Genetic variants associated with skin photosensitivity in a southern European population from Spain

Recent GWAS studies, mostly performed in populations of North European origin, have identified the genetic loci associated with pigmentation, sun sensitivity, freckling and skin cancer susceptibility. Here, we aimed at addressing the genetic determinants of sunlight sensitivity in Spain, a southern European population.


| INTRODUC TI ON
In recent years, there have been important advances in the elucidation of the genetic determinants of human skin colour, skin sensitivity to sunlight and tanning ability. [1][2][3][4][5][6][7][8] Both cutaneous pigmentation and the skin's response to sunshine are extremely relevant to individual susceptibility to developing skin cancers. [9][10][11] Although skin sensitivity to ultraviolet (UV) radiation and the ability to tan are genetically determined, exposure to ambient sunlight is necessary for UV light to reveal its harmful effects on human skin. 12 People with high sensitivity to sunlight have a tendency to develop sunburns and frequently lack the ability to acquire a tan following exposure to UV light. These individuals with diminished sun response tend to display cutaneous hyperpigmented spots such as nevi or ephelides more commonly than people who are able to tan easily. 13,14 Skin sensitivity to sunlight, together with light skin (and light hair and eye colour) and incapacity to tan are all risk phenotypes conferring susceptibility to develop skin cancers-including basal cell carcinoma, squamous cell carcinoma and the more dangerous melanoma.
In fact, several genes involved in skin pigmentation-such as MC1R, IRF4, SLC45A2, and so on-are known to behave as low-penetrance genes for melanoma and other skin cancers. 15 Most of the studies aimed at addressing the genetic determinants of skin photosensitivity and/or tanning ability have been performed in populations of North European origin, both in northern Europe and in North America. [3][4][5][16][17][18][19] However, there is a shortage of studies carried out in darker-skinned populations such as southern Europe or the Mediterranean. [20][21][22][23][24][25] These more southerly populations present lower percentages of highly sun-sensitive individuals and higher percentages of individuals that are capable of tanning with ease. Accordingly, allele frequencies of the variants responsible for pigmentation in southern Europe may differ from those further north. 26 Due to higher UV radiation in South Europe, mainly during the summer months, evolution has probably influenced the distribution of allele frequencies of several human pigmentation genes in a different way to North Europe. For example, the F374 allele in SLC45A2, responsible for darker skin, is significantly more frequent in Spain than in more northerly European populations. 27,28 In this study, we set out to elucidate the genetic determinants of sun photosensitivity in Spain, a population of Mediterranean origin, by means of testing variation in the genes previously associated with sun sensitivity. We have performed SNP genotyping in 8 pigmentation-related genes (HERC2, SLC45A2, SLC24A4, TYR, ASIP, OCA2, BNC2 and IRF4) as well as complete gene sequencing of the MC1R gene in 456 individuals of Spanish origin. The correlation of the genetic variants with skin photosensitivity, and the interaction among all the associated genes, is analysed.

| Study subjects and data collection
A total of 456 cancer-free volunteers (272 controls and 184 cases) were included in this case-control study. Individuals with low sensitivity to sunlight, who never or rarely burn and always tan after sun exposure (Fitzpatrick skin phototypes III-IV-V), were included in the control group. Conversely, individuals presenting high sensitivity to sunlight, who always or often burn and never or seldom tan after sun exposure (phototypes I-II), were defined as cases. All individuals were Europeans of Spanish origin. Written informed consent was provided by all participants, and the study was approved by the Ethics Committee of the Jaume I University of Castellon (Castellon, Spain).
A standardised questionnaire was used to collect information on sex, age, place of birth, pigmentation traits (skin, hair and eye colour), freckling degree, history of childhood sunburns, skin sensitivity to sunlight (tanning ability vs tendency to burn), Fitzpatrick's skin type classification and sun exposure habits. To avoid misclassification, all questionnaires were completed by a professional dermatologist.

| DNA isolation, SNP selection and genotyping assays
Genomic DNA was isolated from saliva samples using QIAamp DNA Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's protocol.
We selected genes previously associated with skin sensitivity to sunlight and tanning ability. 4,5,16,18,19 Nine SNPs located in 8 pigmentation-related genes were finally genotyped: rs4911442 Unsuccessful genotyping rate was lower than 5% in all SNPs analysed.
The MC1R gene was studied by direct genetic sequencing with the Sanger method, as previously described. 29 A sample with known MC1R genotype per 96-well plate was added for quality control.

| Statistical analysis
Statistical analyses were performed using the R software (http:// www.R-project.org). All analyses were two-sided, and a significance level of 0.05 was considered for rejection of the null hypothesis.
Correction for multiple hypothesis testing was carried out using the Bonferroni method (P-value < 4.54 × 10 −3 = 0.05/11). Unknown and missing values were excluded at each specific analysis.
Descriptive statistics (means, standard deviations, sample size and percentages) were used to report demographic characteristics, sun sensitivity traits and sun exposure habits of participants (Table 1). For categorical variables, Fisher's exact test was used to evaluate differences between cases and controls. Mann-Whitney U test was used to account for differences in age between cases and controls, as age was not normally distributed according to the Shapiro-Wilk test.
For genetic association analyses, nonsynonymous MC1R mutations were classified as previously described. 25 Individuals were then classified according to the number of R and/or r alleles carrying.
The individual effect of each MC1R variant was evaluated.
Association between all variants studied and sunlight sensitivity was assessed according to the additive model of inheritance via logistic regression adjusted by sex, estimating odds ratios (ORs) per variant carried, their corresponding 95% confidence intervals (CIs) and associated P-values. All genetic analyses were performed estimating the effect of the minor allele in the Spanish population.
Fisher's exact test was used to study combinations of protective and risk genotypes in different genes for skin photosensitivity. Interactions among sunlight sensitivity-associated genes were tested using the multifactor dimensionality reduction (MDR) method (MDR software v.1.2). 30 The MDR method tests all potential locuslocus combinations to create a model that sorts cases and controls with the lowest possible classification error. That procedure is repeated 10 times to obtain both the cross-validation (CV) consistency, the number of times that a particular model is chosen as the best one, and the balanced accuracy (BA), the proportion of occurrences correctly classified using the model.

| RE SULTS
Evidence of association with skin sensitivity to sunlight for the genetic variants studied was assessed ( Table 2). All associations observed were significant after Bonferroni correction. R variants of the TA B L E 1 Mean ages and distribution of sex, pigmentation and sun sensitivity traits, and sun exposure habits among Spanish cases and controls MC1R gene were the most strongly associated with sunlight sensitivity (P-value = 2.00 × 10 −6 ), being the OR estimated per variant carried of 2.55. The rs16891982 polymorphism in SLC45A2 presented the second strongest association with sunlight sensitivity (P-value = 6.79 × 10 −5 ), observing a protective effect of the G minor allele regarding tolerance to sunlight (OR = 0.57). Significant associations were also observed for IRF4 rs12203592 (P-value = 9.35 × 10 −5 ) and HERC2 rs12913832 (P-value = 4.75 × 10 −4 Moreover, interaction effects between SLC45A2 rs16891982 and the other genetic variants influencing sunlight sensitivity were assessed. A significantly increased percentage of HERC2 rs12913832*C carriers (P-value = 9.97 × 10 −3 ), IRF4 rs12203592*T carriers (P-value = 3.62 × 10 −3 ) or MC1R R carriers (P-value = 1.31 × 10 −3 ) were detected in highly photosensitive individuals harbouring at least one copy of the rare protective allele in SLC45A2 rs16891982 (Figure 1). According to the MDR method, sunlight sensitivity was best explained assuming interaction between polymorphisms in all four sun sensitivity-associated genes (BA = 0.679, CV consistency = 10/10, P-value < 0.0001). Significant redundant interaction was observed between MC1R R variants, IRF4 rs12203592 and SLC45A2 rs16891982 (indicated by black lines in Figure 2).

| D ISCUSS I ON
In the current study, nine pigmentation-related genes were exam- The MC1R gene has been previously described as one of the major contributors to the diversity of human pigmentation. 32 The melanocortin-1 receptor is a G-protein-coupled receptor expressed on the melanocyte's membrane surface that is involved in the regulation of melanocyte proliferation and function. 33  However, our findings show an absence of a significant effect of the r variants on sunlight sensitivity. These results are consistent with those reported by Latreille et al, 38 which found that r variants were significantly associated with freckling (usually a signal of high sensitivity to sunlight) but not with sunburn event frequency nor suntan intensity.
The SLC45A2 gene, which encodes a membrane-associated transporter protein involved in arranging melanogenic enzymes during melanosome maturation, has been previously defined as a human pigmentation-related and melanoma-susceptibility gene. Genetic variants of the SLC45A2 gene, located in both the promoter region and the coding region, have been significantly associated with dark pigmentation and with protection against melanoma in southern European populations. [40][41][42] In this study, we focused on the rs16891982 polymorphism in SLC45A2, which results in a nonsynonymous change (F374L IRF4 has also been associated with pigmentation characteristics in several studies. 4,44 This interferon regulatory factor cooperates with MITF to activate the expression of tyrosinase in melanocytes, a function that seems to be impaired in carriers of the rs12203592*T-derived minor allele. 45 From the results of our study, we can confirm that this gene is important in the skin sensitivity to sunlight in the Spanish population, as it is highly associated with this phenotype. The derived T allele has previously been associated with tanning ability, 5 and melanoma susceptibility. 21 As human pigmentation has been demonstrated to be a highly other sun sensitivity-associated genes. The MDR interaction analysis found that sunlight sensitivity is better explained by taking into account the effect of MC1R, IRF4, SLC45A2 and HERC2. As seen in Figure 1 and Table 3, risk alleles in MC1R, IRF4 and HERC2 are able to mitigate the protective effect of carrying at least one SLC45A2 rs16891982*C allele.
It is important to keep in mind that there is still considerable ambiguity in the definition of the term "skin sensitivity to sunlight" or "skin photosensitivity," being difficult to perform a comprehensive comparison with previous published studies. Data standardisation should be addressed in future investigations by employing consistent definitions to allow for meta-analysis studies. Another possible caveat is that this study may indeed be ignoring the effect of other unknown genetic variants on skin photosensitivity, as we only focused on pigmentation-related genes previously associated-although through several potent GWAS-with skin sensitivity to sunlight and tanning ability.
In short, this study shows that the MC1R gene is the major contributor to skin photosensitivity in the Spanish population, specifically by way of its R variants. Besides, genetic variants in SLC45A2, HERC2 and IRF4 are also significantly associated with sunlight sensitivity in individuals of Spanish origin, independently of sex. However, large-scale replication collecting phenotypic traits related to photosensitivity in different southern European populations is necessary to confirm our findings.

ACK N OWLED G EM ENTS
We are extremely grateful to all volunteers for taking part in this study, as well as to all medical specialists for supervising phenotype collec-

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