OF MELANOCYTIC PIGMENTATION

3.1. Melanocortin 1 Receptor (MC1R)

The MC1R signaling cascade is one of the major systems involved in the regulation of mammalian pigmentation, by influencing the type and quantities of synthesized melanin. The human MC1R protein consists of 317 amino acids and there are approximately 700–1000 protein molecules expressed per melanocyte (Donatien et al., 1992; Roberts, Newton, Beaumont, Helen Leonard, & Sturm, 2006). Even though it is mostly expressed in melanocytes, low levels of MC1R protein have also been reported in keratinocytes, fibroblasts, and immune cells (Bastonini, Kovacs, & Picardo, 2016). MC1R belongs to the melanocortin receptor subgroup, the members of which are the shortest polypeptides in the G protein-coupled seven membrane-spanning receptor (GPCR) superfamily. Upon stimulation by its endogenous agonists – α-MSH and adrenocorticotropic hormone (ACTH) – MC1R activates adenylyl cyclase, leading to an increase in the intracellular concentration of cyclic AMP (cAMP). The cAMP then activates protein kinase A (PKA), which in turn phosphorylates cAMP responsive-element-binding protein (CREB) transcription factor family members (Busca & Ballotti, 2000a; reviewed in Garcia-Borron, Abdel-Malek, & Jimenez-Cervantes, 2014). CREB transcriptionally activates a variety of downstream targets, including MITF that controls the expression of several pigmentation enzymes, such as TYR, TYRP1, TYRP2, and melanocytic differentiation markers (Lin & Fisher, 2007).

There are numerous MC1R polymorphisms and these are associated with variations in hair color. The red-hair variants of MC1R are thought to produce complete loss of function of the cAMP-inducing activity of the receptor (Beaumont et al., 2007; Beaumont, Shekar, Cook, Duffy, & Sturm, 2008; Ringholm et al., 2004). Variants of MC1R with progressively more potency for cAMP induction are associated with progressively darker hair color. This progressive darkening is caused by increased eumelanin incorporation into the hair matrix. Mechanistically, it is likely that weak activity of MC1R variants produce low levels of cAMP and correspondingly low levels of MITF. In turn, weak expression of TYR causes production of only low levels of dopaquinone, which are chemically reduced by either cysteine or glutathione, leading to synthesis of pheomelanin. In contrast, strongly signaling variants of MC1R produce higher cAMP levels, higher MITF expression, and stronger transcriptional expression of TYR, resulting in larger quantities of dopaquinone synthesis, which is more likely to deplete intracellular stores of the reducing thiols from cysteine or glutathione. After depletion of these intracellular thiols, dopaquinone metabolism occurs via an alternative pathway to produce dopachrome, which is subsequently converted to brown/black eumelanin pigments via the actions of the TYRP1 and TYRP2 enzymes (Figure 1).

The best studied endogenous antagonist of MC1R is the agouti signaling protein (ASIP), which serves as a biological switch from eumelanin to pheomelanin synthesis in melanocytes (Lu et al., 1994; reviewed in Voisey & Van Daal, 2002). ASIP inhibits eumelanin synthesis by directly binding to MC1R and functioning as a reverse agonist, blocking even ligand-independent activity of the receptor, thereby preventing activation of the signaling cascade, including MITF expression. The Asip gene has been known to be responsible for the pheomelanotic banding pattern observed in wild-type mice (Lu et al., 1994). Loss of function at the Asip gene leads to lack of the yellow banding of fur (giving rise to the “black” name in the C57BL/6 mouse strain). In contrast, mice with increased and ectopic expression of ASIP caused by the A^vy mutant exhibit yellow coat color (Lu et al., 1994; Voisey & Van Daal, 2002).

3.2. Microphthalmia-associated Transcription Factor (MITF)

MITF has been identified as a master transcriptional regulator that orchestrates key developmental and differentiation programs in the melanocyte lineage. Almost 70 years ago, the microphthalmia phenotype was first observed in mice having significantly reduced eye sizes and a depigmentation phenotype due to the mutant locus mi (later changed to Mitf) (Hertwig, 1942). Loss of pigmentation in these mice is caused by lack of melanocytes rather than a defective melanogenesis pathway within viable melanocytes, suggesting this gene’s role in early melanocytic developmental processes. Subsequently, multiple mutations occurring within the Mitf locus were discovered to confer additional phenotypes, including a decreased number of mast cells, defective osteoclasts, and early onset of deafness. Through transgenic insertional mutagenesis, the Arnheiter group determined that this locus encodes a novel transcription factor, MITF, that contains a basic helix-loop-helix leucine zipper (bHLH-ZIP) DNA binding motif (Hodgkinson et al., 1993). MITF together with transcription factor E3 (TFE3), transcription factor EB (TFEB), and transcription factor EC (TFEC) form the MiT transcription factor family, which binds to E-box motifs consisting of the consensus sequence CA[T/C]GTG and dictates tissue-specific gene expression of critical pigmentation enzymes (Hemesath et al., 1994). In humans, heterozygous mutation in MITF leads to autosomal dominant Waardenberg syndrome type IIA, characterized by sensorineural deafness and pigmentation abnormalities of hair and eyes (Hughes, Newton, Liu, & Read, 1994). COMMAD syndrome is associated with biallelic mutations in MITF that include at least one dominant-negative mutation. The clinical phenotypes are abnormal development of the eyes, albinism, hearing loss, osteopetrosis, microphthalmia, and macrocephaly (George et al., 2016).

Multiple MITF isoforms are known to be expressed from distinct promoters of the MITF gene, with different tissue expression profiles. Among those, the M-isoform of MITF (MITF-M) is expressed almost exclusively in melanocytes (Fuse, Yasumoto, Suzuki, Takahashi, & Shibahara, 1996), though it has also been detected in retinal pigment epithelium (Bharti, Liu, Csermely, Bertuzzi, & Arnheiter, 2008; Masuda & Esumi, 2010) and likely in mouse olfactory bulb projection neurons (mitral cells and tufted cells) (Ohba, Takeda, Yamamoto, & Shibahara, 2015). Besides CREB mentioned previously, multiple other transcription factors are involved in the regulation of MITF-M at the transcriptional level, including paired box gene 3 (PAX3), SRY (sex-determining region Y)-box 10 (SOX10), lymphoid enhancer-binding factor 1 (LEF1 or TCF), one cut domain 2 (ONECUT-2), and MITF itself (reviewed in (Kawakami & Fisher, 2017)). PAX3 functions synergistically with SOX10 to activate and induce MITF transcription (Bondurand N et al., 2000). However, interestingly, PAX3 can also compete with MITF to bind to an enhancer required for the expression of TYRP2, a critical enzyme in melanin synthesis pathway (Lang et al., 2005). SOX10, expressed almost exclusively in neural crest lineages, has been found to interact tightly with CREB to drive the melanocyte-specific expression of MITF-M (Huber et al., 2003). In addition, MITF-M is a direct downstream target of Wingless-type (Wnt) signaling, which is critical for melanocyte development, particularly in the neural crest. Upon the stimulation of this cascade, β-catenin is stabilized and translocated to the nucleus, where it functions as a transcriptional coactivator of LEF1 to activate the MITF-M promoter (Takeda, Yasumoto, et al., 2000; Widlund et al., 2002). ONECUT-2 has been shown to enhance MITF-M promoter activity in a cis-acting manner through genetic and biochemical studies (Jacquemin et al., 2001).

At the post-transcriptional level, MITF expression is modulated by multiple chemical modifications including phosphorylation, sumoylation, and ubiquitination (Hsiao & Fisher, 2014). MITF-M is phosphorylated by mitogen-activated protein kinase (MAPK), ribosomal S6 kinase (RSK), glycogen synthase kinase-3β (GSK3β), and p38, in response to different environmental stimuli (Hemesath, Price, Takemoto, Badalian, & Fisher, 1998; Mansky, Sankar, Han, & Ostrowski, 2002; Takeda, Takemoto, et al., 2000; Weilbaecher et al., 2001). Dual phosphorylation of MITF by c-Kit results in ubiquitination and proteasomal degradation of MITF (Wu et al., 2000). The protein inhibitor of activated STAT3 (PIAS3), which causes sumoylation of MITF represses MITF activity by inhibiting its DNA-binding ability (Levy, Nechushtan, & Razin, 2002; Miller, Levy, Davis, Razin, & Fisher, 2005). Previously, various studies have revealed other proteins that potentially repress MITF transcription, such as FOXD3, POU3F2, transforming growth factor β (TGF- β), and ALX3 through suggested mechanisms (reviewed in (Kawakami & Fisher, 2017)).

Advances in technologies have provided substantial insights into identification of downstream target genes of MITF and their roles in major cellular processes, particularly in melanocytic differentiation and growth/survival functions. MITF is the master regulator of three primary pigmentation enzymes required for melanin synthesis: TYR, TYRP1, and TYRP2 (Bentley, Eisen, & Goding, 1994; Bertolotto et al., 1998). Additionally, MITF controls the expression of PMEL17 (gp100) and MLANA (MART-1) factors (Du et al., 2003) that are essential for the formation of melanosomal matrix and maturation of melanosomes (Berson, Harper, Tenza, Raposo, & Marks, 2001; Hoashi et al., 2005; Raposo, Tenza, Murphy, Berson, & Marks, 2001). Other genes involved in melanosome biogenesis and trafficking have been revealed to be direct targets of MITF, including GPR143 and RAB27A (Hoek et al., 2008; reviewed in Cheli, Ohanna, Ballotti, & Bertolotto, 2010). Our group reported that phosphodiesterase 4D3 (PDE4D3) is a direct target of MITF that creates a negative feedback circuit to modulate cAMP signaling activity and melanocytic differentiation . In addition, MITF controls cell cycling and survival of melanocytes by regulating transcription of various target genes, including cyclin-dependent kinase 2 (CDK2), T-box transcription factor 2 (TBX2), CDK inhibitors p21^CIP1 and p16^INK4A, B-cell lymphoma 2 (BCL2), and BCL2A1 ( Hoek et al., 2008; reviewed in Cheli et al., 2010). More recently, direct regulation by MITF of genes required for DNA replication and genome maintenance has been shown (Strub et al., 2011). Additionally, MITF has been shown to regulate metabolism and oxidative stress through the master mitochondrial regulator peroxisome proliferator-activated receptor γ coactivator 1 α (PGC1α) and apurinic/apyrimidinic endonuclease I/redox factor-1 (APEX1/REF1) (Haq et al., 2013; Hsiao & Fisher, 2014; Liu, Fu, & Meyskens, 2009). Proteins that interact with MITF have been identified using tandem affinity purification and mass spectrometry, including proteins involved in DNA damage, repair, and replication, as well as components of the PBAF chromatin remodeling complex (Laurette et al., 2015).

4. UV RADIATION EXPOSURE AND PIGMENTATION

4.1. Characteristics of UV Radiation and Mutagenesis

UVR, spanning a region between visible rays and X-rays within the electromagnetic spectrum, is classified into three subgroups based on wavelength: UVA (400–320 nm), UVB (320–280 nm), and UVC (280–200 nm). Even though the sun emits a significant amount of UVR, only 5% of the radiation actually reaches the earth’s surface, of which approximately 95% is UVA and 1–10% is UVB. Almost 100% of UVC is filtered out by the atmosphere and the ozone layer (Valejo Coelho, Matos, & Apetato, 2016). Therefore, research has paid most attention to the biological consequences of UVA and UVB exposure to the skin. The depth of penetration of UVR primarily depends on the wavelength: UVB reaches the basal layer of the epidermis, while UVA can be absorbed deeper, into the dermis. Macromolecules including proteins, lipids, and nucleic acids are direct targets of UVR. Absorption of photons from the radiation by chromophores results in excited electron states and potentially deleterious reactions, causing direct and indirect DNA damage (Tran, Schulman, & Fisher, 2008).

UV-induced mutagenesis has been found to produce signature mutations in genes participating in key cellular processes such as proliferation, differentiation, and survival. With shorter wavelengths, UVB can cause direct DNA damage by photochemically inducing the formation of DNA photoproducts, which are dimers linked by a covalent bond between two adjacent pyrimidines in the same polynucleotide chain. Cyclobutane pyrimidine dimers (CPDs) and pyrimidine 6–4 pyrimidones are two major DNA photoproducts induced by UVB (Cadet, Grand, & Douki, 2015). Upon production, these bulky damaged DNA lesions can be repaired by the nucleotide excision repair (NER) system. Inadequate repair can result in signature mutations for UVB mutagenesis, particularly C$\to$T transitions in the setting of an adjacent pyrimidine (about 70%) and CC$\to$TT tandem mutations (about 10%). On the other hand, UVA indirectly causes cellular DNA damage through two mechanisms. In type I photosensitized reactions, energy transfer from UVA-excited chromophores to DNA triggers formation of CPDs. In type II reactions, with the presence of singlet oxygen (¹O₂), UVA can lead to oxidation of guanine bases, generating 7,8-dihydro-8-oxoguanine (8-oxoG) lesions (Valejo Coelho et al., 2016). These lesions are repaired by a different pathway called base excision repair (BER), which is initiated by the enzyme 8-oxoguanine DNA glycosylase-1 (OGG1). Common UVA-induced mutations due to inefficient repair include G$\to$T transversions and G$\to$A transitions (Chen et al., 2014).

Typically, production of CPDs is completed within picoseconds from direct UV absorption (Brash, 2016). However, the Brash group discovered that after UVA or UVB exposure, melanin-containing murine melanocytes continued to generate CPDs for at least 3 hours, while CPD induction lasted only picoseconds in fibroblasts and albino melanocytes (Premi et al., 2015). CPDs from delayed production in pigmented melanocytes were called “dark CPDs” and most prominent in skin containing pheomelanin. Melanin was demonstrated to be an active participant in the formation of these dark CPDs as well as delayed DNA damage arising from CPDs upon UV exposure. Mechanistically, when UV-induced reactive oxygen and nitrogen species combine, they will excite an electron in fragments of the pigment melanin, creating a quantum triplet state to induce CPDs by radiationless energy transfer to DNA (Premi et al., 2015).

4.2. UV-dependent Tanning Pathways

The tanning response is probably the most striking photo-protective mechanism against detrimental effects of UVR exposure to the skin. In keratinocytes, DNA damage due to UVR stabilizes the tumor suppressor protein p53, the central regulator of cell cycle and DNA damage repair. Our group previously demonstrated that loss of p53 in mice interrupted UV-induced tanning ability (Cui et al., 2007). Stabilized p53 subsequently promotes the transcriptional activation of the pro-opiomelanocortin (POMC) gene. The POMC pro-peptide is known to be enzymatically processed by proconvertase 1 to generate ACTH, which is cleaved by proconvertase 2 to form corticotropin-like intermediate peptide (CLIP) and α-MSH. In addition, the opioid β-endorphin is a highly potent derivative of POMC involved in behavioral and mood-related effects (reviewed in Bicknell, 2008). Mutations in the POMC gene in humans can result in skin tanning inability, adrenal insufficiency, and early-onset obesity (Krude et al., 1998). Stimulation of the MC1R signaling pathway by α-MSH ligand activates the cAMP-CREB-MITF cascade in melanocytes, resulting in the synthesis of melanin and eventual transfer of melanosomes to keratinocytes for protection against UV (Cui et al., 2007; D’Orazio et al., 2006) (Figure 2). In addition, the stress-responsive p38 kinase pathway has been suggested to be involved in contributing to the tanning response by phosphorylating the ubiquitous bHLH-ZIP transcription factor USF1 to induce TYR expression (Galibert, Carreira, & Goding, 2001). In Usf-1^−/− mice, melanocytes were found to be unable to activate Pomc and Mc1r gene expression upon UVR treatment (Corre et al., 2004). α-MSH also acts synergistically with UVR to activate both p38 kinase and p53 in melanocytes expressing wild-type MC1R (Kadekaro et al., 2012; Newton, Roberts, Leonard, & Sturm, 2007). Furthermore, there are other pathways involving endothelin-1, β-fibroblast growth factor, nitric oxide, p-locus, and stem cell factor that potentially contribute to UV-mediated tanning responses (reviewed in (Lin & Fisher, 2007). Human red-haired/light-skinned individuals harbor non-functional polymorphic variants of MC1R and are known to display an inability to tan in response to UV, a phenotype that is corroborated in mice harboring a loss of function mutation in Mc1r (D’Orazio et al., 2006). Based on this phenotype, along with the ability of topical cAMP agonists to rescue cutaneous pigmentation, it appears likely that the POMC-MSH-MC1R-cAMP-MITF pathway plays a rate limiting role in the UV-tanning response.

Figure 2. UV-induced cutaneous tanning response.

DNA damage induced by UVR stabilizes p53, which subsequently promotes transcriptional activation of the pro-opiomelanocortin (POMC) gene. POMC is cleaved to yield multiple peptides, including α-MSH and the opioid β-endorphin. Stimulation of the MC1R signaling pathway by α-MSH ligand activates the cAMP-CREB-MITF cascade in melanocytes, resulting in the synthesis of melanin and eventual transfer of melanosomes to keratinocytes for protection against UV. β-endorphin has been shown to be involved in mood-related effects and analgesia dependency. In addition, salt inducible kinases (SIKs) phosphorylates the cAMP-regulated transcriptional co-activator (CRTC) family of proteins, thereby preventing translocation of CRTC to the nucleus and eventually leading to inhibition of transcriptional activation of MITF. UV absorption by 7-dyhrocholesterol (7-DHC) forms vitamin D₃, an inactive form of vitamin D in keratinocytes.

4.3. Photo-protective Roles of Melanin Against UV

Melanin, synthesized upon activation of the tanning pathway, possesses shielding effects against UVR exposure by scattering or absorbing the radiation to prevent its penetration through the epidermis (Kaidbey, Agin, Sayre, & Kligman, 1979). Of note, melanin is thought to absorb up to 50% to 75% UVR contacting the skin in certain settings, and transform the energy into heat through internal conversion (Garibyan & Fisher, 2010). Of the two types of melanin, eumelanin has substantially more photo-protective properties compared with pheomelanin. Melanosomes in darker skin, having high eumelanin content, have been shown to be intact within the epidermis due to their resistance to lysosomal degradation (Brenner & Hearing, 2008). In addition, eumelanin functions as a free radical scavenger with superoxide dismutase-like activity that reduces reactive oxygen species (ROS), while pheomelanin can generate ROS through UV-dependent and UV-independent pathways (Mitra et al., 2012; Napolitano, Panzella, Monfrecola, & d’Ischia, 2014; Bustamante, Bredeston, Malanga, & Mordoh, 1993). This anti-oxidant activity of eumelanin may represent one of its most important protective functions. Typically, stimulation of the UV-dependent α-MSH-MC1R pathway results in skin darkening by eumelanin synthesis through increased activity of TYR, TYRP1, and TYRP2 enzymes. It has been suggested that high tyrosinase activity and dopaquinone synthesis due to strong MC1R signaling depletes the supply of cysteine and glutathione, which are necessary for the generation of pheomelanin, leading to the occurrence of eumelanogenesis as the default pathway. In mice, recessive mutations in the extension locus of the Mc1r gene (Mc1r^e/e) lead to yellow or pheomelanic hair, similar to the red-hair variants of MC1R in humans, whereas wild-type C57BL/6 mice have black or eumelanic hair (Chen et al., 2014).

5. UV RADIATION AND THE RISK FOR MELANOMA

Cutaneous melanoma is one of the most aggressive and treatment-resistant malignancies among all human cancer types (Tsao, Chin, Garraway, & Fisher, 2012). Even though melanoma only accounts for about 1% of skin cancers, it leads to a majority of skin cancer deaths. According to the American Cancer Society, there will be approximately 87,000 new patients diagnosed with melanoma and almost 10,000 predicted deaths within the United States in 2017 (Siegel, Miller, & Jemal, 2017). Cumulative exposure to UVR, either from direct sunlight or indoor tanning beds, is considered the major environmental risk factor of melanoma. Historically, there was a positive correlation between childhood sunburn and subsequent risk of cutaneous melanoma (Oliveria, Saraiya, Geller, Heneghan, & Jorgensen, 2006; Whiteman, Whiteman, & Green, 2001). However, through a comprehensive meta-analysis pooling data from 51 independent study populations, the risk was suggested to correlate with increasing number of sunburns for all life-periods (childhood, adolescence, adulthood, and lifetime) (Dennis et al., 2008). In addition to sunlight exposure, there is growing evidence from case-control studies supporting an increase in risk of melanoma from the use of indoor tanning beds. In a study with a large cohort of 141,045 women, ever users of indoor tanning devices were more likely to develop melanoma than never users (adjusted relative risk (RR) = 1.24, 95% confidence interval (CI): 1.05, 1.46). Increased melanoma risk was also significantly associated with longer duration of use and cumulative number of tanning sessions (Ghiasvand et al., 2017).

Although a comprehensive understanding of the mechanisms through which UV induces melanoma has yet to be fully defined, there is significant evidence demonstrating how UVR contributes to the initiation and progression of cutaneous melanoma. There is some controversy over the question of whether UVR induces the common valine-to-glutamic acid substitution at codon 600 (V600E) of BRAF within the MAPK pathway, which is the most commonly mutated oncogene, identified in 40–60% of melanoma patients (Davies et al., 2002). However, this mutant allele does not likely arise from classical UVB-associated, pyrimidine dimer-mediated, directly targeted mutagenesis, though perhaps it may occur via an oxidative damage mechanism (Venza et al., 2015). Furthermore, UV-independent BRAF^V600E mutation alone is often unable to initiate the malignant transformation of melanocytes (Michaloglou et al., 2005). Over 90% of BRAF^V600E mutations occur in sun-exposed skin, and these lesions are primarily benign melanocytic nevi. Considerable information has been obtained regarding the acquisition of additional mutations within cutaneous melanomas arising from adjacent benign nevi, from studies of Bastian and colleagues (Shain et al., 2015). A very high fraction of the additional mutations are UV-induced and likely to synergize with BRAF^V600E to drive the development of invasive melanoma. A minority of melanoma patients harbor UV-signature mutations of TP53. UV-induced DNA damage signatures have been found in approximately 46% of known driver mutations in human BRAF^V600E melanomas, including CDKN2A, TP53, NF1, RAC1, and PTEN (Michaloglou et al., 2005; Hodis et al., 2012; Xia et al., 2014). Separate from its mutagenic activity, UV acts as an immunosuppressant by disturbing skin antigen presentation and recruiting suppressive immunological cells through cytokines and other mediators, contributing to tumor growth and survival (reviewed in (Valejo Coelho et al., 2016). Recently, the White group demonstrated UVB as a critical extrinsic factor to promote the initiation of melanoma through melanocyte stem cells in an inflammation-dependent manner (Moon et al., 2017).

6. UV AND BEHAVIORAL EFFECTS

Since UVR exposure is considered a predominant risk factor for all common cutaneous malignancies, minimizing tan-seeking behavior is strongly recommended as a prevention approach. Nonetheless, there is still an alarming increase in skin cancer incidence annually. Several studies have been performed to determine whether UV-seeking behavior (sunbathing or indoor tanning) could become addictive upon frequent and excessive exposure, based on the CAGE (Cut down, Annoyed, Guilty, Eye-opener) questionnaire for alcoholism screening and the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) (DSM-IV) criteria for substance-related disorders. From a cohort of 145 beachgoers, Warthan et al. revealed that 53% of subjects met proposed tanning addiction criteria and 26% had problematic tanning behavior (Warthan, Uchida, & Wagner, 2005). A modified study of 100 indoor tanners also showed a similar trend with a total of 74% of subjects exhibiting aberrant tanning behavior (Harrington et al., 2011). Recently, Mays et al. found 22.6% among a community sample of 389 non-Hispanic white young women screened positive for tanning dependency, which was suggested to associate with younger age at indoor tanning initiation, usage of more than 20 times, stronger beliefs in benefits of indoor tanning, and depression (Mays, Atkins, Ahn, & Tercyak, 2017).

There are proposed mechanisms underlying physiological reinforcing properties of UV to drive tanning addiction, from both preclinical and clinical studies. In 2014, our group demonstrated in mice that β-endorphin, produced upon UV exposure from cleavage of the POMC propeptide, induced an opioid receptor-mediated addiction-like response to UV light (Fell, Robinson, Mao, Woolf, & Fisher, 2014). When mice were chronically treated with low doses of UVR, we observed elevating plasma levels of β-endorphin in parallel with increased analgesic thresholds. These responses were reversed by an opioid antagonist, naloxone, and absent in β-endorphin knockout mice. Mice treated with naloxone following chronic UVR were found to exhibit typical murine opioid withdrawal symptoms (wet dog shake, paw tremor, teeth chatter, and rearing). Chronically UV-irradiated mice also displayed opioid dependency-like responses in a conditioned place aversion assay that tested them for avoidance of a naloxone-paired environment. The other support for opiate-like effects from UV induction of β-endorphin included significantly higher doses of morphine being required in chronically UV-irradiated mice relative to non-UV controls to achieve comparable thermal analgesia. Furthermore, expression of p53 in keratinocytes was demonstrated to be essential for increased circulating β-endorphin levels and the observed behavioral responses upon chronic UV exposure.

Recently, a study from the Adinoff group investigated the striatal dopaminergic efflux in addicted indoor tanners to understand the effect of UV on the central nervous system (CNS) (Aubert et al., 2016). The mesostriatal dopamine pathway has been known to associate with reward (providing enjoyment and arousal) and addiction (characterized by persistent, compulsive, and controlled behaviors) (Adinoff, 2004). Using single photon emission computerized tomography (SPECT) and ¹²³I-iodobenzamide, the group showed that exposure to UV triggered increased striatal dopaminergic efflux in addicted, but not in infrequent indoor tanning users. In addition, the level of UV-induced dopamine release was demonstrated to correlate with severity of tanning dependence. As dopamine efflux is one of the downstream responses upon elevating β-endorphin plasma levels, this study has posited a potential connection between the cutaneous tanning pathway and the CNS in driving UV-seeking addictive behavior.

Despite its detrimental effects, UVR plays a critical role in the synthesis of vitamin D. UVB absorption by 7-dehydrocholesterol in the skin generates the previtamin D molecule, which exists in two conformations. The less thermodynamically favored cis form of previtamin D is then isomerized to form vitamin D₃ (Figure 2) (Holick, 2008). Vitamin D₃ is biologically inert, but is hydroxylated in the liver by the enzyme D-25-hydroxylase (25-OHase) to become 25(OH)D. A second hydroxylation reaction catalyzed by 25(OH)D-1-OHase (CYP27B1) in the kidneys is required for conversion of 25(OH)D to its biologically active form, 1,25(OH)2D. The vitamin D receptor is expressed almost ubiquitously in the body. 1,25(OH)2D plays a vital role in various functions, particularly in maintaining calcium and phosphate homeostasis as well as bone metabolism. Vitamin D deficiency not only leads to severe defects in bone mineralization, but has also been suggested to contribute to risks for cancer, heart diseases, type 2 diabetes, and autoimmune conditions (reviewed in (Nair & Maseeh, 2012). Considering the suggested wide-ranging health benefits of vitamin D, there are potentially natural mechanisms reinforcing UV-seeking behavior that may have arisen during certain stages of evolution to support the synthesis of vitamin D. Unfortunately, the indoor tanning industry has emphasized UVR’s induction of vitamin D synthesis without clarifying its known carcinogenic risks. In fact, rather than excessive exposure to carcinogenic UVR, a much safer approach to maintaining healthy levels of vitamin D is consumption of inexpensive oral supplements, which do not carry carcinogenic risk.

7. UV-INDEPENDENT SKIN DARKENING AND POTENTIAL APPLICATIONS

Epidemiological studies have revealed that the MC1R gene is highly polymorphic with over 200 coding allelic variants, significantly contributing to pigmentation diversity among different ethnic populations (reviewed in Makova & Norton, 2005; Dessinioti, Antoniou, Katsambas, & Stratigos, 2011). Certain single amino-substitutions in these variants result in the red hair color phenotype characterized by fair skin, red hair, propensity to freckle, inability to tan, and higher risk of sunburn (Lin & Fisher, 2007; Valverde, Healy, Jackson, Rees, & Thody, 1995). The red hair color phenotype is associated with an increased incidence of melanoma through both UV-dependent and UV-independent pathways (Mitra et al., 2012; Williams, Olsen, Hayward, & Whiteman, 2011). In fact, the poor tanning responses of those individuals promote vulnerability to melanoma development due to insufficient photo-protective eumelanogenesis through α-MSH/cAMP (Rouzaud, Kadekaro, Abdel-Malek, & Hearing, 2005). Hence, induction of skin darkening without UV exposure may plausibly offer an effective melanoma prevention strategy, especially for individuals having MC1R non-signaling variants.

Previously, our group demonstrated that topical application of forskolin, a cAMP agonist, could rescue eumelanization in Mc1r-deficient mice and confer significant protection against carcinogenic UV exposure (D’Orazio et al., 2006). From identifying PDE4D3 as a negative regulator of cAMP homeostasis, we also reported that inhibition of PDE4D3 resulted in a skin darkening phenotype in red-haired mice (Khaled, Levy, & Fisher, 2010). However, topical application of these small molecules was limited by their poor penetration through human skin, which is approximately 5 times as thick as murine skin. We recently showed that topical application of small molecule inhibitors of the salt inducible kinase (SIK) family to both human and mouse skin induced eumelanin synthesis in the absence of UV exposure, by inducing MITF and pigment gene expression (Mujahid et al., 2017). Deletion of Sik2 in yellow-haired, light-skinned Agouti mice was previously shown to promote eumelanization (Horike et al., 2010). SIK kinases were shown to phosphorylate the cAMP-regulated transcriptional co-activator (CRTC) family of proteins, thereby preventing translocation of CRTC to the nucleus, where unphosphorylated CRTC normally co-activates CREB, an inducer of MITF in melanocytes (Altarejos & Montminy, 2011; Clark et al., 2012). In a mouse model of red hair/fair skin (Mc1r^e/e), with epidermal melanocytes from transgenic expression of stem cell factor in keratinocytes, we showed that multiple small molecule pan-SIK inhibitors induced dose-dependent upregulation of MITF and MITF target genes, as well as profound reversible eumelanization (Figure 2) (Mujahid et al., 2017).

Since robust epidermal pigmentation is one of the strongest epidemiologic factors associated with prevention of UV-associated skin cancers and photoaging, topical application of SIK inhibitors could have widespread use as a substitute for UV tanning. Such an approach could be especially beneficial for people at a high risk of skin cancer, such as patients with a previous occurrence of skin cancer, immunosuppressed organ transplant patients, and individuals with DNA repair defects or certain forms of albinism. In the United States, there were an estimated 9.7 million non-melanoma skin cancer survivors in 2009 and over 1 million melanoma survivors in 2014. A systematic review and meta-analysis examining the risks of second primary skin cancers after cutaneous squamous or basal cell carcinoma found 17-, 3.2, and 2.4-fold increased risks of basal cell carcinoma, squamous cell carcinoma, and melanoma, respectively, after basal cell carcinoma, and 4.2-, 15-, and 2.7-fold respective increased risks after squamous cell carcinoma (Flohil, van der Leest, Arends, de Vries, & Nijsten, 2013). An analysis of the National Cancer Institute Surveillance, Epidemiology, and End Result (SEER) database determined that melanoma survivors developed second primary cutaneous melanomas at 10 times the rate that the general SEER population developed first primary melanomas (Goggins & Tsao, 2003).

Approximately 126,670 solid organ transplants were performed worldwide in 2015, of which two thirds were kidney transplants (Data from the WHO-ONT Global Observatory on Donation and Transplantation). An increased risk of skin cancers in organ transplantation recipients due to immunosuppressive treatments has been well documented, including 65–250 fold and 10–16 fold higher frequencies of cutaneous squamous cell carcinoma and basal cell carcinoma, respectively, which account for 90% of all skin cancers in organ transplant recipients (Euvrard, Kanitakis, & Claudy, 2003; Lindelöf, Sigurgeirsson, Gäbel, & Stern, 2000). A slightly increased risk of melanoma has also been reported in these patients, as compared to the general population (Brewer et al., 2011; Kempf, Mertz, Hofbauer, & Tinguely, 2013). Presumably, immunosuppressive treatments inhibit immune system scavenging of neoantigen-containing skin cells created by UV-induced mutations. The frequency of such mutations might be reduced by sunless eumelanization induced by SIK inhibitors, together with traditional UV-shielding sunscreens. Although the frequency of squamous cell carcinoma among Norwegian organ transplant patients decreased 2–3 fold from 1998–2012 compared with the 1980’s, likely due to less damaging immunosuppressive drugs, the risk of skin cancer in transplant recipients remains substantial (Rizvi et al., 2017).

Werner syndrome is an autosomal recessive disorder characterized by deficient repair of double stranded DNA breaks and premature aging (Yannone et al., 2001). The prevalence of Werner syndrome has been estimated at 1 per 380,000–1,000,000 worldwide and 1 per 20,000–40,000 in Japan. Among residents of Japan, Werner syndrome was found to confer a 53-fold greater risk of melanoma, although most of the tumors were acral lentiginous and mucosal melanomas (Lauper, Krause, Vaughan, & Monnat, 2013). Xeroderma pigmentosum (XP) is an autosomal recessive disorder characterized by a deficiency in nucleotide excision repair, a process responsible for repair of UV-induced DNA damage, The frequency of XP has been estimated to be approximately 1 per million and 1 per 40,000–100,000 births in Western countries and Japan, respectively (Moriwaki & Kraemer, 2001). Studies published in 1984 and 2011 reported that XP patients had 10,000-fold and 4,800-fold higher rates, respectively, of non-melanoma skin cancers before age 20 compared with non-XP populations, and both studies reported 2,000-fold higher rates of melanoma in the same cohorts (Bradford et al., 2011; Kraemer, Lee, & Scotto, 1984).

Of the four types of oculocutaneous albinism, OCA2, due to a defective OCA2 melanosomal transmembrane protein, is the most prevalent at 1 per 36,000 among white Europeans and 1 per 3,900–10,000 among Africans. OCA1, resulting from mutations that abolish tyrosinase activity completely (OCA1A) or partially (OCA1B) is found in 1 per 40,000 individuals worldwide. OCA3 is caused by defective TYRP1 and is found in 1 per 8,500 Africans, but is very rare in Caucasian and Asiatic populations. OCA4, due to defects in membrane-associated transporter protein, is found in 1 per 85,000 Japanese, but is very rare in Caucasians (Grønskov, Ek, & Brondum-Nielsen, 2007). Among African albinos, the risk of developing squamous cell carcinoma is up to 1000-fold higher than in the general population (Lekalakala et al., 2015). Melanoma among albinos is rare (Kiprono, Chaula, & Beltraminelli, 2014; Mabula et al., 2012; Perry & Silverberg, 2001), though a slightly increased risk has been reported among individuals carrying the temperature-sensitive TYR R402Q variant (Gudbjartsson et al., 2008). It has been suggested that the risk of skin cancer may be higher among OCA1B, 2, 3, and 4 individuals, in which a low level of reactive oxygen species-promoting pheomelanin is maintained (de Vijlder, de Vijlder, & Neumann, 2013), and for which sunless tanning might yield varying degrees of protection, compared with OCA1A, consistent with murine data on pheomelanin-associated carcinogenesis (Mitra et al., 2012).

Sunless tanning with SIK inhibitors might also provide benefits to conditions such as photosensitive disorders and vitiligo. The prevalence of polymorphous light eruption (PLE), the most common photodermatosis, has been estimated at 10–20% among Northern U.S. and European populations (Hönigsmann, 2008). In addition, a growing list of drugs have been shown to be photosensitizers, including widely used nonsteroidal anti-inflammatory drugs (Monteiro-Steagall, Steagall, & Lascelles, 2013). While a trend toward lower prevalence of skin cancer was found in PLE patients, UV protection provided by sunless tanning might foster skin health maintenance in individuals previously afflicted with PLE or being treated with photosensitizing drugs.

The prevalence of vitiligo worldwide has been estimated at 0.5 to 2.0% (Krüger & Schallreuter, 2012). Patients with vitiligo have decreased susceptibility to development of melanoma, due to autoimmune destruction of melanocytes, and also have not shown significant increases in nonmelanoma skin cancers compared to the general population, presumably owing to general aspects of the autoimmune response against melanocytes (Rodrigues et al., 2017). SIK inhibitors might restore pigmentation to areas of vitiligo that retain a small number of melanocytes.

Our findings that SIK inhibitor-induced MITF expression in human melanocytes and pigmentation in mouse skin are reversible suggest that it is a potentially safe approach to induce tanning without UV exposure. It is not expected that transient topical application of SIK inhibitors would lead to development of hyperpigmentation conditions such as melasma (Lee, 2015), or induction of potentially oncogenic genomic mutation or amplification of the MITF gene (Garraway et al., 2005). However, human use of small molecule SIK inhibitors will require careful consideration dosage regimens and other safety factors, including the potential for adverse effects from excessive or prolonged application.

8. CONCLUDING REMARKS

UVR exerts various physiological consequences on the skin, from genetic alterations, pigmentation, and carcinogenesis, to vitamin D synthesis. The tanning response is a powerful photo-protective mechanism against damaging effects upon UVR exposure. Despite growing evidence supporting that UVR is a well-established carcinogen and a major extrinsic factor contributing to the initiation and progression of cutaneous malignancies, the incidence of skin cancer has been rising rapidly, especially among young adults. Furthermore, there are both preclinical and clinical studies suggesting that UV possesses addictive potential to drive tanning dependency. The indoor tanning industry has previously convinced some users that cosmetic and health benefits of tanning might actually outweigh its significant skin cancer risks. In fact, since UVR is biologically dangerous and addictive, it is tremendously important to educate the general public about the deleterious consequences of chronic UV-seeking behavior.

Recently, we demonstrated that inhibition of the SIK family independently activates the tanning pathway without exhibiting DNA-damaging effects of UV. Topical application of these small molecule inhibitors might serve as a novel approach in place of UV tanning to induce dark pigmentation, especially for individuals having higher risks of skin cancer, and possibly for vitiligo.