Skin colour and vitamin D: An update

Homo sapiens evolved in East Africa and had dark skin, hair, and eyes, in order to protect against deleterious consequences of intensive UV radiation at equatorial latitudes. Intensive skin pigmentation was thought to bear the risk of inefficient vitamin D3 synthesis in the skin. This initiated the hypothesis that within the past 75 000 years, in which humans migrated to higher latitudes in Asia and Europe, the need for vitamin D3 synthesis served as an evolutionary driver for skin lightening. In this review, we summarize the recent archeogenomic reconstruction of population admixture in Europe and demonstrate that skin lightening happened as late as 5000 years ago through immigration of lighter pigmented populations from western Anatolia and the Russian steppe but not primarily via evolutionary pressure for vitamin D3 synthesis. We show that variations in genes encoding for proteins being responsible for the transport, metabolism and signalling of vitamin D provide alternative mechanisms of adaptation to a life in northern latitudes without suffering from consequences of vitamin D deficiency. This includes hypotheses explaining differences in the vitamin D status and response index of European populations.


| INTRODUC TI ON
Most anthropomorphic properties of humans, such as skin, hair and eye colour, [1] are based on the genotype of individuals, that is on the function and expression of their genes. Nowadays, these traits vary a lot between human populations, but they are the result of the rather recent evolution of anatomically modern humans (homo sapiens). Skin, hair and eyes were dark, when some 50-75 000 years ago homo sapiens left its origin in equatorial East Africa and migrated first to Asia and Oceania and later to Europe and the Americas. [2,3] In particular in northern Europe and Siberia, human skin and hair turned lighter and blue eyes occurred leading throughout Europe to a North-South gradient in pigmentation. [4] This leads to the question, whether in new geographic environments, such as regions of rather high latitude, there were evolutionary drivers for these changes in pigmentation intensity.
Vitamin D 3 is a prehormone that is essential for proper bone formation and a functional immune system. [5] The main way of providing the human body with vitamin D 3 is its synthesis in UV-B-exposed skin. [6] At higher latitudes, UV-B exposure is significantly lower and displays prominent seasonal changes. [7] Accordingly, in winter there is no endogenous production of vitamin D 3 possible so that it needs to be taken up by diet or pills, whereupon the molecule can act as a dietary vitamin. The fact that there are only few food sources naturally rich in vitamin D 3 makes its endogenous production essential.
This led to the hypothesis that the need for a more efficient synthesis of vitamin D 3 may have acted as an evolutionary driver for skin lightening, in order to avoid deleterious consequences vitamin D deficiency, such as musculoskeletal diseases and immunosuppression. [8,9] Interestingly, a recent large twin study estimated that half of the genetic variance in the vitamin D status, that is the 25-hydroxyvitamin D 3 (25(OH)D 3 ) serum levels, may be explained by skin colour and sun exposure behaviour. [10] A more detailed study with nearly 80 000 European-descent unrelated participants indicated that genes related to vitamin D endocrinology but not to skin colour are contributing to the vitamin D status. [11] On the basis of some 400 000 participants, the number of involved genomic regions increased to more than 100 but still skin colour was not found as a major trait. [12] In this review, we challenge the hypothesis of skin lightening due to vitamin D by summarizing recent archeogenomic data on changes of skin, hair and eye colour in European populations within the past 10 000 years. Moreover, we discuss changes in the allele frequency of single nucleotide polymorphisms (SNPs) in genes related to vitamin D metabolism and function by comparing present European populations with Africans and East Asians. In this context, we present our own hypotheses explaining differences in the vitamin D status and response index of individuals.

| MEL ANIN AND MODUL ATI ON OF ITS SYNTHE S IS
Since the beginning of life on Earth some 3.9 billion years ago [13] species had to protect themselves against the harmful consequences of excessive UV radiation being part of the solar spectrum. [14] Basically, all sun-exposed species synthesize molecules that efficiently absorb UV radiation, in order to protect sensitive macromolecules, such as DNA and proteins, from mutagenesis and destruction. In humans and many other species, melanin is used as the main UV-absorbing pigment, since it has a very high refractive index [15] as well as a very broad absorption spectrum ranging from gamma rays via UV and visible to infrared. Melanin is produced via the oxidation and polymerization of the aromatic amino acid tyrosine. [16] This process happens in melanosomes, which are specialized organelles within melanocytes that can be transferred to other cells, such as keratinocytes and hair follicles, in order to pigment them [17] (Figure 1, left). The process of melanogenesis can be induced by UV radiation, such as sun tanning of human skin. [18] The most common form of melanin, eumelanin, is brown to black, while the cysteine derivative pheomelanin is yellow to red. [19] Accordingly, skin and hair colour depend on the absolute amount of eumelanin and the addition of pheomelanin, such as in red hair.
The ratio between eumelanin and pheomelanin depends on the activity of the enzyme tyrosinase and the availability of cysteine. About 2 million years ago, when hominins lost most of their body hair, [20] in order to improve via more efficient heat dissipation their physical performance, [21] their initial pale skin developed an intensive pigmentation, in order to protect from sunburn and UV-induced cancer. [22,23] Variations in the pigmentation intensity of skin, eyes and hair can be explained by SNPs in gene encoding for key proteins in melanogenesis. [16] For the European populations, the most relevant SNPs are those related to the genes SLC24A5 (solute carrier family 24 member 5), SLC45A2 and OCA2 (OCA2 melanosomal transmembrane protein) [3,24] encoding for proteins functioning as a potassium-dependent sodium/calcium exchanger, ion transporter and pH regulator in melanosomes [16] (Figure 1, left). Since ion concentrations and pH values within melanosomes are critical for tyrosinase activity and efficient melanogenesis, loss of function at derived alleles, such as at SNP rs1426654 causing an alanine to threonine exchange at position 111 of the SLC24A5 protein (Ala111Thr) and at SNP rs16891982 leading to a leucine to phenylalanine exchange at position 374 (Leu374Phe) of the SLC45A2 protein, significantly reduce skin pigmentation. Moreover, SNP rs12913832 is located on a transcription factor binding site within an enhancer residing in an intron of the HERC2 gene but affecting the expression of the OCA2 gene. The homozygous G/G-derived allele of this SNP is associated with blue irises in Europeans. [25] Another critical gene is KITLG (KIT ligand), which encodes for a ligand produced in keratinocytes that stimulates the tyrosine kinase KIT-inducing melanogenesis. The derived allele of SNP rs12821256 controlling the expression of the KITLG gene significantly reduces hair pigmentation and leads to blond hair [26] (Figure 1, left).
In summary, a few variations in genes involved in melanogenesis explain well the differences in skin, hair and eye pigmentation, in particular in European populations.

| SYNTHE S IS OF VITAMIN D AND ITS ME TABOLITE S
The secosteroid vitamin D 3 is synthesized in human skin when the direct precursor of cholesterol, 7-dehydrocholesterol, is exposed to UV-B [27] (Figure 1, right). Although this reaction is non-enzymatic, the 7-dehydrocholesterol level, and by this amount of producible vitamin D 3 , is critically dependent on the activity of the enzyme 7-dehydrocholesterol reductase (DHCR7). [27] Individuals with less active DHCR7 have a higher amount of substrate in their skin, that is they can produce vitamin D 3 more efficiently even at less intense UV-B exposure. Moreover, vitamin D synthesis takes place predominantly in keratinocytes of the upper epidermis layer stratum spinosum [28] ( Figure 1, centre), where highest concentrations of 7-dehydrocholesterol are found. [29] In contrast, 70-80% of total melanin is found in the basal epidermis layer irrespectively of the skin phototype, [23] ensuring photodamage protection to the most cancer-susceptible skin layer. [28] Thus, a dark pigmentation of the skin, which is mainly due to eumelanin-laden melanosomes in keratinocytes of lower layers, seems not to inhibit significantly vitamin D 3 synthesis taking place in the upper layers. [6,30,31] Accordingly, dark skinned hunter-gatherer populations of Africa that live traditionally and expose most of their skin to intensive sun, such as the Tanzanian Hadza, have a rather high vitamin D status with average 25(OH)D 3 serum levels in the order of 110 nmol/L. [32] Like other dietary fats, vitamin D 3 obtained from diet or supplements is assembled in intestinal enterocytes into chylomicrons and passed through the lymphatic system to the bloodstream, where the vitamin is unloaded to the serum glycoprotein GC (GC vitamin D binding protein). [33] Similarly, vitamin D 3 obtained from endogenous synthesis is released from keratinocytes into the bloodstream and is as well transported via the GC protein. In the liver, vitamin D 3 is converted to 25(OH)D 3 , primarily by the enzyme CYP2R1 [34] (cytochrome P450 family 2 subfamily R member 1). Due to a halflife of some 3 weeks, 25(OH)D 3 is the most stable and abundant vitamin D metabolite and often used as a biomarker for the vitamin D status of a person. [35] Kidneys and a few other cell types, such as immune cells and keratinocytes, express the enzyme CYP27B1 mediating further hydroxylation of 25(OH)D 3 into 1α,25-dihydroxyvitamin D 3 (1,25(OH) 2 D 3 ). [36] The GC protein is responsible for the transport of all vitamin D metabolites but shows higher affinity for 25(OH)D 3 than for 1,25(OH) 2 D 3 . Accordingly, in serum vitamin D metabolites are mainly protein-bound and only a very small fraction circulates as "free" molecules. [36] 1,25(OH) 2 D 3 binds with high affinity (K D = 0.1 nM) to the transcription factor vitamin D receptor (VDR). [37] Thus, vitamin D regulates via 1,25(OH) 2 D 3 and the VDR the expression of hundreds of primary target genes. Examples of important vitamin D target genes are CYP24A1, CD14 (CD14 molecule) and CARD9 (caspase recruitment domain family member 9) [38] ( Figure 1, right).
Vitamin D endocrinology emerged some 550 million years ago in the boneless fish lamprey. [39] Thus, rather than bone metabolism, the prime function of vitamin D was the regulation of energy metabolism. [40] Since the immune system requires substantial amounts F I G U R E 1 Biology of skin pigmentation and vitamin D synthesis. Melanocytes residing in the basal layer of the epidermis provide keratinocytes and hair follicles (not shown) with melanosomes (left). In keratinocytes, melanosomes are taken up by phagocytosis and reorganized in the supranuclear area to form melanin caps. In the process of melanogenesis, melanin is produced in melanosomes under the control of the proteins KITLG, SLC24A5 and SLC45A2. Vitamin D 3 is produced when the cholesterol precursor 7-dehydrocholesterol is exposed to UV-B radiation (right). This non-enzymatic reaction produces previtamin D 3 (not shown) that rapidly isomerizes to vitamin D 3 . In serum, vitamin D 3 and its metabolites are transported bound to the GC protein. In the liver, the enzyme CYP2R1 converts vitamin D 3 into 25(OH)D 3 and in the kidneys (and other cell types) CYP27B1 produces 1,25(OH) 2 D 3 , which is the high-affinity ligand to the transcription factor VDR. Examples of VDR target genes are indicated of energy, [41] vitamin D and its receptor obtained via the control of immunometabolism a modulatory impact on immunity. [42] Since immune cells are the most rapidly growing cells of the human body, vitamin D extended its functional profile to the control of cellular proliferation, differentiation and apoptosis. [43] When around 400 million years ago bony fish left the calcium-rich ocean, at calcium-poor land they were facing the need to control the homeostasis of this ion being essential for a stable skeleton. [44] Thus, at that time vitamin D endocrinology gained the additional role of regulating calcium homeostasis. In this physiological function, vitamin D obtained key importance, [45] that is in contrast to its other tasks no other regulatory molecules are able to replace it. This explains why the prime phenotype of vitamin D deficiency is bone malformation, such as occurring in rickets. [46] Nevertheless, the rather ubiquitous expression of the VDR gene in more than half of the 400 human tissues and cell types indicates that vitamin D has a wider physiological role than the regulation of calcium homeostasis. [47] Taken together, vitamin D and its metabolites are important biomolecules, the sufficient synthesis of which is critical for a number of physiological functions of the human body.

| HIS TORY OF S KIN LI G HTENING IN EUROPE ANS
The genetic history of today's European populations is based on continuous migrations over the past 40 000 years. Homo sapiens arrived in Europe from Near East some 42 000 years ago. [48,49] Like in their African origin, these humans had dark skin but due to variations of their OCA2 gene (causing iris depigmentation) many of them had blue eyes [48,50] (Figure 2, left). At that time, still ancestral Neanderthal hominins lived in Europe, whom had arrived from Africa at least some 400 000 years earlier. [50][51][52] Through interbreeding with modern humans, the Neanderthalers were outnumbered [53] and disappeared a few thousand years after the arrival of the modern humans. [52] In net effect, the genome of today's Europeans contains in average 2.3% Neanderthal DNA. [54] The first European homo sapiens lived as hunter-gatherers in ice-free south-western Europe. [55] Following the retreating ice-sheet some 11-12 000 years ago, these hunter-gatherers started to colonize also northern Europe. [48] Interestingly, eastern and Scandinavian hunter-gatherers had light skin, [48] in contrast to Baltic hunter-gatherers who kept their dark skin only until 3800 years ago when farming was introduced in this region by the Bronze Age expansion of people of Russian steppe origin. [56,57] These migrants were associated with the Yamnaya culture and emerged out of a mixture between light-skinned Caucasus and eastern hunter-gatherers. [58,59] In addition, north-eastern speakers of Uralic languages, such as Estonians, Finns and most notably Saami, have a genetic component most similar to modern Siberians who arrived 3500 years ago. [56,60] Thus, Siberians and Yamnayas had a large impact on the phenotype of today's northern Europeans.
Although large genome-wide association studies highlighted skin colour as a highly polygenic trait, [20,61,62] evidence for selection of pigmentation variants in present-day Europeans was found only for a limited number of large-effect variants, the top two of which turned out to be the loci of the genes SLC24A5 and SLC45A2. [24,63] Thus, the light skin colour of today's Europeans is primarily based on non-synonymous SNPs of the genes SLC24A5 and SLC45A2, the geographical origin of which is western Asia and the Near East [48] (Figure 2).
Initial estimations based on haplotype diversity of microsatellites in modern humans suggested that derived alleles of these SNPs were common in Europeans already some 11 000 to 19 000 years ago. [64,65] However, recent archeogenomic analysis, that is whole genome sequencing of ancient DNA, allowed a more precise description of the evolution and timing of trait changes within European populations. [48,56,57,[65][66][67][68][69][70] The SNP rs1426654 within the SLC24A5 gene has the single largest effect on skin lightening of all gene variants identified to date. [66] It is located within a large (78 kb) haplotype block C11 shared by virtually all carriers of this allele, including ancient Scandinavian, eastern and Caucasus hunter-gatherers, [48] suggesting that this light skin variant derives from a single carrier who lived 22 000-28 000 years ago [48,71] in the Middle East. [72] Similarly, haplotype analysis of the SNP rs16891982 of the SLC45A2 gene concluded that this skin-lightening mutation occurred only once in the ancestry of Caucasians. [48,73] Migrations across the Caucasus and Eastern Europe would have brought both alleles to Scandinavia, in contrast to southern and central Europe, where they were introduced by farmers from western Anatolia [48] expanding 8500 to 5000 years ago [3] (Figure 2, right). This was the start of the Neolithic revolution in these regions, characterized by a more sedentary lifestyle and the domestication of certain animal and plant species. [3] The Anatolian farmers had rather short body stature and predominantly brown eyes, which explains the key anthropomorphic traits of today's southern Europeans, in contrast to Yamnayas, who had a high body stature and settled preferentially in northern Europe. [3,74] Moreover, these steppe pastoralists brought the horse, the wheel and Indo-European languages. [66,[74][75][76] Interestingly, ancient North Eurasian derived populations, such as eastern hunter-gatherers and Yamnayas, carried the blond hair allele rs12821256 of the KITLG gene to Europe. [66] Its first evidence was described in an 18 000 years old ancient North Eurasian west of Lake Baikal (Figure 2, right). It is important to note that the four major founding populations of Eurasians, which were farmers of the Fertile Crescent (including western Anatolia), farmers of Iran, hunter-gatherers of central and western Europe as well as of eastern Europe (Figure 2, right), genetically differed from each other probably as much as today's Europeans to East Asians. [77] Thus, the classic light phenotype of Europeans became frequent only within the past 5000 years [3,56,70] and owes its origin to migrants from Near East and western Asia. [48] Differences in the relative admixture of ancient hunter-gatherers, Anatolian farmers, Yamnaya pastoralists and Siberians explain the variations in skin and hair pigmentation, eye colour, body stature and many other traits of present Europeans. [60,74,78,79] The rapid increase in population size due to the Neolithic revolution, [64,80] such as the use of milk products as food source for adults and the rise of agriculture, [81] as well as the massive spread of Yamnaya pastoralists likely caused the rapid selective sweep in European populations towards light skin and hair. Interestingly, there had been an increase in the frequencies of the alleles of the genes SLC45A2 and SLC24A5 for light skin even after the admixture event indicating that selection for depigmentation variants was ongoing until rather recent times. [24,48,82] Phenotype predictions of ancient individuals have been recently based not only on the genes SLC24A5, SLC45A2 and OCA2 but on a whole array of marker pigmentation SNPs used in forensics [56,67,68,70,83] suggesting that light skin was for many thousand years not essential for survival at higher latitudes of Europe. [84] In addition, there is no mechanistical link between vitamin D and the expression of the main light skin alleles, such as primary vitamin D target genes carrying VDR binding sites in their regulatory regions. [85] In summary, there is no convincing evidence of a direct evolutionary effect of vitamin D on skin lightening of Europeans.

| G ENE TI C VARIANTS A SSO CIATED WITH VITAMIN D ME TABOLIS M AND S IG NALLING
A genome-wide association study (GWAS) based on nearly 80,000 European-descent subjects indicated that SNPs in the genes DHCR7, GC, CYP2R1 and CYP24A1, which are related to vitamin D metabolism and transport, may explain some of the genetics behind their vitamin D status. [11] For example, the derived allele of  hunter-gatherers [63] (Figure 2, left). Interestingly, DHCR7 is a risk gene for autoimmune diseases, type 1 diabetes [87] and multiple sclerosis, [88] which are both highly associated with vitamin D deficiency. Thus, evolution rather used the adaption of DHCR7 expression than that of skin-lightening genes as a mechanism of avoiding vitamin D deficiency and its medical consequences. [89] Like the minor allele of the DHCR7 SNP rs12785878, also the frequencies of the derived alleles of the vitamin D transport protein (GC, rs3755967), the 25(OH)D 3 synthesizing enzyme (CYP2R1, rs10741657), VDR (rs2228570 (commonly known as FokI polymorphism), rs1544410 (Bsm1) and rs731236 (Taq1)) and the VDR target genes CYP24A1 (rs17216707), CD14 (rs2569190) and CARD9 (rs4077515) increased up to 4-fold in European populations compared with the reference from Africa and even up to 8.6-fold in relation to data from Asian populations (Figure 3). In addition, also variants in the genes CYP11A1 and CYP27A1, which are encoding for enzymes involved in vitamin D metabolism, as well as the genes RXRA and RXRG, encoding for VDR partner proteins retinoid X receptors α and γ, had been shown to be enriched in European populations. [90] The mechanistic details how the minor alleles of these gene variants affect vitamin D metabolism and signalling are not fully understood, but in net effect the European populations seem to be more sensitive in their response to vitamin D. For example, the translation start site polymorphism of the VDR gene (rs2228570, FokI) results in a three amino acid shorter protein mediating higher transcriptional activity to its target genes. [91] Moreover, increased circulating levels of the fibroblast growth factor 23 proteins are associated with the minor allele of SNP rs17216707 close to the CYP24A1 gene, [92] a beneficial elevated pro-inflammatory response in the contextof sepsis is linked to the derived allele of SNP rs2569190 of the CD14 gene [93] and the minor allele of SNP rs4077515 of the CARD9 gene leads to a lower likelihood of progressive liver disease. [94] Thus, all thirteen example SNPs underwent positive selection within the European population.

VDR target genes
Taken together, within the past 40 000 years of the presence of homo sapiens in Europe populations were selected for genetic variants that allowed to cope via different mechanisms with lower levels of circulating 25(OH)D 3 .

| S TATUS AND RE S P ON S E INDE X OF VITAMIN D
The vitamin D status of an individual is most commonly deter- individuals over the year [7] (Figure 4, centre). The US Institute of Medicine [95] considers a 25(OH)D 3 serum level of 50 nmol/L sufficient, while the US Endocrine Society [96] suggests at least 75 nmol/L, that is there are different recommendations how much daily vitamin D 3 supplementation should be used, in order to reach which vitamin D status. Depending of these threshold levels, some 25%-50% of the world population can be considered as vitamin D deficient. [97] However, it can be questioned, whether a threshold level of the vitamin D status is the best reference for calculating the vitamin D needs of individuals. Based on the vitamin D response index concept [98] persons are distinguished into high, mid and low responders to vitamin D [99,100] (Figure 4, centre). Accordingly, they have a personal need for vitamin D 3 supplementation, which depends on their vitamin D status in relation to their response index. The index describes the molecular response to vitamin D 3 and its metabolites and is measured via the change in the expression of a few or all vitamin D target genes in a given tissue, such as peripheral blood mononuclear cells (PBMCs), after a significant change in the vitamin D status [38] ( Figure 4, centre). The response index is a variable genetic trait of each individual that is constant over time or may change only slowly, such as during the development of an age-related disease.
Importantly, high vitamin D responders benefit even from a low vitamin D status, that is they should tolerate conditions of low or no endogenous vitamin D 3 production during European winters.
Accordingly, these persons should be less affected by disorders, such as infections, [101] autoimmune diseases [102] and cancer, [103]  In summary, analysing the vitamin D response index may be the better parameter for describing vitamin D needs than referring to the vitamin D status.

| E VOLUTIONARY VITAMIN D ADAP TION HYP OTHE S E S
The central point of this review is the question how during the 40,000 year history of homo sapiens in Europe they adapted their vitamin D system of synthesis, transport, metabolism, receptor binding and target gene efficiency so that they did not suffer from malfunction due to deficiency. The first answer to this question was the hypothesis of Jablonski [8,9] that the need for efficient vi- tation. [65,82] Accordingly, the dark skin of some of today's Arctic Native people, such as Inuits, has been traditionally explained for their marine diet rich in vitamin D 3 . [8] However, nitrogen and carbon isotope analysis of northern and western Scandinavian hunter-gatherers revealed their had an extreme marine diet as well and yet developed light skin. [48] Thus, archeogenomic data as well as comparisons of today's populations did not provide any indication for an evolutionary pressure for light skin created by the need for vitamin D.
The second hypothesis is based on SNPs of the genes DHCR7, GC and CYP2R1 that show association with the vitamin D status of individuals [11] (Figure 4, left). The minor alleles of the SNPs of all three genes accumulated in today's European populations ( Figure 3). Importantly, archeogenomic analysis found the derived alleles of the DHCR7 gene rs12785878, rs7940244 and rs7944926 to be present already in western hunter-gatherers, highlighting rs7944926 as the 7th most significantly selected SNP in ancient Eurasians. [63] Moreover, in contrast to previous beliefs, dark pigmentation seems not to be a prominent inhibitor of vitamin D 3 synthesis, since the latter happens largely in upper layers of the skin that are less pigmented. [6,30] In addition, other SNPs affecting the vitamin D system, such as those discussed in Figure 3, may improve vitamin D's actions, so that individuals can better handle a lower vitamin D status at higher latitudes. Interestingly, based on a recent study with more than 400 000 Europeans [12] in total 143 SNPs were found to be associated with 25(OH)D 3 serum levels.
These genetic variants are located in 112 genomic regions carrying genes that are, for example involved in dermal development and integrity, such as FLG encoding for the cytoskeletal protein filaggrin. [104] Thus, future studies should provide further mechanistic explanations how these SNPs affect the vitamin D system and when they occurred during the evolution of homo sapiens in Europe.
The third hypothesis suggests that populations at higher latitudes managed to live with a lower vitamin D status, because their vitamin D system (enzymes, transporter and receptor) developed a higher affinity for vitamin D 3 and its metabolites, so that it can be activated by a lower vitamin D status (Figure 4, right). This concept has the advantage that individuals would not have to aim for a high vitamin D status. However, although similar affinity adaptions are known from different biological systems, such as sensing glucose or adapting to cold temperature, [105] there is so far no experimental proof for this hypothesis.
Since for high vitamin D responders a rather low vitamin D status is still sufficient to activate all vitamin D target genes, [38] there may have been a selection for high vitamin D responders in northern Europe. Accordingly, hypothesis 4 assumes that Yamnaya ancestry (including ancestry from Siberian hunter-gatherers) had a dominant effect on the phenotype of northern Europeans, [3,106] that is together with other traits Yamnayas also brought high vitamin D sensitivity to Europe, which may have originated predominantly from ancient North Eurasians (Figure 4, right). In a north-south gradient, Taken together, the concept of increasing frequency of SNPs affecting the vitamin D system may serve as a replacement of the skin-lightening hypothesis.

| CON CLUS ION
The 1000 Genomes Project [107] indicated that humans of European decent have a number of variants in their genome that have important effects on traits forming their phenotype. The master example of these traits is lactose tolerance, where due to SNPs within an enhancer regulating the LCT gene adults are able to digest milk sugar.
Furthermore, anthropomorphic traits, such as skin, hair and eye colour, are now well understood on the basis of SNPs affecting the genes SLC24A5, SLC45A2, OCA2, KITLG and others. The skin-lightening hypothesis of Jablonski [8] was formulated first 20 years ago and assumed that the need for efficient vitamin D synthesis acted as evolutionary driver for skin lightening in populations living at high latitudes. However, archeogenomic data indicated that Europeans got pale as predominant skin colour just some 5000 years ago through interbreeding with migrating populations originating from Anatolia and the Russian steppe. The latter populations likely accumulated the light skin alleles via genetic drift [24] rather than through an effect of vitamin D.
In this review, we provided alternative explanations for the adaption of the vitamin D system during the migration of homo sapiens to geographic regions of low or no endogenous vitamin D 3 production during winter. These concepts are based on SNPs affecting the function of genes involved in vitamin D synthesis, transport, metabolism, receptor binding and target gene efficiency. In conclusion, ancient European populations dealt with their "vitamin D problem" primarily via higher sensitivity for vitamin D 3 and its metabolites but less likely by getting lighter skin.

ACK N OWLED G EM ENTS
The authors would like to thank Jani Miettinen for guidance in R programming and Juha Kekäläinen and Teemu Kuulasmaa for bioinformatic support, as well as Mauro Scaravilli for valuable discussions.

CO N FLI C T S O F I NTE R E S T S
The authors have no conflicts of interest.