GWAS OF SKIN AGING IN KOREAN POPULATION 75 therefore that ADSS2 may have an effect on wrinkles. In addition, rs7042102, which is another SNP identifi ed in this study, is a variant located downstream of the SPTLC1, for which its differences in expression, based on genotype in skin tissues, were confi rmed in the eQTL database. Previous reports suggest that SPTLC1 may be a causal gene that affects psoriatic lesions (41) and may also affect skin phenotype by functioning as a regulator of skin fi broblasts (42). SPTLC1 is involved in the synthesis of sphingolipids, which are sub- stances involved in maintaining the skin’s barrier and moisture (43) and exerts skin im- provement effects through skin moisturization and maintenance (44,45). From this perspective, SPTLC1 may affect skin aging and the generation of wrinkles, if it causes ab- normalities in skin maintenance. MOISTURE CONTENT The moisture content of the skin is closely related to the contraction of the stratum cor- neum. This outermost layer of the epidermis is important for the skin’s barrier function because it has various proteolytic effects related to skin changes as well as to exfoliation, lipid synthesis, and infl ammatory reactions (46). Dry skin and frequent infl ammatory reactions in the stratum corneum may lead to diseases, such as atopic dermatitis and pso- riasis. Internal factors involved in modulating the important characteristic of the skin moisture content include lipids of the stratum corneum, natural moisturizing factors, and external factors (47). However, there are still individual differences regarding changes in skin hydration, for which genetic variants have been highlighted as a likely explanatory factor and are currently being studied (15,48). Our results identifi ed six SNPs, rs9873353, rs34567709, rs1362404, rs7853290, rs143938096, and rs12955989, which were associ- ated with the skin moisture content. The SNP identifi ed as the most signifi cant marker of skin moisture in our results was rs9873353. This locus has several pseudogenes in the surrounding area, but no functional genes have been identifi ed. However, there are SNPs that are considered to have other potential functions. Potential functional effects are sug- gested for rs7853290 and rs143938096, which are associated with the CEMIP2 (TMEM2) and CTSH genes, respectively. Both of these genes can affect skin changes, especially CEMIP2, which regulates hyaluronan, a key component of physical functions of the skin, such as moisturization and lubrication (49). Regarding the differences in expression ac- cording to the SNPs in the eQTL database, although only adipose tissue eQTLs are in- cluded among the single-tissue eQTLs, the available multiple-tissue eQTLs confi rmed that altered expression of CEMIP2, based on the rs7853290 genotype, also appears in the skin tissue. Therefore, it is possible that this SNP is capable of providing moisture and functionally controlling the barrier, which may explain our results. PIGMENTATION Changes in pigmentation are largely visible and can be found in the hair and eyes, in addi- tion to local phenotypic variations in the skin. Pigmentation is the most studied topic among the various phenotypic variations of the skin and is caused by differences in melanin, where differences between individuals are caused by changes in the type, amount, or other parameters of melanin (13). These individual melanin differences are due to genetic

JOURNAL OF COSMETIC SCIENCE 76 characteristics, and many previous studies have reported melanin-related genes and gene mutations. In addition, differences and correlations between genes involved in pigmenta- tion have been reported (17). In the present study, we identifi ed new candidate loci, TSN1, RUFY4, NCLN, and CDC42BPA, as well as the previously reported OCA2 gene, which is associated with various skin pigment-related diseases, such as albinism (13) and melanoma (50). In particular, an amino acid substitution (His615Arg) found in Asian populations has been reported to present at a high prevalence in cases of skin whitening and pigment changes (13). According to our results, rs74653330 is a missense (Ala481Thr) SNP that is re-associated with pigmentation in the East Asian population (31,32). Although the other genes are not functionally related to genes that directly affect melanin or skin tone, TSN1 encodes collagen-associated intercellular adhesion proteins (51), and RUFY4 causes skin infl ammatory responses (52). NCLN and CDC42BPA are not reported to be associated with pigmentation of the skin. However, our single-tissue eQTLs show differences in genotype expression in skin tissue. This suggests that our analysis and results are reproducible com- pared with previous studies, which indicates accuracy of our correlation model with skin measurements and suggests that we used effective analytical methods. OIL CONTENT Oil forms a protective fi lm on the skin to prevent various problems, such as water evapo- ration and infection. However, when various factors result in abnormal control of oil content, problems such as infl ammation, acne, and sebaceous gland hypersecretion may occur (53). In general, the amount of oil in the skin decreases with aging, which causes the skin to lose elasticity, moisture content, and dryness (54). In addition, recent studies have reported that skin diseases can be caused by excessive oil and that air pollutants may be absorbed by skin oil that can adversely affect skin health. As such, oil acts as a source of protection for the skin and prevents skin aging, in addition to preserving the moisture content (55). Therefore, rs308971, rs9577919, rs8107564, and rs6490805, which are associated with skin oil, may cause skin changes. The most signifi cant SNP in our analy- sis was rs908971, which is located at intron 1 of the SYN2 gene. Traditionally, SYN2 is associated with neurotransmitters (56). Recently, the association between the lipid me- tabolism and the SYN2 gene has been reported (57). Based on this, SYN2 is involved in the synthesis of synaptic vesicles and is associated with the formation of lipids and apoli- poproteins (58), which may affect the formation of lipids in the body. There have also been reports of associations of SNPs with triglycerides (57), which may lead to associa- tions in the lipid metabolism and further suggests that it is a process that can have a measurable effect on the skin. Furthermore, rs9577919 is located in intron 1 of GAS6, which mediates the infl ammatory response and affects the development of psoriasis (59). In addition, rs8107564 is located downstream of the INSR gene. Insulin receptors en- coded by INSR are involved in various mechanisms, including the regulation of infl am- mation (60), cancer development (61), and keratinocyte proliferation (62). Although rs6490805 shows a correlation with skin oil content, there are no relevant research re- ports or reported functional correlations. However, because single-tissue eQTLs show differences in expression levels in skin tissue according to genotype, potential oil abnor- malities can be inferred from observed skin changes. rs7334780, the SNP with the stron- gest signifi cance in sensitivity of the skin, is located in the intergenic region, and there is no gene present nearby therefore, further studies are needed to determine its function.