Skin pigmentation is one of the most variable and noticeable phenotypes in humans, but the genetic basis and evolution of this polygenic trait has not yet been fully unraveled. The available evidence strongly suggests that the observed variation in this trait within and between populations has been shaped by natural selection. The general geographic patterns of skin pigmentation show a strong correlation with latitude and ultraviolet radiation (UVR) intensity. Skin pigmentation tends to be darker in equatorial and tropical regions (Sub-Saharan Africa, South Asia, Australia and Melanesia) where UVR levels are higher than in regions distant to the equator (Figure 1a,b). The current evolutionary hypothesis for skin pigmentation variation is the vitamin D/folate hypothesis, where a compromise may exist between the requirements for photoprotection on the one hand, and vitamin D3 synthesis on the other hand [1]. Dark skin would have a biological advantage under high UVR to protect from UV-induced sunburn, skin cancer, immune suppression and photolysis of folate, a metabolite essential for the development of embryonic neural tube and fertility, while light skin would have a biological advantage in regions far from the equator exposed to lower levels of UVR where UVB is needed and corresponds to the effective wavelengths for transformation of 7-dehydrocholesterol to vitamin D3 in the skin, and results in multiple effects on health, ranging from bone metabolism, innate immune response, cell proliferation and differentiation to fertility [2].

As described in the next section, skin color is determined mainly by epidermal melanin pigments, with a minor contribution from carotenoids and (de)oxyhemoglobin from dermal capillaries [4].

Melanin pigments are produced in the epidermis by highly specialized cells, the melanocytes, neural crest-derived cells migrating as melanoblasts, during embryogenesis, into the epidermis and hair follicles via the mesenchyme. Melanogenesis is a very complex process with multiple regulatory agents and pathways that could interact within the melanocyte but also in cooperation with the other cutaneous cells types (for a general review on melanogenesis, see [5]). Localized at the dermal–epidermal junction, each interfollicular melanocyte is connected through its dendrites to approximately 40 keratinocytes, establishing the so-called epidermal melanin unit [6]. Within melanocytes, melanin is synthesized in distinct, specialized membrane-bound organelles called melanosomes. Variations in constitutive pigmentation are attributed to differences in quality and quantity of the melanin content as well as in the size, amount, distribution and fate of melanosomes within the epidermal melanin unit rather than to differences in melanocyte density, as the latter is quite comparable among different ethnicities [7,8,9]. Melanocytes produce two chemically distinct types of melanin: the brown-black eumelanin, and the yellow-red pheomelanin [10,11,12,13,14]. The type of melanin produced depends on the function of melanogenic enzymes and the availability of substrates. Tyrosinase (TYR) is the key enzyme in melanogenesis and mediates the first steps in melanin synthesis, which involve the hydroxylation of the amino acid tyrosine to l-3,4-dihydroxyphenilalanine (DOPA) and subsequent oxidation to DOPAquinone. The latter is a key compound that undergoes further modification in two different pathways depending on the availability of the amino acid cysteine [15,16]. Cysteine reacts with DOPAquinone to produce 5-S-cysteinyldopa (5SCD) and 2-S-cysteinyldopa (2SCD). Cysteinyldopas are then oxidized by DOPAquinone to give benzothiazine intermediates, which gradually polymerize to form pheomelanin pigment. In the late stage of pheomelanin production, the benzothiazine moiety is gradually converted to benzothiazole moiety [17]. When cysteine is consumed in melanosomes, DOPAquinone spontaneously reacts to give, via DOPAchrome, 5,6-dihydroxyindole (DHI) and 5,6-dihydroxyindole-2-carboxylic acid (DHICA). The production of DHICA is accelerated by DOPAchrome tautomerase (DCT), also known as tyrosinase-related protein 2 (TRP2) or copper ions [18]. These 5,6-dihydroxyindoles are then further oxidized by tyrosinase and tyrosinase-related protein 1 (TRP1) to produce the eumelanin polymer.

It is generally accepted that eumelanin is photoprotective by limiting the extent of UV penetration within the epidermis and scavenging reactive oxygen radicals. In contrast, pheomelanin is not only weakly photoprotective against UV but also highly phototoxic and can enhance the UV-induced production of reactive oxygen species and further damage the cells [19,20,21,22].

Melanin synthesis takes place in melanosomes, a major advantage as it isolates melanogenesis from the rest of the cell and avoids the toxic effect of reactive quinoid melanin intermediates. There are four different maturation stages (I–IV) with increasing degree of melanization [23]. Rising from endoplasmic reticulum, Stage I melanosomes correspond to perinuclear and spherical premelanosomes where the internal matrix formed by the structural protein Pmel17 (gp100) begins to assemble. Stage II melanosomes are ovoid organelles containing longitudinally organized matrix fibers but without the presence of melanin. Stage III melanosomes show deposits of melanin along the matrix striations and tyrosinase activity is maximal. Stage IV melanosomes are opaque, filled with melanin and have no longer tyrosinase activity.

Besides tyrosinase and related proteins, melanosomal pH is also important in the regulation of skin pigmentation. Melanosomes derived from light skinned melanocytes are more acidic and have lower tyrosinase activity than melanosomes from dark skin which have more neutral pH and higher tyrosinase activity [24,25]. Actually, the activity of tyrosinase is greatly enhanced at neutral pH. Melanosomal pH regulating tyrosinase activity is thus a key parameter in the melanogenesis process. In darker skin types, it has been shown that Tyrosinase activity is increased [26] but without up-regulation of TYR gene expression [9]. To date, several ion channels, pumps and transporters have been identified as effectors of melanosomal pH among them ATP7A, a copper transporting ATPase, V-ATPase, a H⁺ pump, and the more specific OCA2/p protein, a chloride anion channel, SLC45A2/MATP a sugar transporter-like membrane protein, SLC24A5-encoded NCKX5 protein, a putative NA⁺/Ca²⁺ion exchanger pump and the Two Pore Channel 2 TPC2 (TPCN2 gene), a cation (Na⁺) ion channel [27].

More than 125 genes are known to regulate pigmentation. They regulate differentiation, survival of melanocytes, biogenesis and melanosome function which require a number of specific enzymes and structural proteins to mature and produce melanin.

The critical enzymes include tyrosinase (TYR), TRP1 and DCT, mutations of which dramatically affect the quality and quantity of melanin produced. Critical structural proteins include Pmel17 and MART1 both required for maturation of melanosomes l-tyrosine and l-DOPA, besides serving as substrates and intermediates of melanogenesis, are described to be also bioregulatory agents acting as positive regulators of melanogenesis enzymes and as regulators of melanocyte functions (e.g., proliferation) [28].

Hormonal regulation of pigmentation is crucial and one of the major determinants of pigment phenotype of the skin is the melanocortin 1 receptor (MC1R), a G protein-coupled receptor that plays a key role in the switch between eumelanin and pheomelanin synthesis within the melanocytes [29,30]. MC1R function is controlled by the agonists α-melanocytes-stimulating-hormone (α-MSH) and adrenocorticotropic hormone (ACTH) and by an antagonist, Agouti signaling protein (ASIP) [5]. Binding of MC1R by an agonist activates adenylyl cyclase, and increases levels of intracellular cAMP that activates protein kinase A (PKA) which in turn leads to increased transcription of microphthalmia-associated transcription factor (MITF) and, as a consequence, increased transcription of tyrosinase. Ultimately, this leads to an increased tyrosinase activity, resulting in eumelanin synthesis. Conversely, when ASIP binds to MC1R, there is a decreased tyrosinase activity resulting in pheomelanin production. Of note, MITF acts as a master regulator of melanocyte development, function and survival [31] and plays a role in the control of melanogenesis that leads to increased transcription of a range of genes including TYR, TRP1 and TRP2 involved in the control of the type and amount of melanins produced.

Constitutive skin pigmentation is a polygenic trait and, in recent years, the number of genes and the allelic variants that affect human skin pigmentation that have been identified has significantly increased. Studies of human pigmentation disorders, animal pigmentation models combined with the identification of regions in the human genome under positive selection in different populations, and genome-wide association studies (GWAS) have identified several single nucleotide polymorphisms (SNP) involved in human pigmentation, including SNPs that had not previously been implicated in pigmentation [32] (Table 1).

Approximately 15 genes have been associated with skin pigmentation variation in humans and some of the major ones are described as follows.

The melanocortin-1 receptor gene (MC1R) is the most well studied. This short intronless gene plays a key role in the switch between eumelanin and pheomelanin synthesis within the melanocytes. MC1R is highly polymorphic in European populations with more than 60 variants having been described. Several MC1R variants are associated with fair skin and red hair, sun-sensitivity and melanoma risk in Europeans [43]. In East and Southeast Asian populations, some polymorphisms have been found. Conversely, MC1R polymorphisms are absent in sub-Saharan African populations and very low in other dark skinned populations such as Papuans or South Asians showing that MC1R is under strong functional constraint in regions of high UVR [44].

The solute carrier family 24 member 5 (SLC24A5) gene is a cation exchanger that has a key role in melanosome morphogenesis and melanogenesis. Variants in SLC24A5 play a central role in skin lightening in Europeans, explaining approximately 30% of the difference in skin pigmentation between Europeans and West African populations [45]. The ancestral SNP (rs1426654 G) is present at very high frequencies in Sub-Saharan Africa, East Asia, Southeast Asia, the Americas, and Melanesia while the derived SNP (rs1426654 G>A) is found in Europeans and in geographically near populations in the Middle East, North Africa and Pakistan [35,45].

Another remarkably polymorphic gene is the membrane-associated transporter protein (MATP) also known as SLC45A2. In humans, mutations at MATP are responsible for oculocutaneous albinism type 4 (OCA4). SLC45A2 shows a pattern of polymorphisms very similar to that of SLC24A5: the derived SNP (rs16891982 C>G) is found in European and closely related populations while the ancestral SNP (rs16891982 C) is associated with darker pigmentation [35].

The oculocutaneous albinism type 2 (OCA2) gene has been under positive selection in Europe and East Asia, but different alleles have been selected in each region. The derived allele OCA2 rs1800404 C>T associated with light pigmentation occurs at high frequencies in European and East Asian populations whereas the ancestral allele (rs1800404 C) associated with dark pigmentation is highly frequent in Sub-Saharan Africans and Melanesians [35]. Furthermore, OCA2 variant rs1800414 A>G is prevalent in most East Asian populations and has reached very high frequencies in Chinese, Japanese and Korean populations but is absent in African and European populations [39,46]. Another hypofunctional OCA2 variant allele (rs74653330 G>A) was reported in East Asian populations living at highest latitudes (northern China and Mongolia) but was not observed in South Asian, European and African populations [47,48]. OCA2 is a major gene conferring skin lightening among East Asians, indicating that the evolution of light skin occurred, at least in part independently, in East Asians and European populations [39,48].

The agouti signaling protein (ASIP) is the antagonist of MC1R and upon binding to this receptor, ASIP promotes the synthesis of pheomelanin. The ASIP ancestral allele (rs6058017 G) resulting in decreased antagonism of α-MSH binding to MC1R and thus increased eumelanin synthesis has been reported to be associated with dark hair and brown eyes in European-Americans [49] and was found also associated with dark skin in African Americans [41], while rs6058017 G>A polymorphism leads to lighter skin color.

An interferon regulatory factor-4 (IRF4) SNP (rs12203592) is strongly associated with skin color, skin tanning response, freckling and nevus count [34,50,51,52]. The rs12203592 C>T polymorphism leads to reduced IRF4 gene expression and consequently reduced TYR expression, leading to sun sensitivity and blue eyes [53]. While Chinese, Japanese and African populations are homozygous for the rs12203592 C allele, only European populations harbor rs12203592 C>T polymorphisms [42]. This SNP shows a north–south gradient allele frequency distribution across Europe [54].

The tyrosinase gene (TYR) encodes a key enzyme in melanogenesis and is responsible for oculocutaneous albinism type 1 (OCA1). It is also involved in normal pigmentation variation. Two polymorphisms (rs1042602 C>A and rs1126809 G>A) associated with reduced tyrosinase enzyme activity appear at high frequency in Europeans and are absent in African populations [33].

In East Asian populations, ADAM metallopeptidase domain (ADAM17) SNP (rs4328603 C>T), ADAM metallopeptidase with Thrombospondin type 1 motif 20 (ADAMTS20) SNP (rs11182091 C>T, rs11182085 A>G, rs1510523 C>T) and attractin (ATRN) show strong specific signatures comparable to the SLC24A5 and SLC45A2 signatures in European populations [39]. Additional examples of East Asian-specific pigmentation-associated alleles include TRP1 (rs10809814) and DCT (rs1407995) [32,55].

Nevertheless, despite distinct genes and derived alleles in Europeans and East Asians, some genes, i.e., KIT Ligand (KITLG), ASIP and basonuclein (BNC2), have variants highly frequent in both populations.

Based on clear evidence for the selection of pigmentation-related genes in different populations, an evolutionary tree model for human skin pigmentation has been proposed in three populations East Asian, European and West African [3] (Figure 1c). This model suggests that skin lightened independently in Europeans and East Asian populations. The human tree starts with a hominin ancestor in Africa who probably had lightly pigmented skin covered with dark black hair that protected his skin from UVR [56]. As the lineage leading to the early Homo sapiens began to lose body hair, a strong selection for skin darkening emerged [57,58] (Branch 1). Microphtalmia-associated transcription factor (MITF) and endothelin 3 (EDN3) are the genes hypothesized to have been subject to positive selection. Later, the African and the European/East Asian populations diverged upon the “out of Africa” expansion (Branch 2), followed by the splitting of the East Asians and Europeans (Branches 3 and 4, respectively). The genetic mechanisms resulting in skin lightening in both Europeans and East Asians occurred independently, and after their divergence [35,45]. In Europeans, genetic selection for the evolution of light skin color has been confirmed for SLC24A5 [35,45] and SLC45A2 (MATP) [35,59]. In addition, myosin VA (MYO5A), dystrobrevin binding protein 1 (DTNBP1), TYRP1, ectodysplasin A (EDA), OCA2 and KITLG may have undergone European-specific changes and may thus have also affected skin lightening in Europeans (Branch 4). A parallel mechanism driving skin lightening (convergent evolution) may have occurred in East Asians (Branch 3). In particular, ADAM17 and ADAMTS20 show strong selection signatures in East Asian populations, such as SLC24A5 and MATP in European populations. DCT, MC1R, lysosomal trafficking regulator (LYST), EDA, OCA2 and ATRN also show relatively strong signatures of distinct East Asian selective events (Branch 3). However, it seems that gene selection was not that straight forward. Two additional genes, ASIP and BNC2, show evidence of selection in both Europeans and East Asians. This would suggest that these genes were selected prior to the divergence of these populations (Branch 2). Two other genes, LYST and KITLG, appear to have had selective events that may have started on Branch 2 and then continued on to Branches 3 and 4. West African populations also experienced adaptive pressure for skin darkening relative to the ancestral human population, but the authors found less evidence for this and have placed no genes on Branch 5.

Recently, two studies have shed light on the architecture of skin pigmentation in African skin and on regions of the genome not previously associated with skin color [37,38]. Four regions of the human genome associated with skin color in African populations living in Ethiopia, Tanzania and Botswana have been identified in or near SLC24A5, MFSD12 (Major facilitator superfamily domain containing 12), DDB1 (Damage specific DNA binding protein 1)/TMEM38 (Transmembrane protein 38) and OCA2/ HERC2 (HECT And RLD domain containing E3 ubiquitin protein ligase 2) [37]. The SNPs with the strongest association with skin color in Africans was SLC24A5. The light variant allele (rs1426624A) already known to frequently occur in Europeans, Pakistani, Central Asians and North Indians [36] was shown to be common (28–50%) in Ethiopians and Tanzanians with Afroasiatic ancestry and at moderate frequency (5–10%) in South African and Botswanan populations with low levels of East Asian ancestry and recent European admixtures. This light pigmentation variant at SLC24A5 was thus likely introduced into East Africa by gene inflow from non-Africans. The second strongest association with skin pigmentation is due to the MFSD12 gene, a transmembrane solute transporter that locates to late endosomes and/or lysosomes in melanocytes and not to eumelanosomes. Depletion of MFSD12 increases eumelanin content. A third genomic region associated with pigmentation contains the DDB1/TMEM38 genes. DDB1 plays a role in DNA repair. The fourth region of significantly associated SNPs encompasses the OCA2/HERC2 loci. Variants associated with dark pigmentation at MFSD2, DDB1, OCA2 and HERC2 are found in Africans, which are also shared by southern Asian and Australo-Melanesian populations. In contrast to the lack of variation at MC1R in Africans [60], both light and dark alleles at MFSD2, DDB1, OCA2 and HERC2 arose in hominid and continued to evolve. In southern African populations with considerably lighter skin than equatorial Africans, novel variants associated with skin pigmentation have been identified in canonical SLC24A5, TYRP1 and non-canonical pigmentation genes, including near SMARCA2 (SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily A, member 2)/VLDLR (Very low density lipoprotein receptor) and SNX13 (Sorting nexin 13) [38]. VLDLR and SMARCA2 have been implicated in pigmentation in model organisms and in vitro studies. SNSX13 regulates lysosomal degradation and G protein signaling but has not previously been associated with skin pigmentation. Thus, although genetic sequencing has shed light on many but probably not all the genes, more variants involved in variation of skin pigmentation in human populations still remain to be discovered.

There is a need for an objective tool to classify skin pigmentation in dermatological and cosmetic research. Furthermore, an accurate description of the color phenotype is necessary to compare different genetic studies of human skin pigmentation or to determine the role of skin color in responses to UVR.

Measurements of melanin in human skin and particularly in skin with different constitutive pigmentation are scarce [94,95,96,97,98], as compared to melanin analysis in mice, cell culture or hair [12].

Apart from melanogenesis regulation within melanocyte, much evidence is now available on the regulation of melanogenesis through the interactions and cross talk between melanocytes and other cutaneous cell types, mainly keratinocytes and dermal fibroblasts [118,119].

Paracrine regulation is now well established. One of the major mechanism of pigmentation regulation acts through the POMC local production and processing to CHR-related peptides, α-MSH, ACTH and β-endorphin [120]. Many other keratinocytes derived soluble factors, growth factors and cytokines have been identified such as basicFGF (bFGF), endothelin 1, IL1α, SCF, GM-CSF, HGF, GROα, PGE2/PGF2a and LIF (for review see [121]). Most of them have been clearly associated with melanocyte or melanogenesis activation upon UV exposure and some have also been suggested to play a role in the development of hyperpigmentary disorders, such as lentigo senilis [122]. On the other hand, some of them, such as cytokines such as IL-1, are also regulators of POMC expression [120]. However, few data have identified specific keratinocyte actors or derived factors associated with the level of constitutive pigmentation. A study from Yin et al. analyzed gene expression in the epidermis obtained from skins of African, Asian and Caucasian subjects. After separating the epidermis using the suction blister technique, differences between Asian and Caucasian skin could not be highlighted except for the microtubule associated gene NINL which was upregulated in Asian skin [123]. In contrast, a differential expression profile between Asian/Caucasian and African skin was found. Among the genes differentially expressed, some were already known to be associated with skin color type such as FRZB, CDH12 or KITLG. Many of the modulated genes encode receptors and their ligands. This emphasizes the interactions between epidermal cell types, but at that stage no functional validation has been provided. Among the candidates, CCL18 or S100A4 showed a weaker expression in dark skin phototypes. The expression of protease-activated receptor-2 (PAR-2), has also been shown to be higher in darker skin types compared to fair skin phototypes, suggesting a role for melanosome phagocytosis in keratinocytes [105,124]. The role of transfer of melanosome in keratinocytes has also been highlighted by the higher expression of Rab27 in darker skin types [125].

The dermal compartment has also been shown to play a role in pigmentation regulation, through the extracellular matrix proteins such as laminins and collagen IV but also fibroblastic soluble factors. Some of these are common with keratinocytes i.e., SCF and HGF but others are specifically produced by dermal fibroblasts and can enhance melanin synthesis and melanocyte proliferation such as KGF/FGF7. The contribution of dermal fibroblasts to skin pigmentation has been clearly shown in three dimensional reconstructed skin models performed under various fibroblast conditions [126]. Dermal factors have also been identified to be abnormally expressed in pigmentary lesions, such as KGF/FGF7, HGF and SCF in lentigo senilis [127,128] and an increased secretion of SCF and HGF by dermal fibroblasts in neurofibromatosis type 1 (NF-1) disease [129]. Dermal soluble factors have also been demonstrated to contribute to constitutive pigmentation. Dermal fibroblasts from the lighter anatomical skin sites of the body, the soles and palms, were shown to produce higher levels of Dickkopf 1 (DKK1), a Wnt pathway antagonist compared to fibroblasts from other parts of the body [130,131] and a role of DKK1 in the down-regulation of melanocyte proliferation and melanin synthesis was demonstrated. The group of V. Hearing also addressed the question of the role of dermal fibroblasts in the level of constitutive pigmentation. They performed a cDNA array gene expression analysis on dermal fibroblasts isolated from skin of phototypes SPT I, III and VI. They were unable to find any difference in the expression of well-known fibroblastic regulators of melanogenesis such as SCF, bFGF or DKK1 but identified among the differentially expressed genes, Neuregulin 1 as a positive regulator for melanogenesis to be highly expressed in fibroblasts from skin type VI compared to those of lighter skin phototypes [132,133]. Among the mechanisms of cell–cell interactions, there is a growing interest in the role of secreted vesicles called extracellular vesicles (EVs), including the exosomes, the endosome-derived vesicles. Exosomes are secreted in the intercellular space and contain proteins, lipids and RNAs among them miRNAs, the small non coding RNAs known to regulate major biological functions [134,135] (for miRNAs, see [136]). Epidermal keratinocytes have been shown to release exosomes and their content differs depending on the stage of keratinocyte differentiation. Keratinocyte-derived exosomes may also be responsible for the dialog with dermal fibroblasts [134,135,136]. More recently, a characterization of EVs from keratinocytes from Caucasian and African skin types has been reported [137]. Apart from the classical exosomal component (histocompatibility complex class I, CD9, CD81, and Hsc70) and specific keratinocyte proteins such as keratins and stratifin, the authors analyzed the miRNAs profile. Although the amount of released exosomes did not differ between skin of different ethnic origin, exosomes from African skin keratinocytes play a role in the stimulation of pigmentation upon UV exposure but more interestingly, exert a melanogenesis stimulation on Caucasian melanocytes. This effect was associated with an increase in Tyr activity and MITF expression suggesting a role of Black keratinocytes in the regulation of pigmentation in Black skin. Among the components of these exosomes 30 miRNAs were differentially expressed between both ethnic skin origins among them miR203, the major miRNA already known to be implicated in keratinocyte differentiation [138] which showed a higher expression in exosomes from Black skin. MiR 203 has also been demonstrated to regulate pigmentation in melanoma cells but also in normal melanocytes by targeting Tyr and Kif5b without effect on MITF [139]. Other miRNAs have been shown to interfere with pigmentation and melanogenesis such as miR3196 or miR145 involved in the UVB response [137,140]. In patients with melasma, the down-regulation of miR675 stimulates pigmentation. MiR675 has been shown to be secreted in keratinocytes derived exosomes and has MITF as a direct target [141].

Although an increasing number of studies revealed the role of some miRNAs in the regulation of pigmentation and pigmentary disorders, the number of identified miRNAs involved in the regulation of constitutive skin color is still quite low. Some data have reported their functional role in animal coat color such as miR137 in mice [142] or miR25 in the skin of alpacas [143].

Erythema or sunburn is a skin reddening caused by increased blood flow. It is the most widely used clinical endpoint in photobiology and is used as an indicator of UVR sensitivity by estimating the minimal erythemal dose (MED) as the lowest UVR dose able to induce a just perceptible skin redness (or redness with defined borders) 24 h post-exposure to a series of increasing UV doses [147].

Several studies have shown a relationship between MED and skin color, i.e., increased MED with skin type or skin color [96,148,149,150,151,152,153,154,155,156]. A powerful association between MED and skin colorimetric parameters and particularly the L* parameter was shown [157].

Dark skin has been shown to have an intrinsic sun protection factor (SPF) of 13.4, compared to 3.4 for light skin [158,159]. Additionally, dark skin transmits 7.4% of UVB compared to 29.4% in light skin [160]. These studies show that approximately four times as much UV reach the upper dermis of light skin compared to dark skin.

The relationship between various measures of constitutive skin pigmentation and erythema caused by solar-simulated radiation (SSR) UV, 290 and 310 nm UV was analyzed and a significant relationship (p < 0.001) between MED and skin color (SPT I-IV) was found using SSR and 310 nm UVB but not for 290 nm, thus suggesting that basal layer melanin only protects against erythema with wavelengths >290 nm [161].

Erythema is clinically associated with the induction of sunburn cells (SBC), i.e., apoptotic keratinocytes histologically characterized as dyskeratotic and vacuolated keratinocytes 24 h after UVR exposure [162]. SBC can be used as endpoints for photoprotection studies and are the basis of the biologically efficient dose (BED) which is the minimal UVR dose required to induce one SBC per 0.45 µm of the epidermis [163]. BED in vitro is equivalent to MED in vivo because SBC become apparent at 1 MED. SBC were quantified in skin samples of variable constitutive pigmentation 24 h after exposure to increasing SSR doses and the results showed a statistically significant correlation between skin color measured by ITA and BED (R² = 0.70; p < 0001), i.e., the darker the skin the higher the BED (Figure 3a,b), similar to minimal erythemal dose MED [71].

Cyclobutane pyrimidine (CPD) dimers are the major UV-induced DNA lesions in human skin and are indisputably the causal events in the development of skin cancers [164,165,166,167]. Several studies have demonstrated that UVR exposure induces DNA damage in all skin types and that the extent of such damage is inversely related to constitutive pigmentation of the skin [96,168,169,170]. CPD may trigger erythema, immunosuppression and photoaging that may increase the risk of skin cancer [171]. The action spectra for DNA damage (CPD), melanogenesis and erythema are very similar, suggesting that DNA is the chromophore for melanogenesis and erythema [147,172] and erythema can be considered as a clinical surrogate for DNA damage.

Skin color has been shown to influence the amount and distribution of DNA damage in the epidermis. An inverse relationship between melanin content and DNA damage induced by exposure to 1 MED of UVA/UVB was reported in individuals of different ethnic origin [96]. DNA damage-inducing effectiveness of a fixed sub-erythemal dose (100 J/m²) of UVA/UVB was compared on three ethnic groups Caucasian, Asian and African-American and showed a uniform distribution of CPD throughout the epidermis in lighter skin, while CPD declined with increasing distance from the skin surface in dark skin [168]. Dark skin provided a protection factor of approximately 3 for CPD. The DNA damage in the epidermis of different ethnic groups exposed to 1 MED of UVA/UVB was investigated and showed that CPD damage was similar in the upper and lower epidermis of Caucasian skin, while in the lower epidermis it was 1.5-fold and 2.1-fold lower than in the upper epidermis of Asian/Hispanic and African-American skin respectively, meaning that the upper epidermis of dark skin is more photoprotective than that of fair skin [173]. The erythema and CPD protection factor (PF) was determined by exposure to 2 MED SSR one week after stimulating tanning in skin type II and IV with repeated sub-erythemal SSR UV doses for two weeks [174]. The results showed that for both skin types PF were equivalent to 2–3, supporting the hypothesis that DNA is the chromophore for erythema as suggested by Young et al. [147]. Recently one study [175] determined the PF against DNA damage of constitutive pigmentation by comparing the SSR-induced CPD in buttock skin of fair skinned (SPT I/II n = 6) and dark skinned (West Africans, SPT VI, n = 6) participants and showed that melanin in dark skin provides protection against CPD by 8-fold in the whole epidermis and by 59-, 16.5- and 5-fold in the basal, middle and upper epidermis respectively. Although based on a small sample size and on limited skin color types, the basal and middle layer DNA protection factor are aligned with the relative differences in BCC (×60) and SCC (×20) of light and dark skin [176] (see Skin cancers section).

An inverse relationship between constitutive pigmentation measured by ITA and CPD on samples collected immediately after SSR UVR exposure was shown [71] (Figure 3c). Whereas for Light, Intermediate and Tan skin, DNA lesions were detected in all epidermal layers including the basal layer and the upper dermis, they were present only in the suprabasal layers in Brown and Dark skin. Lesions at the origin of potential mutations in the basal proliferative layer of the epidermis may explain, from a biological standpoint, the higher susceptibility of light skin types to develop UVR-induced skin cancers. The results also show that more pigmented skin, namely Intermediate and Tan skin types, comprising notably people of Hispanic, African-American and Asian origins, are also subject to this type of damage.

DNA damage occurring specifically within melanocytes was also investigated in skin with different ethnic origin or constitutive pigmentation. A 7.5-fold higher CPD levels in melanocytes of Caucasian skin, and 2.9-fold higher in Asian/Hispanic skin compared to African-American skin was shown one day after a single 1 MED UV dose [173]. A dose-dependent increase in DNA lesions within melanocytes of skins with variable constitutive pigmentation measured by ITA (Figure 3d) and prevalence of CPD-positive melanocytes in Light, Intermediate and Tan skin reaching 79–100% at the BED was reported [177]. Lesions in melanocytes may explain the higher risk of lighter skin types of developing melanoma. In addition, they may explain the high prevalence of pigmentation disorders in Tan skin types in individuals of Asian, Hispanic/Latin American and African descent [178,179,180,181,182,183].

Although many in vivo and in vitro studies in human and animal models found no difference in DNA repair capacity in skin of different constitutive pigmentation [96,184,185,186,187,188,189], one study [174] showed a greater loss of CPD in skin type IV compared to type II after exposure to sub-erythemal doses suggesting faster repair in skin type IV. In addition, there is now evidence that MC1R can contribute to a protective effect through activation of DNA repair [190,191,192].

The tanning response of human skin to sun exposure can be categorized into three different phases: immediate pigment darkening (IPD), persistent pigmentation (PPD) and delayed tanning (DT). IPD is a reversible transient grayish coloration that occurs during and after UV exposure and fades within minutes to a maximum of 2 h. When the skin is exposed to a sufficient UV dose, IPD is more intense and is followed by PPD that may last 24 h or longer and may blend with the DT response which requires melanin neo-synthesis and occurs between 3–5 days after sunlight exposure [193,194,195,196]. IPD typically appears gray to black while PPD is brown. Both IPD and PPD do not require any new pigment synthesis and are thought to result from oxidation and/or polymerization of pre-existing melanin or melanogenic precursors and metabolites [197,198,199,200]. Melanosomes have also been shown to redistribute within both keratinocytes and melanocytes during the IPD/PPD response [201,202]. PPD is not protective against neither UVB-induced erythema [203] nor UVB-induced DNA lesions [204,205]. For some authors, the IPD/PPD response can develop in all skin phototypes [206,207,208] (Figure 4).

In volunteers from three different ethnic origins (European, Indian, African) and ITA values ranging from 13° to −50°, exposure to 45–50 J/cm² UVA1 induced a similar ΔITA around 16–18° irrespective of the original skin color type [208]. For other authors, there is a lack of IPD in skin phototypes I and II [209]. This may be due to difficulties in evaluating color change in fair-skinned individuals compared to easier observation in higher skin phototypes [206,207,209]. In very fair skinned individuals, this may also be due to the irradiation dose needed to produce IPD/PPD being greater than the dose required to induce sunburn. A lack of IPD response in fair skinned individuals was suggested to be correlated to “non-brown” eye color rather than light hair color [148]. Even if some slight discrepancies could be found between the articles, IPD/PPD action spectrum studies have shown that, in the UV range, the UVA wavelengths are the most efficient with a peak between 340 and 370 nm [193,199]. However, when the source of exposure extends up to 620 nm, short wavelengths of visible light (400–500 nm) appear to be also highly effective [210]. Upon UVA exposure eumelanin undergoes a reversible oxidation of 5,6-dihydroxyindole (DHI and DHICA) moiety to 5,6-indolequinone moiety (for review, see [211]). IPD may fade away through reduction of indolequinone to the original dihydroxyindole moiety. PPD develops by a more complex, irreversible chemical process involving oxidative cleavage of indolequinone to form free pyrrole-2,3,5-tricarboxylic acid (PTCA) [212] and cross-linking of dihydroxyindole that gives pyrrole-2,3,4,5-tetracarboxylic acid (PTeCA) [213]. Both oxidative modifications of eumelanin and cross-linking occur in tandem and lead to increased absorption in the visible range. Pheomelanin also undergoes photodegradation during PPD as UVA exposure induces oxidative conversion of benzothiazine (4-AHP) to benzothiazole (TTCA) moiety. However, in terms of skin pigmentation, pheomelanin may not contribute much to PPD due to its lower content compared to eumelanin [78].

DT is a brown dark yellow pigmentation that appears several days after UV exposure and reaches a maximum after 2–3 weeks depending on the UV dose and skin type [214,215]. It is maximally stimulated by the sunburn/erythema spectrum (290–320 nm) and to a lesser extent by UVA exposure. It results from neo-melanization (de novo melanogenesis) due, to increased activity and number of melanocytes, dendrite branching, tyrosinase activity and number and transfer of melanosomes. Multiple photobiological events initiate melanin synthesis, such as DNA damage/DNA repair [216] membrane damage and factors secreted by surrounding keratinocytes. Activation of P53 protein in keratinocytes as a result of UV-induced DNA insult leads to upregulation of proopiomelanocortin (POMC) which is processed into αMSH. However, the direct regulation of POMC by P53 is still controversial since no P53 responsive element has been identified in the POMC promoter region [217]. Keratinocyte derived αMSH then binds to melanocyte membrane through MC1R resulting in activation of melanogenesis and increased melanin production. Pigmentation resulting from the stimulation of the cutaneous hypothalamic-pituitary axis (HPA) has been studied with regard to the contribution of the different UV wavelengths [218]. UVB was highly effective in stimulation of CRH (Corticotrophin Releasing Hormone), POMC, MC1R, MC2R and β-endorphin whereas UVA was able to stimulate β-endorphin and to a lesser extend POMC. Other keratinocyte derived factors, such as cytokines play a role similar to endothelin 1 (ET-1) or interleukin 1. Membrane effects upon UV exposure are also contributing through release of diacylglycerol (DAG) and subsequent activation of protein kinase C (PKC), a major protein kinase involved in tyrosinase activation. Arachidonic acid metabolites such as leukotrienes C4 or D4 are also known to be mitogenic for melanocytes.

DT due to resulting increased melanin content has been shown to be slightly photoprotective leading to SPF around 2 [219], based on its protective value from erythema and DNA damage. DT response differs depending on skin color type, as clearly evidenced by the Fitzpatrick classification, at least for skin types I–IV, taking into account the individual tanning ability from “never tan” to “always tan”. Facultative skin color has been shown to be proportional to constitutive skin color [70]. From a mechanistic point of view some articles have addressed the question of differences between skin color types or ethnic origin of the skin. From a global point of view, tanning response whatever the skin color type does not lead to a huge increase in melanin amount. In Caucasians, Asians and Africans, after exposure to 1 MED, the melanin content was not statistically increased one week later, but mostly redistributed in the epidermal layers [220]. The authors suggest that this phenomenon highly contributes to the visible induced pigmentation. In another study, even after three weeks of repetitive UV exposure of subjects phototypes II/III, melanin content increased only slightly without statistically significant levels despite the fact that visible tanning of the skin had increased fivefold, as measured by a chromameter [221]. Alaluf et al. [10,222] confirmed this moderate increase in melanin content in repetitively UV-radiated skin compared to protected skin. In Asian skin, UV tanning response was also reported to be associated with a small increase in pigmentation [94]. Melanocyte density was found not to be increased at seven days after a single 1 MED exposure [220], while after repetitive exposures it has been shown to reach a threefold increase at five weeks [223]. Although pheomelanin synthesis within the melanosome seems to occur prior to eumelanin [224], the amount of melanin observed after UV exposure, appears to result from a slight increase in both eumelanin and pheomelanin in a constant proportion [94,96].

Visible light (VL) represents almost half of the solar spectrum and its role in IPD/PPD response has been shown since a long time with a contribution of shorter VL wavelengths around 420–460 nm as shown in IPD/PPD action spectra studies [193,210,225]. Using a solar simulator delivering wavelengths from 385 to 690 nm, the pigmentary doses were established at 80–120 J/cm² in skin phototypes II-IV [226]. More recently, the effects of VL using more controlled VL sources has been revisited. A study compared UVA1 and VL in pigmentation induction [227]. Using a halogen VL source with a higher intensity compared to the sun, the minimal VL pigmentary dose was found at 40 J/cm² and dark skin individuals (SPT IV–VI) responded to VL while no pigmentation was induced in lighter skin type subjects (SPT II). The pigmentation was browner compared to UVA1-induced pigmentation (more greyish) at early time points after exposure and VL was responsible for a more sustained effect. In natural sunlight exposure conditions (Bangalore, India, noon time), the pigmentary effect of VL has been assessed after blocking the UV wavelengths using physical filters. Twenty-five minutes of exposure were sufficient to induce a measurable pigmentation in subjects with phototypes IV-V. In this study, the authors indicate that the contribution of UV and VL seems comparable but with a 25 times higher efficacy of UV (per J/cm²) [228]. To address the respective role of the different VL color subdomains, Duteil et al. used Light Emitting Diodes (LED) with a peak within the shorter wavelengths corresponding to blue-violet light (LEDs emitting around 415 nm) and others with a peak at 630 nm corresponding to red light. They demonstrated that induced-hyperpigmentation in melanocompetent subjects (SPT III–IV) was only observed under blue light exposure with a change in ITA values up to 20 ITA units [229]. They also showed that the mechanisms underlying this pigmentation were different from the UVB-induced pigmentation without contribution of P53 or oxidative stress. To better understand the contribution of VL and UVA1 in induced pigmentation, 10 subjects SPT IV-VI were exposed to pure VL or VL with 0.5% UVA1 and assessments were performed from 24 h to 14 days post exposure. Pigmentation induced by both types of exposure indicate that, when combining the two wavelength domains an additive effect could be observed with regard to pigmentation intensity but erythema was only seen where skin was exposed to VL + UVA1 [230]. The mechanisms underlying blue light-induced pigmentation has been recently reported to involve the presence in the melanocyte membrane of a light photoreceptor belonging to Opsins family, the OPN3 well known to detect blue light through the retina [231]. Activation of OPN3 by blue light stimulates the calcium flux within the melanocyte which in turns activates MAP kinase pathways leading to increased melanogenesis. Interestingly, OPN3 activation was also responsible for the formation of TYR complexes with other melanogenic proteins, leading to a sustainable Tyrosinase activity, and this specifically in dark skin melanocytes (SPT III–VI) compared to lighter ones (SPT I and II).

The role of solar exposure is considered in the dermatological practice to be an important contributor to pigmentary disorders and efficient photoprotection is usually recommended. The development of protocols aiming at testing the efficacy of VL blocking agents has also been recently published [232].

UVR exposure causes skin cancer including carcinomas, i.e., basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) which are keratinocyte-derived cancers and, more dramatically, malignant melanoma (MM), which arise from melanocytes, and are responsible for the vast majority of skin cancer deaths due to metastasis proneness (American Cancer Society, 2016). Epidemiologic studies show higher incidences of BCC and SCC as well as MM in Caucasians compared to African-Americans [233,234,235,236,237]. Recent data from South African population demonstrated that populations of African and Asian ancestry had a lower incidence of SCC, BCC and MM than people of mixed European descent and considerably lower than individuals of European descent [176]. A 60-fold and 20-fold higher incidence of BCC and SCC, respectively, was reported in individuals of European ancestry compared to African ancestry. In the United States, the rates of melanoma have been shown to vary with ethnicity and, between 2010 and 2014, the age-adjusted incidence rates per 100,000 in men and women were 34.4 and 20.9, respectively, in white individuals; 5.0 and 4.7, respectively, in Hispanics; 4.3 and 4.9, respectively, in American Indian/Alaskan natives; 1.6 and 1.2, respectively, in Asian/Pacific Islanders; and 1.1 and 1.0, respectively, in black individuals [238].

Although the risk of skin cancer is lower in darker skin compared to lighter skin types [239], UVR exposure is a risk factor for skin cancer even in darker skin types including African Americans, Asians and Hispanics [235,240,241,242,243] and epidemiologic studies have shown that sunburn occurs even in darkly pigmented skin [188,197].

The lower rates of skin cancers found in pigmented skin can be explained by photoprotection provided by melanin. Dark skin has been shown to have an intrinsic UVB protection factor of 13.4, and a UVA protective factor of 5.7 compared to 3.4 and 1.8, respectively, in light skin. Additionally, dark skin transmits 7.4% of UVB and 17.5% of UVA, compared to 29.4% and 55.5%, respectively, in light skin [158,159,160].

Even though skin cancer is less frequent in nonwhite/dark skinned individuals, it is often associated with increased morbidity and mortality due to the fact that some perceive themselves to be at a lower risk for developing sunburn and skin cancers [244,245]. There also is a lack of awareness of these diseases among these populations, associated with delayed diagnosis [240,246,247,248,249,250,251,252].

Skin aging involves two types of processes, intrinsic aging (chronological aging) and extrinsic aging also called photoaging, considered as a premature skin aging and resulting from the impact of environmental stresses, essentially solar UV exposure. The overall aging signs have been shown to differ depending on the ethnic origin and related to facial structural differences (for reviews, see [253,254]). In this section a focus on photoaging is made. Although clinical manifestations of skin photoaging include pigmentary disorders, those aspects are detailed in a specific section. One of the major clinical traits of photoaging is the progressive appearance of coarse wrinkles associated with the formation of dermal solar elastosis, (Figure 5a,b), which results from alterations and reorganization of the whole dermal structure.

The entire UV spectrum is involved in photoaging. UVA wavelengths have been shown to highly contribute due to their high penetration properties, which allows them to directly reach the dermal compartment. Most of the literature on this subject is related to ethnic origin rather than to properly classified skin color phototypes. More than differences in the nature of clinical signs, photoaging differences can be found regarding the age of onset and the severity of damage that can be explained by the fact that UV and especially UVB rays penetrate in different amounts depending on skin color type and melanin content. Wrinkling process appears to highly depend on skin color type being delayed in darker skin types (African Americans) compared to Caucasians [255]. In addition to lower amounts of UV reaching the dermis, darker skin types have been shown to have a thicker dermis, which has more elasticity and resistance [254]. Clinical manifestations of aging appear 10–20 years later in pigmented skin populations compared to age-matched white populations [158]. A 10 year delay has also been found between Chinese and European populations [178]. In Asians, although the principal clinical signs of skin phototoaging are pigmentary changes, wrinkle patterns can also develop [256,257] even if due to differences in sun exposure habits it is not possible to strictly compare European and Asians photoaging process. Wrinkle patterns in Asians are characterized by coarser, thicker and deeper wrinkles compared to Caucasians [258]. Wrinkling in Asians women is reported as not noticeable before the age of 50 and that the degree is less severe compared to European women [259]. However, in both Caucasian and Asian populations the biological events related to the photoaging process seem quite similar with an upregulation of matrix metalloproteinases (MMPs), especially MMP1 and decreased collagen content resulting from down-regulation of expression of main structural collagens notably Col1 and Col3 [260,261]. Although the mechanisms seem quite similar, one study showed a higher MMP1 increase in lighter skin phototypes (I/II) compared to darker skin types (V/VI) after UV exposure [262]. Another study found that the induction of MMP-1 and MMP-9 to erythemal doses of SSR is similar in all skin types from I to VI but different when the same physical doses were delivered with increase in MMPs in lighter skin types and not in skin type VI [169].

Beyond the tanning effect induced by acute sun exposure, pigmentary disorders can be triggered or exacerbated by mid or long-term sun exposure. Pigmentary disorders, even benign, cause psychosocial distress and negative impact on self-image and the quality of life of individuals [263]. Type, onset, frequency and severity of hyperpigmented lesions depend on the skin complexion and on individual ancestry. High prevalence of pigmentary disorders in dark skinned people (South/southeast Asian, Hispanic/Latin American, African descent) has been reported [179,181,182,183,264]. Pigmentary disorders are the third and fourth more common dermatoses in Blacks and Hispanics respectively [181]. However the special importance of skin hyperpigmentations in Asian populations, from lighter to darker skin tone has also been however highlighted [178,259,265,266,267]. Hyperpigmented spots develop earlier and/or are more pronounced in Asian than in Caucasian population. A higher spot prevalence was reported in Chinese women than French women, both of them having mostly intermediate skin complexion [178]. In addition, over the age of 40, Chinese women who have pigmented spots, have many spots, whereas, in France, women have only a few spots. These results indicate that pigmented spot intensity is an earlier ageing feature in Chinese women than in French women.

A focus is made on four major hyperpigmented disorders linked to the light exposure: melasma, post-inflammatory hyperpigmentation (PIH), seborrheic keratosis (SK) and solar lentigo (SL) (Figure 5c–f).