Merel A. Hamer

A Genome-Wide Association Study Identifies the Skin Color Genes IRF4, MC1R, ASIP, and BNC2 Influencing Facial Pigmented Spots

Facial pigmented spots are a common skin aging feature, but genetic predisposition has yet to be thoroughly investigated. We conducted a genome-wide association study for pigmented spots in 2,844 Dutch Europeans from the Rotterdam Study (mean age: 66.9±8.0 years; 47% male). Using semi-automated image analysis of high-resolution digital facial photographs, facial pigmented spots were quantified as the percentage of affected skin area (mean women: 2.0% ±0.9, men: 0.9% ±0.6). We identified genome-wide significant association with pigmented spots at three genetic loci: IRF4 (rs12203592, P=1.8 × 10(-27)), MC1R (compound heterozygosity score, P=2.3 × 10(-24)), and RALY/ASIP (rs6059655, P=1.9 × 10(-9)). In addition, after adjustment for the other three top-associated loci the BNC2 locus demonstrated significant association (rs62543565, P=2.3 × 10(-8)). The association signals observed at all four loci were successfully replicated (P<0.05) in an independent Dutch cohort (Leiden Longevity Study n=599). Although the four genes have previously been associated with skin color variation and skin cancer risk, all association signals remained highly significant (P<2 × 10(-8)) when conditioning the association analyses on skin color. We conclude that genetic variations in IRF4, MC1R, RALY/ASIP, and BNC2 contribute to the acquired amount of facial pigmented spots during aging, through pathways independent of the basal melanin production.

Prediction of male-pattern baldness from genotypes

The global demand for products that effectively prevent the development of male-pattern baldness (MPB) has drastically increased. However, there is currently no established genetic model for the estimation of MPB risk. We conducted a prediction analysis using single-nucleotide polymorphisms (SNPs) identified from previous GWASs of MPB in a total of 2725 German and Dutch males. A logistic regression model considering the genotypes of 25 SNPs from 12 genomic loci demonstrates that early-onset MPB risk is predictable at an accuracy level of 0.74 when 14 SNPs were included in the model, and measured using the area under the receiver-operating characteristic curves (AUC). Considering age as an additional predictor, the model can predict normal MPB status in middle-aged and elderly individuals at a slightly lower accuracy (AUC 0.69–0.71) when 6–11 SNPs were used. A variance partitioning analysis suggests that 55.8% of early-onset MPB genetic liability can be explained by common autosomal SNPs and 23.3% by X-chromosome SNPs. For normal MPB status in elderly individuals, the proportion of explainable variance is lower (42.4% for autosomal and 9.8% for X-chromosome SNPs). The gap between GWAS findings and the variance partitioning results could be explained by a large body of common DNA variants with small effects that will likely be identified in GWAS of increased sample sizes. Although the accuracy obtained here has not reached a clinically desired level, our model was highly informative for up to 19% of Europeans, thus may assist decision making on early MPB intervention actions and in forensic investigations.

Validation of image analysis techniques to measure skin aging features from facial photographs

Accurate measurement of the extent skin has aged is crucial for skin aging research. Image analysis offers a quick and consistent approach for quantifying skin aging features from photographs, but is prone to technical bias and requires proper validation.

Perceived skin colour seems a swift, valid and reliable measurement

DEAR EDITOR, Skin colour is an important trait in dermatological research because it modifies the risk of many skin diseases. In clinical practice dermatologists evaluate skin colour by a quick visual assessment of the sun-unexposed skin. Because visual assessment might be subjective, Fitzpatrick proposed a sun-sensitivity skin-type (Fitzpatrick skin type, FST) classification as a better alternative to quantify skin colour. However, this self-reported skin type might change in time and could be biased because light-skinned individuals tend to overestimate their skin colour. Alternatively, a spectrophotometer measures skin colour objectively, but might be influenced by the seasonal variations and colour inequalities. Visual assessment of perceived skin colour (PSC) allows physicians to combine different clues besides the colour (such as freckles), and to exclude tanning influences, into one skin-colour category. PSC has been used in several studies, but it has not yet been validated. We investigated the reliability of physicians’ perception of skin colour, and how PSC relates to the widely used FST and to unexposed skin colour quantified by a spectrophotometer. This study included adult volunteers visiting the dermatology department of Erasmus MC Rotterdam, the Netherlands, in December 2014 (exclusion criteria: albinism and erythroderma). A sample-size calculation showed 80% power to detect an inter-rater correlation of 0 8 with a maximum deviation of 0 1 when 80 participants were included. The local medical ethics committee approved this study and all participants signed informed consent. Skin colour was assessed using a six-point scale: 1 very white, 2 white, 3 white to olive, 4 light brown, 5 brown and 6 dark brown to black. Three physicians (L.C.J., M.A.H. and J.A.C.V.) independently graded the colour of sun-unexposed skin of the upper body (abdomen and inner upper arm) of each participant, without discussing their assessments. The agreement in grading was tested using the intraclass correlation coefficient (ICC; two-way-mixed, single measures). Subsequently, skin colour was measured as colour saturation at the inner upper arm using a spectrophotometer (CM-600d; Konica-Minolta, Osaka, Japan), and the FST was assessed by combining the answers to the questions: ‘What is your skin colour?’ (type I–IV white, type V brown, type VI black) and, ‘If white, how does your unprotected skin react to sunlight?’ (type I always burns and never tans, type II usually burns and afterwards lightly tans, type III sometimes burns and always tans, type IV never burns and tans deeply). Spearman’s q was used to test the correlation of the spectrophotometer skin colour with PSC and FST. We studied 117 individuals (mean age 45 7 18 6 years, 31% men) of different ancestry. The mean PSC of the three graders included 28 individuals (24%) as very white, 58 (50%) white, 14 (12%) white to olive, six (5%) light brown, 11 (9%) brown and none dark brown to black. The three physicians showed an excellent agreement in grading (ICCabsolute agreement = 0 90). In 114 individuals (three answered ‘unknown’) the self-reported FST included 14 individuals (12%) of type I, 46 (40%) type II, 28 (25%) type III, four (4%) type IV, 22 (19%) type V and none type VI. Colour saturation of unexposed skin ranged from 0 17 (very white) to 0 45 (dark) and was highly correlated with mean PSC (q = 0 82, Fig. 1a, and less so with FST (q = 0 63, Fig. 1b). The FST was generally higher than the mean PSC (mean difference = 0 5, P < 0 001; Fig. 2), but consistency between PSC and FST was still high (ICCconsistency = 0 82). Five individuals were white or white to olive according to the physicians but judged themselves as brown. However, none of the individuals with light brown or brown PSC judged their own skin as white (Fig. 2). After excluding these five individuals, the consistency between PSC and FST increased to an ICC of 0 86. Comparing PSC in younger (≤ 45 years, n = 60) vs. older (> 45 years, n = 57) individuals, a similar agreement between the three physicians was seen (ICCabsolute agreement young = 0 89, ICC old = 0 90). However, the correlation between the mean PSC and the FST was higher in the younger individuals (ICCconsistency young = 0 89, ICC old = 0 76). To assess differences in sun reaction between the present and the teenage years, we asked the volunteers whether their answers to the FST questions would have been different for their teens. Seven individuals reported a higher FST in the past and 11 a lower FST. Most individuals with a higher FST in the past were aged < 45 years (six of seven), and most individuals with a lower FST in the past would have had an FST of I in their teens (seven of 11). Our results show that physician assessment of skin colour is reliable, because minimal interobserver variability was observed. The self-reported FST is higher than the PSC, but largely consistent with it. However, the spectrophotometer colour saturation correlated less with the FST than with the PSC. This suggests that PSC is a mix between true colour (spectrophotometer) and sun sensitivity (FST). Moreover, our data confirm that light-skinned individuals tend to overestimate

Genome-wide association meta-analysis of individuals of European ancestry identifies new loci explaining a substantial fraction of hair color variation and heritability

Hair color is one of the most recognizable visual traits in European populations and is under strong genetic control. Here we report the results of a genome-wide association study meta-analysis of almost 300,000 participants of European descent. We identified 123 autosomal and one X-chromosome loci significantly associated with hair color; all but 13 are novel. Collectively, single-nucleotide polymorphisms associated with hair color within these loci explain 34.6% of red hair, 24.8% of blond hair, and 26.1% of black hair heritability in the study populations. These results confirm the polygenic nature of complex phenotypes and improve our understanding of melanin pigment metabolism in humans.Genome-wide meta-analysis identifies >100 loci associated with hair color variation in humans of European ancestry. These loci explain a large portion of the heritability of this trait & provide insights into pathways regulating hair pigmentation.

Genome-wide association study in 176,678 Europeans reveals genetic loci for tanning response to sun exposure

The skin’s tendency to sunburn rather than tan is a major risk factor for skin cancer. Here we report a large genome-wide association study of ease of skin tanning in 176,678 subjects of European ancestry. We identify significant association with tanning ability at 20 loci. We confirm previously identified associations at six of these loci, and report 14 novel loci, of which ten have never been associated with pigmentation-related phenotypes. Our results also suggest that variants at the AHR/AGR3 locus, previously associated with cutaneous malignant melanoma the underlying mechanism of which is poorly understood, might act on disease risk through modulation of tanning ability.The skin’s tanning response to sun exposure shows great interindividual variability. Here, Visconti et al. perform a genome-wide association study for ease of skin tanning and identify 20 genetic loci, ten of which had not previously been associated with pigmentation-related traits.

A healthy diet in women is associated with less facial wrinkles in a large Dutch population-based cohort

Background: Little is known about the effects of different dietary patterns on facial wrinkling. Objective: We aimed to investigate the association between diet and facial wrinkles in a population‐based cohort of 2753 elderly participants of the Rotterdam study. Methods: Wrinkles were measured in facial photographs by digitally quantifying the area wrinkles occupied as a percentage of total skin area. Diet was assessed by the Food Frequency Questionnaire. Adherence to the Dutch Healthy Diet Index (DHDI) was calculated. In addition, we used principal component analysis (PCA) to extract relevant food patterns in men and women separately. All food patterns and the DHDI were analyzed for an association with wrinkle severity using multivariable linear regression. Results: Better adherence to the Dutch guidelines was significantly associated with less wrinkles among women but not in men. In women, a red meat and snack–dominant PCA pattern was associated with more facial wrinkles, whereas a fruit‐dominant PCA pattern was associated with fewer wrinkles. Limitations: Due to the cross‐sectional design of our study, causation could not be proven. Other health‐conscious behaviors of study participants could have influenced the results. Conclusion: Dietary habits are associated with facial wrinkling in women. Global disease prevention strategies might benefit from emphasizing that a healthy diet is also linked to less facial wrinkling.

Dermatological screening of a middle‐aged and elderly population: the Rotterdam Study

DEAR EDITOR, The prevalence of several skin diseases increases with age, partly due to age-related physiopathological alterations. The global burden of disease project (GBD) presented a comprehensive overview of the burden due to skin diseases across different age groups. However, the point prevalences of many common skin diseases, especially in older people and noninstitutionalized settings, are still unknown. In this study, we estimated the point prevalence and ageand sex-adjusted standardized prevalence rates of common inflammatory and (pre)malignant skin diseases in a cross-sectional study in a middle-aged to elderly population (Appendix S1; see Supporting Information). In 2010, full body skin examinations (FBSEs) were introduced in the Rotterdam Study, a prospective populationbased cohort study of people aged ≥ 45 years. Since the start of study, in 1989, all citizens of Ommoord (a district of Rotterdam) were invited to participate, with an overall response of 72% during three invitation cycles. The study design and follow-up are detailed in Hofman et al. At FBSE, all participants were aged ≥ 50 years. The FBSEs were carried out by dermatology-trained physicians who checked for common skin diseases, including (pre)malignancies, eczema, psoriasis, seborrhoeic dermatitis and clinical signs of venous insufficiency (Appendix S2; see Supporting Information). Ageand sex-adjusted standardized prevalence rates (PR) were calculated per 100 000 persons aged ≥ 50 years and standardized to the Revised European Standard Population (Appendix S3; see Supporting Information). In total, 5365 participants (median age 67 2 years; interquartile range 61 6–74 8) were examined. The ageand sex-adjusted prevalence of skin diseases are presented in Table 1. Xerosis cutis was present in 66% of participants. Premalignant and malignant skin diseases were very common, comprising 48% of all identified skin diseases (Fig. S1; see Supporting Information). Actinic keratosis (AK) was found in 1399 (26 1%) participants; 234 (4 4%) showed one or more cutaneous malignancies. Basal cell carcinomas (BCCs) were most common but melanomas, squamous cell carcinomas (SCCs) and mycosis fungoides were also diagnosed (Table 1). Seborrhoeic dermatitis was the most common nonmalignant disorder, with a standardized prevalence rate of 17 685 per 100 000 men and 9588 per 100 000 women. Skin disorders are often not identified, or neglected. In this cohort, we found a high prevalence of (pre)malignant disorders. About a quarter of participants were diagnosed with AK, a potential precursor of SCC. The point prevalence of SCCs is lower than expected compared with melanomas. This could be because melanomas more often develop on less visible body sites and are often subclinical, contrasting with SCC. Therefore, melanomas are more likely to be detected during screening. About one in 25 participants was diagnosed with at least one cutaneous malignancy. This high prevalence suggests that primary care practitioners should be more alert to detect suspicious skin lesions in the middle-aged and elderly population. The U.S. Preventive Services Task Force recently concluded that there is insufficient evidence for skin cancer screening in asymptomatic adults. However, screening via self-examination, case finding by physicians or screening of high-risk subgroups were considered. The German skin screening programme showed that screening of all adults had a limited impact on melanoma-related mortality. However, the lower number needed to treat in elderly patients, and the higher skin cancer prevalence in men, suggest that elderly men are potentially a target group for skin screening. Also, skin cancer awareness campaigns should not be restricted to melanomas but should include keratinocytic cancers given their prevalence and associated disease burden. Although most prevalent diseases identified in this study (e.g. seborrhoeic dermatitis, psoriasis, venous disease and eczema) have low mortality rates, they can severely affect quality of life. Simple measures such as education on general skin care (e.g. using mild cleansers instead of water and soap in xerosis cutis) might be a fast and easy way to manage some of these conditions. Also, treatment of varicose veins in patients with oedema or skin changes might reduce symptoms and prevent the development of venous leg ulcers. Most skin diseases may only cause a relatively small burden on the individual patient level, but due to the high prevalence, the societal burden is substantial. As stated in a recent report from the World Health Organization on ageing and health, changes in health care will be needed to adapt to the continuing increase in life expectancy and the subsequent higher proportion of the elderly population living with comorbidity, including skin diseases. The high prevalence of skin malignancies found in this population-based screening indicate that the total burden of skin diseases is beyond the recent estimates presented by the GBD project. Therefore, the diagnosis, prevention and treatment of skin disease should get sufficient priority in general healthcare education and policies.

Genome-wide association studies and CRISPR/Cas9-mediated gene editing identify regulatory variants influencing eyebrow thickness in humans

Hair plays an important role in primates and is clearly subject to adaptive selection. While humans have lost most facial hair, eyebrows are a notable exception. Eyebrow thickness is heritable and widely believed to be subject to sexual selection. Nevertheless, few genomic studies have explored its genetic basis. Here, we performed a genome-wide scan for eyebrow thickness in 2961 Han Chinese. We identified two new loci of genome-wide significance, at 3q26.33 near SOX2 (rs1345417: P = 6.51×10−10) and at 5q13.2 near FOXD1 (rs12651896: P = 1.73×10−8). We further replicated our findings in the Uyghurs, a population from China characterized by East Asian-European admixture (N = 721), the CANDELA cohort from five Latin American countries (N = 2301), and the Rotterdam Study cohort of Dutch Europeans (N = 4411). A meta-analysis combining the full GWAS results from the three cohorts of full or partial Asian descent (Han Chinese, Uyghur and Latin Americans, N = 5983) highlighted a third signal of genome-wide significance at 2q12.3 (rs1866188: P = 5.81×10−11) near EDAR. We performed fine-mapping and prioritized four variants for further experimental verification. CRISPR/Cas9-mediated gene editing provided evidence that rs1345417 and rs12651896 affect the transcriptional activity of the nearby SOX2 and FOXD1 genes, which are both involved in hair development. Finally, suitable statistical analyses revealed that none of the associated variants showed clear signals of selection in any of the populations tested. Contrary to popular speculation, we found no evidence that eyebrow thickness is subject to strong selective pressure.