NRF2 in the Epidermal Pigmentary System

Melanogenesis is a major part of the environmental responses and tissue development of the integumentary system. The balance between reduction and oxidation (redox) governs pigmentary responses, for which coordination among epidermal resident cells is indispensable. Here, we review the current understanding of melanocyte biology with a particular focus on the “master regulator” of oxidative stress responses (i.e., the Kelch-like erythroid cell-derived protein with cap‘n’collar homology-associated protein 1-nuclear factor erythroid-2-related factor 2 system) and the autoimmune pigment disorder vitiligo. Our investigation revealed that the former is essential in pigmentogenesis, whereas the latter results from unbalanced redox homeostasis and/or defective intercellular communication in the interfollicular epidermis (IFE). Finally, we propose a model in which keratinocytes provide a “niche” for differentiated melanocytes and may “imprint” IFE pigmentation.


Introduction
Oxidative stress results from exposure to reactive oxygen intermediates, such as superoxide anions (O 2 − ), hydrogen peroxide (H 2 O 2 ), and hydroxyl radicals (HO), and can damage proteins, nucleic acids, and cell membranes, leading to mutagenesis or cell death [1]. Reactive oxygen species (ROS) are produced by mitochondria and peroxisomes during cellular physiological metabolism. ROS-induced cumulative damage eventually causes numerous diseases. The skin serves as the interface between an organism and its external environment, protecting against xenobiotics or ultraviolet (UV) radiation, which can cause ROS-mediated tissue damage. Melanocytes produce pigmentation in hair and the interfollicular epidermis (IFE), thereby contributing to photoprotection and thermoregulation [2]. Vitiligo is an autoimmune-driven chronic depigmenting disease in which oxidative stress plays important roles [3,4]. The Kelch-like erythroid cell-derived protein with cap'n'collar homology-associated protein 1 (KEAP1)-nuclear factor erythroid-2-related factor 2 (NRF2) system is a major antioxidative apparatus whose activation largely depends on the oxidation of cysteine residues on the actin-anchored KEAP1 protein [5]. NRF2 not only affects melanocyte proliferation/differentiation [6] but also helps melanocytes survive the cytotoxic immune responses in vitiligo [7]. Thus, because reduction and oxidation (redox) governs myriad biological processes, NRF2 appears to have broad and significant roles in melanocyte biology. Here, we sought to summarize the significant role of the KEAP1-NRF2 system, whose activation is tightly controlled by the redox status of thiol and disulfide.

The KEAP1-NRF2 System as a Master Regulator of Redox Homeostasis
Redox refers to the give-and-take of electrons between molecules (and/or their moieties). Oxidants deprive the target of electrons and are thus called electrophiles. Reductants provide the target with electrons and are therefore called electron donors. Therefore,

Melanocyte Biology
Melanocytes originate from the neural crest and localize in the hair follicles (HFs) and IFE, endowing the body surface with pigmentation [2]. Human melanocytes exist in the IFE and HFs, whereas murine melanocytes exclusively reside in pelage HFs (exceptions are the ears, nose, paws, tail, etc.). Approximately 1200 melanocytes reside per square millimeter of human skin, regardless of race [10]. Epidermal melanocytes exist in the basal IFE layer and form the epidermal melanin unit, in which one melanocyte communicates with 30-40 IFE keratinocytes (KCs) [10]. Melanocytes adhere to KCs via adhesion molecules such as E-cadherin (Ecad) and desmoglein 1 (DSG1) [11]. There are two distinct melanocytic populations in HFs: melanocyte stem cells (McSCs) and their differentiated progeny. McSCs reside in the hair bulge and secondary hair germ (the lower permanent portion of the HFs) and show cyclic activity in parallel with HF stem cells [12].

Redox "Switch" in Skin Pigmentation
In general, the pheomelanin synthetic pathway requires cysteine/GSH, whereas the eumelanin pathway is expedited in the absence of thiols (high cystine/cysteine or GSSG/GSH ratios) [27,28]. Compared with pheomelanin, eumelanin is more polymerized and structurally stable, and is better at protecting against ROS. Thus, darkly pigmented eumelanin serves as a superior antioxidant/photoprotector on the skin surface [10]. Melanogenesis consumes high energy, produces O 2− or H2O2 from mitochondria [24,25], and generates a pro-oxidant state, potentially sensitizing the epidermis to oxidative stress [22] ( Figure 1). An unbalanced extracellular redox milieu can also perturb cellular fate decisions, leading to cell death (apoptosis) [29] or proliferation (melanomagenesis) [30]. A

Redox "Switch" in Skin Pigmentation
In general, the pheomelanin synthetic pathway requires cysteine/GSH, whereas the eumelanin pathway is expedited in the absence of thiols (high cystine/cysteine or GSSG/GSH ratios) [27,28]. Compared with pheomelanin, eumelanin is more polymerized and structurally stable, and is better at protecting against ROS. Thus, darkly pigmented eumelanin serves as a superior antioxidant/photoprotector on the skin surface [10]. Melanogenesis consumes high energy, produces O 2− or H 2 O 2 from mitochondria [24,25], and generates a pro-oxidant state, potentially sensitizing the epidermis to oxidative stress [22] (Figure 1). An unbalanced extracellular redox milieu can also perturb cellular fate decisions, leading to cell death (apoptosis) [29] or proliferation (melanomagenesis) [30]. A rather perplexing fact is that pheomelanin has a potent photosensitizing capacity [31]. Pheomelanin toxicity likely causes premature aging and cutaneous tumorigenesis in fair-skinned individuals. In contrast, eumelanin can enhance permeability barrier function by lowering the surface pH of darkly pigmented human [32]/mouse [33] skin. A proper antioxidative response induced by changes in the intra-/extracellular redox milieu appears to give an important direction regarding the trajectory of the pigmentary pathways. Defective transmembrane cystine transport caused by the subtle gray (sut) mutation in the Slc7a11 gene reduces intracellular GSH levels and increases GSSG levels (thus, the GSH/GSSG ratio decreases). The failure in response against extracellular oxidative signals skews melanogenesis toward the eumelanin synthesis pathway [34]. The mitochondrial redox-regulating enzyme nicotinamide nucleotide transhydrogenase (NNT) controls melanosomal maturation (and eumelanogenesis) [35]. This inner mitochondrial membrane protein regulates NAD(P) + /NAD(P)H homeostasis by mediating electron transfer [36]. Although NNT can bring about an intracellular oxidative milieu in certain circumstances [37], NNT increases GSH/GSSG ratios and negatively regulates eumelanogenesis [35], similar to what has been observed for NRF2 [6]/Slc7a11 (sut mice) [8]. However, the most striking facts are that single nucleotide polymorphisms (SNPs) in the NNT gene clearly differentiate skin pigmentary phenotypes, and inhibition takes place at the post-transcriptional levels [35]. Existing evidence argues against the classic pigmentation pathway dependent on the UV-cyclic adenosine monophosphate-microphthalmia-associated transcription factor (MITF) axis, which in turn transactivates TYRP1 and TYRP2 [35]. We should note that sut melanocytes, which mount suboptimal counter-responses against an extracellular oxidative milieu [8], also exhibit abnormal proliferation and differentiation in vitro [34]. Thus, redox milieus (intracellular or extracellular) can profoundly affect melanocyte biological behaviors and fate decisions [38]. In summary, investigating the redox "switch" in skin pigmentary pathways not only could lead to a profound understanding of skin pigmentation biology but also may pave a way toward repurposing small molecule inhibitors for diverse pigmentary disorders [35].

Vitiligo
Vitiligo is an acquired chronic pigmentary disorder that affects 0.5% to 2% of the world's population without a clear preference for race or sex [39]. The US population-based prevalence estimate of vitiligo in adults was between 0.76% and 1.11% [40]. Vitiligo results from selective melanocyte loss, which leads to pigment dilution in the affected skin and mucosa. Typical vitiligo lesions present as milky-white nonscaly macules with distinct margins. Generally, vitiligo is clinically diagnosed, and no laboratory tests or biopsies are required. Two major forms of the disease are well-recognized according to the distribution of lesions: segmental vitiligo (SV) and non-segmental vitiligo (NSV) [41]. NSV includes acrofacial, mucosal, generalized, universal, mixed, and rare variants. Distinguishing SV from other types of vitiligo is important because of its prognostic implications [42].
Vitiligo pathogenesis involves multiple factors, such as genetic background, metabolic abnormalities, oxidative stress, generation of inflammatory mediators, autoimmune responses, and decreased melanocyte adhesiveness [43]. These multiple mechanisms may function collectively, leading to melanocyte destruction [42].

Genetic Background of Vitiligo
Multiple studies have revealed the genetic background of vitiligo development. Approximately 50 different genetic loci that contribute to the risk of vitiligo have been discovered, principally in European-derived whites and Chinese [44]. A genome-wide association study identified the susceptibility loci for autoimmunity (e.g., HLA classes 1 and 2, PTPN22, IL2R α, GZMB, FOXP3, BACH2, CD80, and CCR6) and melanocyte-specific gene TYR in patients with vitiligo [45]. In addition, altered NALP1 (the gene encoding NACHT leucinerich repeat protein 1), a regulator of innate immunity, was found to be a risk factor for vitiligo [46]. Recently, polymorphic expression of MTHFR (the gene encoding methylene tetrahydrofolate reductase), which regulates homocysteine levels, has been identified in patients with vitiligo [47]. XBP1P1 (the gene encoding X-box binding protein 1) has also been associated with vitiligo. It is pivotal in attenuating the unfolded protein response and driving stress-induced inflammation in vivo [48].
Recently, a molecular mechanism involved in oxidative stress-induced melanocyte degeneration has been proposed. Oxeiptosis is an apoptosis-like, nonclassical, ROS-induced cell death pathway [73]. Because H 2 O 2 induces vitiligo melanocyte cell death, oxeiptosis may contribute to vitiligo pathogenesis [74]. The microRNA (miRNA) miR-25 suppresses MITF levels in melanocytes and SCF and bFGF expression in KCs, thereby contributing to melanocyte degeneration [75].

Immune Activation in Vitiligo
Autoimmunity has been implicated in vitiligo pathogenesis [76]. This is supported by the presence of antibodies against melanocytes, association with polymorphism at immune loci, prominent T cell infiltration in perilesional areas, cytokine expression, and association with other autoimmune diseases (e.g., autoimmune thyroiditis and type 1 diabetes mellitus) [43].
Evidence indicates that antioxidative response aberrations are central to vitiligo pathogenesis. SNPs in the NRF2 promoter may increase vitiligo risk [96], suggesting that aberrant antioxidative responses can be genetically determined. The epidermis of patients with vitiligo harbors increased H 2 O 2 levels [58,61], and the lesional epidermis exhibits higher NRF2, NQO1, GCLC, and GCLM expression levels compared with the non-lesional epidermis [97]. An enhanced oxidative damage may increase the vulnerability of melanocytes to oxidative damage [7,94], which can be counteracted by local HO-1 augmentation using psoralen plus UVA (PUVA) treatment [98]. However, IL-2-induced expansion of circulating melanocyte-specific CD8 + cytotoxic T cells (CTLs) [99] may bring about a reductive environment (thus reduces GSH/GSSG ratios) [100], decreasing serum HO-1 levels [7]. Collectively, the vitiligo lesional epidermis appears to suffer from high oxidative damage, which may in turn dampen normal antioxidative responses. This notion is further supported by the attenuated induction of phase II detoxification genes in the lesional skin after in vitro and ex vivo treatment with the electrophilic compounds curcumin and santalol [97]. In this meticulous experimental setting, isolated KCs were found to be more susceptible to apoptosis, whereas melanocytes were relatively resistant against apoptosis [97]. These results suggest that melanocyte-KC communication sustains overall redox balance in the epidermis [38] as well as circulation [7].

Management of Vitiligo
Vitiligo is not "just a cosmetic condition" but is psychologically devastating and stigmatizing [123]. The psychological impact on quality of life (QOL) is similar to that of other skin diseases, such as atopic dermatitis and psoriasis [124].
The aim of clinical management is to halt the autoimmune-driven depigmentation and restore the homeostatic pigmentation. Treatment options depend on several factors, such as disease subtype, extent, distribution, activity, patient age, phototype, effect on QOL, and motivation for treatment [42]. These treatments include topical therapies (e.g., corticosteroids and calcineurin inhibitors), phototherapies (e.g., photochemotherapies, narrowband UVB [NB-UVB], and excimer lasers or lamps], oral therapies (e.g., steroids and other immunosuppressants), surgery, and combination therapies [125]. Treatments are graded from first-to fourth-line options [125]. First-line treatment consists of topical therapy with corticosteroids and calcineurin inhibitors; second-line treatment, NB-UVB, PUVA, and systemic steroid therapy; third-line treatment, surgical grafting techniques; and fourth-line treatment, depigmentation therapies.
Although the abovementioned roles of oxidative stress in vitiligo rationalize antioxidantbased treatment, evidence of the efficacy of this treatment is quite limited [125]. To achieve repigmentation, pseudocatalase, vitamin E, vitamin C, ubiquinone, lipoic acid, Polypodium leucotomos, catalase/superoxide dismutase combination, and Ginkgo biloba may be administered with or without UV therapy [125]. Since the discovery of the role of the IFN-γ signaling axis, several clinical trials involving JAK inhibitors have been conducted [126]. JAK inhibitors, which target the type II IFN signaling pathway, have been shown to stimulate repigmentation in patients with vitiligo [127][128][129]. Tofacitinib, ruxolitinib, and baricitinib are the three major JAK inhibitors used for vitiligo. Ruxolitinib, an inhibitor of Janus kinase 1 (JAK1) and 2 (JAK2), was recently approved by the FDA to treat NSV in adult and pediatric patients aged ≥ 12 years. Ruxolitinib cream resulted in repigmentation through 52 weeks in phase 2 [130] and 3 [131] trials; however, its use is accompanied by acne and pruritus at the application site. Large-scale, long-term studies are required to elucidate the effects and risks of ruxolitinib cream application for vitiligo treatment.

Conclusions
Recent progress in vitiligo research has paved the way for disease pathway-based therapy. The IFN-γ-JAK-STAT pathway drives vitiligo pathogenesis, and JAK inhibitors, which presumably inhibit the effector function of CD49a + cytotoxic epidermal resident CD8 + T cells efficiently [132], hold promise for better management of this emotionally devastating ailment. In this review, we initially aimed to examine the role of the KEAP1-NRF2 system in melanocyte biology/vitiligo pathogenesis. It has turned out, however, that this role [5] appears too far-reaching to be a disease-specific pathway. Systemic activation of the KEAP1-NRF2 system by the Keap1-null mutation not only augmented phase II detoxification but also led to uncontrolled keratinization of the squamous epithelium (SE) [9]. We and others have characterized the roles of the KEAP1-NRF2 system in skin diseases involving aberrations in inflammation/keratinization (reviewed in [9]). The aggregated evidence underscores the prominent roles of the KEAP1-NRF2 system in epidermal biology. In summary, the NRF2/KEAP1 system is important in vitiligo but far more specific than a therapeutic target.

Future Directions
Cutaneous pigmentary/depigmentary disorders, such as vitiligo or lentigines, do not necessarily accompany aberrant keratinization or acanthosis. Nonetheless, when considering the SE as a pigmentary "unit" [10], similar to classic immune cell components (the epidermal proliferation [differentiation] unit comprising KCs and Langerhans cells) [133], the epidermal "niche" being the ultimate determinant of cellular behavior may be evident [38]. Previous reports support this notion; unlike other minor epidermal residents, melanocytes express the desmosomal cadherin DSG1 [134], one of the critical commitment factors of IFE differentiation [135]. Loss of DSG1 in epidermal KCs may lead to melanocyte loss from the epithelium and promote invasive/metastatic growth of transformed melanocytes (melanoma) [134,136], suggesting that IFE KCs "imprint" (or instruct) melanocytic behaviors. This reasoning is further supported by the classic morphological changes in epidermal melanosomes following the topical application of the antipolymerization agent 4-tertiary butyl catechol [137], which could also augment the response of epidermal KCs to cellular distress (i.e., keratinization) [9]. This treatment blocks the eumelanin synthesis pathway, causing the appearance of immature pheomelanosomes in hairless mice [137]. Compared with pigmentation in the hair, tanning responses (IFE melanogenesis) depend on the nature of differentiated McSC-derived melanocytes; wet-surfaced SE (squamous mucosa), palmoplantar epidermis, or its apparatus (the nail) hardly experiences tanning responses (caused by the predominance of eumelanogenesis over pheomelanogenesis). The "niche" instruction or "structural imprinting" [138] aspect of pigmentogenesis would be further rationalized when IFE differentiation is analogized to sulfur metabolism; thiol groups of the proliferative layer are converted to disulfide polymerized keratins [139] ( Figure 2). The principle, along with the possibility that the eumelanin synthetic pathway is regulated post-transcriptionally [35], tempted us to determine the "niche factor" within the IFE component. We have recently found that the IFE differentiating factor loricrin (LOR), which has a potent disulfide-linking capacity (reviewed in [138]), is indispensable for protection against UV radiation or electrophilic carcinogens (reviewed in [138]). Thus, we hypothesize that LOR could imprint the behaviors of IFE melanocytes and resident leukocytes [138]. The major effector of cornification (LOR) may act as a fate determinant of IFE melanocytes. Although further investigation and validation are required, revealing the hitherto unproven aspects of epidermal cell biology may lead to the development of mechanism-based skin pigmenting/depigmenting measures.
viewed in [138]), is indispensable for protection against UV radiation or electrophilic carcinogens (reviewed in [138]). Thus, we hypothesize that LOR could imprint the behaviors of IFE melanocytes and resident leukocytes [138]. The major effector of cornification (LOR) may act as a fate determinant of IFE melanocytes. Although further investigation and validation are required, revealing the hitherto unproven aspects of epidermal cell biology may lead to the development of mechanism-based skin pigmenting/depigmenting measures. Figure 2. Summary of thiols/disulfides in the epidermal pigmentary system. Interfollicular epidermis (IFE) pigmentation largely depends on the balance between reduction and oxidation (redox) status with internal or external causes. The former refers to mitochondria-derived reactive oxygen species (ROS) during melanogenesis, and the latter may correspond to epidermal differentiation in which keratinocyte (KC) structural proteins undergo extensive disulfide bridge formation upon the initiation of cornification (transition from the stratum granulosum [SG] to the stratum corneum [SC]). Successful cornification largely depends on the biochemical nature of the structural protein loricrin (LOR, indicated as granules). Differentiating layer (stratum spinosum [SS] and SG)-specific desmosomal cadherin desmoglein 1 (DSG1) appears indispensable for functional IFE pigmentogenesis: melanosome transfer and maturation (eumelanogenesis) (left). In vitiligo melanocytes, stress from environmental factors (e.g., toxic chemicals) can cause aberrations in the IFE antioxidant systems in genetically predisposed individuals. Breached immune tolerance recruits melanocyte antigen-specific CD8 + cytotoxic T cells (CTLs) expressing CXC receptor type 3 (CXCR3) or CD49 + resident memory T cells (TRMs) in vitiligo (right), whereas the autoreactive CTLs constitutively become anergic with the help of regulatory T cells (Tregs) in homeostasis (left). The homeostatic gradient of nuclear factor erythroid-2-related factor 2 (NRF2)-mediated epidermal antioxidative defense and ensuing cornification yield a polymerized/pigmented SC, protecting against oxidative damage. However, local clonal expansion of CTLs in the IFE eventually eliminates melanocytes from the IFE niche (the epidermal melanin unit), perturbing the xenobiotic metabolism coordinated by NRF2 and resulting in depigmentation (leukoderma).   Summary of thiols/disulfides in the epidermal pigmentary system. Interfollicular epidermis (IFE) pigmentation largely depends on the balance between reduction and oxidation (redox) status with internal or external causes. The former refers to mitochondria-derived reactive oxygen species (ROS) during melanogenesis, and the latter may correspond to epidermal differentiation in which keratinocyte (KC) structural proteins undergo extensive disulfide bridge formation upon the initiation of cornification (transition from the stratum granulosum [SG] to the stratum corneum [SC]). Successful cornification largely depends on the biochemical nature of the structural protein loricrin (LOR, indicated as granules). Differentiating layer (stratum spinosum [SS] and SG)-specific desmosomal cadherin desmoglein 1 (DSG1) appears indispensable for functional IFE pigmentogenesis: melanosome transfer and maturation (eumelanogenesis) (left). In vitiligo melanocytes, stress from environmental factors (e.g., toxic chemicals) can cause aberrations in the IFE antioxidant systems in genetically predisposed individuals. Breached immune tolerance recruits melanocyte antigen-specific CD8 + cytotoxic T cells (CTLs) expressing CXC receptor type 3 (CXCR3) or CD49 + resident memory T cells (TRMs) in vitiligo (right), whereas the autoreactive CTLs constitutively become anergic with the help of regulatory T cells (Tregs) in homeostasis (left). The homeostatic gradient of nuclear factor erythroid-2-related factor 2 (NRF2)-mediated epidermal antioxidative defense and ensuing cornification yield a polymerized/pigmented SC, protecting against oxidative damage. However, local clonal expansion of CTLs in the IFE eventually eliminates melanocytes from the IFE niche (the epidermal melanin unit), perturbing the xenobiotic metabolism coordinated by NRF2 and resulting in depigmentation (leukoderma).