- Research article
- Open Access
Genome-wide association study identifies novel loci associated with skin autofluorescence in individuals without diabetes
BMC Genomics volume 23, Article number: 840 (2022)
Skin autofluorescence (SAF) is a non-invasive measure reflecting accumulation of advanced glycation endproducts (AGEs) in the skin. Higher SAF levels are associated with an increased risk of developing type 2 diabetes and cardiovascular disease. An earlier genome-wide association study (GWAS) revealed a strong association between NAT2 variants and SAF. The aim of this study was to calculate SAF heritability and to identify additional genetic variants associated with SAF through genome-wide association studies (GWAS).
In 27,534 participants without diabetes the heritability estimate of lnSAF was 33% ± 2.0% (SE) in a model adjusted for covariates. In meta-GWAS for lnSAF five SNPs, on chromosomes 8, 11, 15 and 16 were associated with lnSAF (P < 5 × 10–8): 1. rs2846707 (Chr11:102,576,358,C > T), which results in a Met30Val missense variant in MMP27 exon 1 (NM_022122.3); 2. rs2470893 (Chr15:75,019,449,C > T), in intergenic region between CYP1A1 and CYP1A2; with attenuation of the SNP-effect when coffee consumption was included as a covariate; 3. rs12931267 (Chr16:89,818,732,C > G) in intron 30 of FANCA and near MC1R; and following conditional analysis 4. rs3764257 (Chr16:89,800,887,C > G) an intronic variant in ZNF276, 17.8 kb upstream from rs12931267; finally, 30 kb downstream from NAT2 5. rs576201050 (Chr8:18,288,053,G > A).
This large meta-GWAS revealed five SNPs at four loci associated with SAF in the non-diabetes population. Further unravelling of the genetic architecture of SAF will help in improving its utility as a tool for screening and early detection of diseases and disease complications.
Advanced glycation endproducts (AGEs) are irreversibly glycated proteins that result from complex reactions between sugars and proteins . AGEs are formed and accumulate in the body during life and accumulation is amplified under conditions of glycaemic stress .
The accumulation of AGEs can be evaluated non-invasively by measuring skin autofluorescence (SAF) with the AGE Reader [3, 4]. The method is based on the fact that AGEs present in dermal tissue exhibit fluorescence characteristics, they absorb the illuminated light at a particular wavelength and subsequently emit light of longer wavelength. Previous studies have shown predictive value of SAF in individuals both with- and without diabetes. In non-diabetic individuals SAF is a predictor for the development of type 2 diabetes, cardiovascular events and mortality and in individuals with both type 1 and type 2 diabetes SAF predicts diabetes-related complications and mortality [5,6,7,8,9,10,11]. A recent study reported that SAF may also predict the occurrence of cancer in individuals with type 2 diabetes, thereby suggesting the possibility that SAF may be used in selection of type 2 diabetes individuals for cancer screening .
In addition to SAF, measurements performed with the AGE Reader (Diagnoptics Technologies, Groningen, the Netherlands) yield a value for skin reflectance (SR). SR is a proxy for skin colour and pigmentary phenotype, light skin reflects more light and therefore individuals with light skin have higher SR values. Pigmentary characteristics are associated with susceptibility for melanoma .
Variation in SAF is dependent on environmental influences such as smoking and coffee consumption and is, in part, explained by genetic variation. Heritability of serum AGEs were estimated to be 63–74% [14, 15]. Previous genome-wide association studies (GWAS) have revealed two genetic loci associated with SAF [16, 17]. The first study identified a SNP in NAT2 which is known to tag the NAT2 acetylator phenotype, and explains around 3% of the SAF variance . The second study, in individuals with type 1 diabetes, demonstrated a locus on chromosome 1 which was associated with skin intrinsic fluorescence, a parameter comparable to SAF . Most of the genetic basis of SAF remains undiscovered.
The aim of the current study was to calculate SAF heritability and to identify additional genetic loci associated with SAF by performing GWAS in non-diabetic individuals of the population-based Lifelines Cohort Study.
First, we calculated genetic heritability of SAF. Subsequently, we performed a meta-GWAS of skin autofluorescence (SAF) on genotyped and imputed SNPs from all non-diabetic individuals in whom both results of measurement of SAF and genetic data were available (n = 27,534). Meta-analysis of the data was performed because genotyping of two subgroups from the Lifelines cohort was performed using different arrays, i.e. the Illumina CytoSNP 12v2 chip and the Illumina GSA chip. Genome wide significantly associated SNPs were reported. For the SNPs on chromosome 16, in a region known for its association with skin colour, we examined Reflectance-SNP interaction. In addition, for the SNP on chromosome 15, which showed a clear attenuation of the P-value when coffee consumption was added as a covariate to the model, we performed further investigation. First, we examined SNP-covariate interaction and in a post-hoc analysis we examined SNPs associated with coffee consumption in previous genetic studies for association with SAF. For all genome wide significantly associated loci single-tissue expression quantitative trait loci (eQTLs) were examined and reported .
All previously published loci for skin fluorescence were tested for association with SAF. We conducted a GWAS of SR and calculated shared heritability of SAF and SR. All signals identified for SR were tested for association with SAF and vice versa, all identified SNPs associated with SAF were tested for association with SR. All individuals were of European descent, following exclusion of non-European descent using principal component analyses.
Participants from the Lifelines Cohort Study, a large population-based study in the northern region of the Netherlands (Additional File 1: Supplementary Methods), were included . All individuals provided written informed consent before participating in the study, which was approved by the Medical Ethics Review Committee of the University Medical Center Groningen. At baseline, both physical and laboratory examination were performed, and extensive questionnaire data were collected .
Genotyping and imputation
The first set of participants of the Lifelines Cohort Study comprised 10,110 individuals with SAF measurements available, who were genotyped with the Illumina CytoSNP 12v2 (these overlap with 8,721 subjects previous reported by Eny et al., ). An additional set of 17,424 participants with available SAF measurements were genotyped using the Illumina Infinium Global Screening Array-24 (GSA) version 1 chip (Additional File 1: Supplementary Methods; Additional File 2: Figure S1).
Prior to imputation, genotyped SNPs with a call rate < 99%, monomorphic SNPs and SNPs with a Hardy–Weinberg equilibrium (HWE) p value < 10−6, were excluded. Subjects with a call rate < 99%, heterozygosity rate > 4 standard deviations (SD) from the mean, subjects with mismatch between self-reported sex and sex chromosome composition and outliers in principal component analyses were subsequently excluded (Additional File 2: Figure S1).
Both datasets were separately phased using SHAPEIT2 and imputed to the Haplotype Reference Consortium (HRC) panel, build GRCh37 . All well-imputed SNPs (INFO ≥ 80%) and minor allele frequency (MAF) > 0.1% were included in the analyses.
Phenotype and covariates measurement
At baseline examination, participants completed a self-administered questionnaire on medical history, past and current diseases, and health behaviour. Information regarding smoking behaviour (never, former and current smoking), as well as coffee consumption (cups/day), was collected by questionnaire . Weight was measured to the nearest 0.1 kg and height to the nearest 0.5 cm, with participants wearing light clothing and no shoes. BMI was calculated as weight/height2 in kg/m2.
SAF was measured on the volar side of the forearm with the AGE Reader (Diagnoptics, Groningen, The Netherlands), using emitted light from an internal UVA blacklight source in the range 350–420 nm with a peak around 370 nm . An internal spectrometer measured emitted and reflected light over 420–600 nm and 300–420 nm, respectively. Additionally, a reflectance spectrum was obtained using light from a white LED. SAF was calculated with AGE Reader software (v2.3) from the ratio between the emitted and the reflected light with skin colour variation taken into account . Skin reflectance (SR) was the mean skin reflectance determined with the UVA excitation source in the range 350–420 nm.
Blood samples were taken in the fasting state between 08:00 and 10:00 h and transported to the laboratory facility at room temperature or 4 °C, depending on the sample requirements. On the same day, HbA1c (EDTA-anticoagulated) was analysed using an NGSP-certified turbidimetric inhibition immunoassay on a Cobas Integra 800 CTS analyser (Roche Diagnostics Nederland, Almere, The Netherlands). Serum creatinine was measured on a Roche Modular P chemistry analyser (Roche, Basel, Switzerland) and renal function was calculated as estimated glomerular filtration rate (eGFR) with the formula developed by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) .
By definition, this general population cohort did include individuals with the metabolic syndrome or existing cardiovascular disease. Exclusion criteria for the present analyses were a SAF value higher than 4.5 AU, impaired glucose tolerance defined as elevated fasting blood glucose ≥ 7.0 mmol/l and/or HbA1c ≥ 6.5% (48 mmol/mol) or a previous diabetes diagnosis, and severely impaired renal function indicated by eGFR < 30 ml/min/1.73m2 (Additional file 2: Figure S1).
Heritability of SAF was tested by estimation of the SNP heritability through variance componence analysis using the BOLT- REML algorithm . Using a Monte Carlo average information restricted maximum likelihood algorithm, BOLT-REML is a computationally efficient program for estimation of trait heritability and shared heritability. Distribution of SAF values was right skewed, therefore natural log transformed SAF (lnSAF) was chosen as the outcome in the analyses. We calculated heritability for lnSAF separately without covariates in the model and in a model adjusted for sex, smoking status, mode of inclusion into the Lifelines cohort, age, age2, BMI and eGFR (Additional File 1: Supplementary Methods). The GSA cohort was used to calculate the heritability estimates since it has highest number of subjects with SAF measurements.
The primary outcome of GWAS was lnSAF. SNPs were tested for association with SAF under an additive genetic model in a linear mixed model, using BOLT-LMM v2.3.4 . In this model a genetic relationship matrix is calculated prior to analyses for which the model is adjusted to correct for relatedness and population structure within the cohort.
Since genetic variants may be associated with SAF through effects on covariates associated with SAF such as HbA1c, coffee consumption and SR, covariates were sequentially added in the GWAS models. Model 1 was adjusted for sex, smoking status, mode of inclusion into the Lifelines cohort, age, age2, body mass index (BMI), eGFR, and the NAT2 rs1495741 (indicators for homozygosity for the G allele and heterozygosity) . In model 2 we adjusted for all parameters of model 1 and HbA1c. In model 3, in addition to the covariates of model 2, coffee drinking status (yes/no) and daily coffee consumption in cups per day were added. In model 4 SR was included as an additional covariate.
An additional GWAS was performed for SR, which is a proxy for skin colour/pigmentation. The model was adjusted for age, sex, smoking group, source of inclusion, month of testing, BMI and eGFR. Shared heritability of SAF and SR was calculated in a model without covariates, and in a model adjusted for sex, smoking status, source of inclusion, age, age2, BMI and eGFR. Meta-GWAS of the CytoSNP and GSA cohort was performed using METAL v1.5 with the STDERR method which uses effect size estimates and standard errors . Double GC correction was not performed. A P-value < 5 × 10−8 was considered genome-wide significant. For visualisation of the GWAS results Manhattan and QQ plots were created in R version 4.0.3 using the ‘qqman’ package [27, 28]. In the QQ plot lambda genomic control (λgc) values are visualised . In addition, regional plots were generated using LocusZoom [21, 30].
SNP-covariate interactions were tested in a linear regression model in R version 4.0.3 .
Conditional analysis was performed for all novel loci separately added as a covariate to GWAS model 4, in order to test whether multiple loci in the same region were independently associated. When multiple loci within a genetic region were found, linkage disequilibrium (LD) and recombination rates were examined in the population genotyped on the GSA chip using PLINK v2.0-alpha2 . GWAS model 5 for lnSAF was adjusted for all covariates in model 4 and all novel identified loci to increase power for detection of additional loci.
In total, 27,534 participants from the Lifelines Cohort Study, a large population-based study in the northern region of the Netherlands (Additional File 1: Supplementary Methods; Additional File 2: Figure S1) were available for analysis. Descriptive statistics of the participants are summarized in Table 1. Mean age of the participants was 44 years and 40.9% was male, with a median BMI of 25.0 [22.8, 27.6] kg/m2, median eGFR 99 [88, 109] ml/min/1.73m2 and median coffee consumption of 3.3 [1.9, 4.7] cups per day. Median SR was 0.20 [0.16, 0.25], a Hexbin plot for lnSAF and SR is depicted in Additional File 3: Figure S2.
No participant reported an earlier diagnosis of diabetes or had elevated fasting blood glucose following exclusion of individuals with fasting glucose ≥ 7.0 mmol/l and/or HbA1c ≥ 6.5%
The heritability estimate of lnSAF was 20 ± 2% (SE) without covariates. Adjusted for sex, smoking group, source of inclusion, age, age2, BMI and eGFR, the heritability estimate of lnSAF was 33 ± 2% (SE) in the GSA data. Heritability of SR without covariates in the model was 35 ± 2%. Shared heritability of SAF and SR without covariates was 33 ± 6%. Heritability of SR adjusted for sex, smoking group, source of inclusion, age, age2, BMI and eGFR was 37 ± 2% (SE). Shared heritability of SAF and SR in the model adjusted for the aforementioned covariates was 24 ± 4% (SE).
Genome-wide association analyses.
In meta-analysis of M1 9,900,019 SNPs were tested, in M2 9,897,818 and in M3 and M4 9,889,051 SNPs were tested. SNPs were tested when they passed QC in one of two datasets. In total, five SNPs on chromosomes 8, 11, 15 and 16 exceeded the genome wide significance threshold (Table 2). The top SNPs of these regions were well-imputed or genotyped and of good quality (Additional File 4: Table S1). In Figs. 1 and 2 Manhattan plots and QQ-plots with the λgc of models 1–4 are displayed.
The G-allele of rs12931267 on 16q24.3 (Chr16:89,818,732,C > G, build GRCh37) was associated with higher SAF in all models with the smallest P-value (model 1: ß = 0.025, SE = 0.003, P = 3.0 × 10–20). The SNP is in intron 30 of FANCA (NM_000135.4), near MC1R (Fig. 3). The beta and direction of the effect was similar across all models, adding SR in model 4 led to a slight attenuation of the SNP-effect. The SNP explains 0.11% of the SAF variance in the GSA data in model 4. There was no significant interaction of rs12931267 with SR on SAF P = 0.19, Additional File 5: Table S2.
Conditional analysis in model 4 where rs12931267 was included additively as a covariate to the model revealed another SNP in this region exceeding the genome-wide significance threshold: rs3764257 (Chr16:89,800,887,C > G, build GRCh37, β = -0.010, SE = 0.0016, P = 3.8 × 10–10). rs3764257 is 17.8 kb upstream from rs12931267, these SNPs are on the same haplotype (D’ = 0.998, r2 = 0.03) but there was no significant interaction between the two SNPs (P = 0.07). The minor C-allele of rs3764257, with allele frequency of 28%, was associated with lower SAF. The SNP is an intronic variant in ZNF276. This variant explains 0.02% of the SAF variance in the GSA data in model 4.
Additional conditional analysis adjusted for both rs12931267 and rs3764257 coded additively in model 1 did not reveal other independently associated genome-wide significant loci in the region.
rs2846707 (Chr11:102,576,358,C > T) is a well-imputed SNP, and the alternate T-allele results in a Met30Val missense variant in MMP27 exon 1 (NM_022122.3). The presence of the T-allele was associated with lower SAF. The SNP-effect was similar across the different models (Fig. 4). 0.11% of the variance was explained by this variant in model 4 of the GSA data.
There was no significant association of rs2846707 with SR (P = 0.66).
The T-allele of rs2470893 (Chr15:75,019,449,C > T), in intergenic region on chromosome 15 located between CYP1A1 and CYP1A2 (Fig. 5) was associated with higher SAF (β = 0.011, SE = 0.002, P = 6.6 × 10–13 in model 1). The SNP-effect was attenuated in Model 3 where coffee intake was included as a covariate to the model (β = 0.008, SE = 0.002, P = 3.0 × 10–7). In model 4 of the GSA data the SNP explained 0.04% of the variance of SAF.
The interaction-term of rs2470893 with daily coffee consumption was not significant for SAF (P = 0.33). Also, the interaction-term of the SNP with eGFR was not associated with SAF (P = 0.88, Additional File 6: Table S3).
rs576201050 (Chr8:18,288,053,G > A) is an independent locus situated approximately 30 kb downstream from NAT2 on chromosome 8, which was significantly associated with SAF (Fig. 6). The minor allele A, with frequency of 1%, was associated with a lower SAF value (ß = -0.047, SE = 0.007, P = 3.6 × 10−11). The added variance explained with this SNP when added in model 4 of the GSA data was 0.07%.
Interaction term of rs576201050 and rs1495741 (the NAT2-acetylator tagging SNP) was significantly associated with SAF in the GSA data (Additional File 7: Table S4). The SNPs are in LD (D' = 1 and r2 = 0.003).
Estimation of SAF variance
Of all models, GWAS model 4 explained the highest percentage of the SAF variance, 48.54%. Adding all five newly identified SNPs to the linear regression model increased the adjusted r2 to 48.96%. Final GWAS adjusted for all covariates in M4 including all 5 newly identified SNPs did not reveal additional loci associated with SAF.
The association of the covariates with the different models for SAF in the GSA and CytoSNP datasets are displayed in Tables 3 and 4, respectively. Covariates in the regression model of the GSA data explained 46–49% of the SAF variance while the same covariates in the models explain 37–39% of the SAF variance in the CytoSNP data. Adding more covariates in models 2–4 does not drastically improve adjusted r2. The total variance explained for SAF is increased by 2.4% in the GSA data and 3.6% in the CytoSNP data when all genetic variants (including NAT2) are added to the model, compared with the model only adjusted for the known environmental and lifestyle determinants.
Replication of earlier reported associations
In addition to rs1495741 (NAT2-acetylator tagging SNP), rs7533564 (Chr1:78,825,912, C < T) was associated with SIF (Skin Intrinsic Fluorescence; a comparable measurement to SAF measured with the SCOUT DS device) in a previous study in type 1 diabetes individuals . This SNP was not associated with SAF in Lifelines (β = -0.001, SE = 0.005, P = 0.82, MAF 2.1% in meta-analysis of model 1).
GWAS of skin reflectance
The results of the GWAS of SR are summarized in Additional File 8: Table S5, Additional File 9: Table S6 and Additional File 10: Figure S2. In total, we identified 21 independent signals which were associated with SR. The strongest association was observed with rs35096708 on chromosome 16 near MC1R, intronic to CPNE7. Conditional analyses revealed that there were fourteen independent SNPs associated with SR in this region (Additional File 9: Table S6). All other loci were independently associated with SR but had no genome-wide significant association with SAF. All SNPs for SAF were tested for association with SR (Additional File 11: Table S7), and in addition to rs12931267 and rs3764257, rs1495741 (NAT2-acetylator tagging SNP) was nominally associated with SR with the opposite direction of effect.
SNPs for coffee consumption and SAF
As higher daily coffee consumption is associated with higher SAF levels , and the association of rs2470893 in the current study was attenuated when coffee intake was included as a covariate in our models. We decided to test all known SNPs associated with coffee consumption for association with SAF . The results of these post-hoc evaluations are shown in Additional File 12: Table S8. In addition to rs2470893 near CYP1A1/2, rs10865548, rs4410790, rs73073176, rs1057868 and rs2330783 were nominally associated with SAF with the same directions of effect.
Expression quantitative trait loci (cis-eQTLs)
In Additional Files 13,14,15,16: Supplementary Tables S9-12 cis-eQTLs of the top associated SNPs for SAF are described . Four out of five novel identified SNPs for SAF were cis-eQTLs in GTEx (v8) and were associated with expression of various genes within the chromosomal region and in various tissue. The SNPs near NAT2 i.e. rs1495741 and rs576201050 were not cis-eQTLs.
In this large group of individuals without diabetes of Western European descent, we estimated the heritability of SAF to be 33%, taking into account the known (in part environmental) factors which may influence SAF levels. We identified four novel loci with 5 SNPs, one on chromosomes 11 and 15, two on chromosome 16, and one locus on chromosome 8 near the previously identified locus NAT2, that were associated with SAF.
Most of the studies on SAF have reported on environmental factors associated with this measurement [32, 34]. SAF increases with ageing, and is elevated in people with type 2 diabetes compared with age-matched controls [11, 35]. It was demonstrated that SAF is already elevated in people without diabetes but with metabolic syndrome, and is associated with some of its individual components . Furthermore, SAF is strongly related to current smoking status as well as smoking history, coffee consumption, and renal function . In the general population SAF is a strong predictor of the development of cardiovascular disease and mortality . In addition, higher SAF levels predicted a higher risk of development of type 2 diabetes, and algorithms have been developed to predict individuals with the highest risk of developing type 2 diabetes based on age, BMI, and SAF levels . SAF reflects fluorescent properties present in skin and skin collagen half-life is about 15 years . As a result, SAF is a marker for cumulative accumulation over a much longer period of time than for instance haemoglobin A1c . Compared with excreted molecules or markers, SAF may therefore be a more appropriate tool for screening for diseases that develop over time.
Because of the growing interest to use SAF to screen for type 2 diabetes, diabetes complication risks, cardiovascular disease and mortality [12, 37], more information on possible genetic determinants of SAF is warranted, especially because studies of serum AGEs have estimated heritability of 63–74% [14, 15]. Understanding the genetics of a quantitative trait can improve performance of the tool. Unfortunately, there are only a few genetic studies of SAF [16, 17].
Our previous study reported the genetic association between a SNP in NAT2 and SAF, in a first set of approximately 9000 participants of the Lifelines Cohort Study, that overlaps partly with those reported in the CytoSNP cohort here . We also reported the association of a locus on chromosome 1 and SIF, but this association was only observed in individuals with type 1 diabetes . The present study has combined the earlier dataset of the Lifelines Cohort Study, and added a new set of over 17,000 individuals genotyped recently on the Illumina Infinium Global Screening Array-24 (GSA) version 1.
In meta-GWAS, we identified five additional loci which are associated with SAF in individuals without diabetes.
The association of the locus on chromosome 16 was the most significantly associated with SAF after NAT2. rs12931267 was significantly associated in all GWAS models, however the SNP-effect was slightly attenuated in model 4 where SR was added as a covariate. The SNP is near MC1R, in which multiple independent coding variants have been associated with skin colour and pigmentation traits [40, 41]. In GWAS of SR measures from the same individuals we identified multiple genome wide significant signals in the MC1R region as well. Through internal software in the AGE Reader, SAF is adjusted for variation in skin colour by SR . However, in multivariable models SR and SAF are still correlated. This association is complex and requires further investigation.
On chromosome 11, rs2846707 which results in a Met30Val missense variant in MMP27 exon 1 was associated with SAF. The pathologic significance of the association could be by the altering of an amino acid, however the SNP is also a cis-eQTL and associates with the expression level of several nearby genes. It is not clear whether there is allelic imbalance as the functional impact of this specific variant in MMP27 has not been investigated. Previous research did demonstrate an interplay between matrix metalloproteinases (MMPs) and AGEs, specifically because of the known collagen degrading activity of MMPs [42,43,44]. There was no association of this SNP with SR. In PheWeb, comprising the genetic (TOPMed imputed) and clinical information from the White British participants of the UK Biobank, the trait with the smallest P-value for rs2846707 was ‘Diffuse diseases of connective tissue’ (p = 0.002) and the second most associated trait was ‘Diabetic retinopathy’ (p = 0.002) [45, 46].
The effect of rs2470893 on chromosome 15 was attenuated when coffee intake -an important environmental factor influencing SAF- was included as a covariate. It can be postulated that some SNPs which are associated with daily coffee consumption may also be associated with SAF levels, based on the fact that higher coffee consumption is associated with higher SAF [32, 34]. Indeed, in our post-hoc evaluation, we found that multiple SNPs associated with coffee consumption were also nominally associated with SAF with similar direction of the effect (additional File 12, Table S8). A variety of factors potentially resulting from the roasting process of coffee beans (Maillard Reaction) may be involved in the association of increased coffee consumption and higher SAF, examples are the loss of chlorogenic acid, the presence of melanoidins or caffeine [32, 34, 47]). The relative contribution of each of these fluorescent factors, and potentially others, on SAF is unknown.
We identified an additional independent locus in the NAT2 region that was associated with SAF, which underlines the robust association of NAT2 variants with SAF. Previous research identified that NAT2 is also associated with insulin resistance . This shared association of insulin resistance as well as SAF with NAT2 may, in part, be explanatory for the predictive value of SAF for type 2 diabetes development.
We examined all SNPs in the GTEx project and found that rs12931267, rs2846707, rs2470893 and rs3764257 are all cis-eQTLs for various nearby genes. However, if the SNP-effect on SAF is explained by the association of the SNP with gene-expression, requires further investigation. For rs1495741 and rs576201050 we found no cis-eQTLs effects in GTEx. This may be caused by low minor allele frequency of rs576201050, since cis-eQTLs are only described for variants with MAF > 1% . Since rs1495741 is a tag-SNP for the acetylator phenotype which well explains the association with SAF, it is reasonable that the association is not mediated through cis-eQTL effects.
All newly identified loci explain a small percentage (below 1%) of the variance in SAF. In combination with the previously reported NAT2 SNP they explain less than 5% of the total SAF variation. However, insight into the genetic factors associated with SAF variation does provide more knowledge of the complex pathology of formation and accumulation of AGEs. As the estimated genetic heritability of SAF was 33%, the majority of genetic variants that explain SAF variability remain undiscovered. It seems reasonable that numerous SNPs each explain a small fraction of SAF variation, which could be identified by larger GWAS studies to increase power for detection of loci with even smaller effect sizes.
We also evaluated genome-wide signals for skin reflectance, an additional measure from the AGE Reader. SR is important for the proper measurement of SAF. Using the current AGE Reader only SAF values can be reliably measured when SR measurements are above 6%. As mentioned, SR is a proxy for skin colour or pigmentary phenotype. All genetic regions associated with SR have indeed earlier been shown to be associated with skin colour traits and perceived age, including SLC45A2, IRF4, BNC2, TYR, HERC2, MC1R and RALY/ASIP [41, 50]. Furthermore, variants in the MC1R region have been shown to be strongly associated with higher susceptibility of cutaneous melanoma [13, 51]. In our heritability analyses we found that SAF and SR have some shared heritability. Part of this shared heritability can be explained by the shared clinical determinants of the two traits; i.e. kidney function, coffee consumption, smoking behaviour and BMI. This explains why the shared heritability between SAF and SR is higher in the model not adjusted for any covariates compared to the model adjusted for covariates, 33% and 24% respectively. The remaining shared heritability of the SAF and SR cannot be explained in the current study, it may depend on specific skin structures associated with both SR as well as SAF, this requires further investigation.
Limitations of the current study are the fact that we used two different genotyping arrays on different subjects and we used HRC imputed data while the TOPMed panel comprises a larger reference panel and results in more imputed SNPs as well as better quality imputation . However, for imputation we are dependent on the Lifelines research group who have not yet released the TOPMed imputed Lifelines data. Another limitation is that we only analyzed autosomal SNPs while there is a considerable possibility that some of the SAF variance is explained by variation on the sex chromosomes. Finally, our study comprises only white individuals from European descent and the results are therefore not generalizable to the non-white population.
The estimated heritability of SAF is 33%, considering the known factors which may influence SAF levels. We identified five novel loci which are associated with SAF in the non-diabetes population. Together with the earlier reported NAT2 polymorphism these SNPs explain less than 5% of the SAF variance in the non-diabetes population.
Availability of data and materials
The manuscript is based on data from the Lifelines Cohort Study. Lifelines adheres to standards for data availability, and allows access for reproducibility of the study results. The data catalogue is publicly accessible at www.lifelines.nl. The dataset supporting the conclusions of this article is available through the Lifelines organization, and all international researchers can apply for data access at the Lifelines research office (firstname.lastname@example.org). For data access, a fee is required.
Advanced glycation endproducts
Body mass index
Estimated glomerular filtration rate
Expression quantitative trait loci
Infinium Global Screening Array-24
Genome-wide association studies
Haplotype reference consortium
Skin intrinsic fluorescence
Singe Nucleotide Polymorphism
Type 2 diabetes
Schmidt AM, Hori O, Brett J, Shi Du Yan, Wautier JL, Stern D. Cellular receptors for advanced glycation end products: Implications for induction of oxidant stress and cellular dysfunction in the pathogenesis of vascular lesions. Arterioscler Thromb 1994;14:1521–8. https://doi.org/10.1161/01.atv.14.10.1521.
McCance DR, Dyer DG, Dunn JA, Bailie KE, Thorpe SR, Baynes JW, et al. Maillard reaction products and their relation to complications in insulin-dependent diabetes mellitus. J Clin Invest. 1993;91:2470–8. https://doi.org/10.1172/JCI116482.
Mácsai E, Takáts Z, Derzbach L, Körner A, Vásárhelyi B. Verification of skin autofluorescence values by mass spectrometry in adolescents with type 1 diabetes: Brief report. Diabetes Technol Ther. 2013;15:269–72. https://doi.org/10.1089/dia.2012.0251.
Meerwaldt R, Hartog JWL, Graaff R, Huisman RJ, Links TP, den Hollander NC, et al. Skin Autofluorescence, a Measure of Cumulative Metabolic Stress and Advanced Glycation End Products, Predicts Mortality in Hemodialysis Patients. J Am Soc Nephrol. 2005;16:3687–93. https://doi.org/10.1681/ASN.2005020144.
Orchard TJ, Lyons TJ, Cleary PA, Braffett BH, Maynard J, Cowie C, et al. The association of skin intrinsic fluorescence with type 1 diabetes complications in the DCCT/EDIC Study. Diabetes Care. 2013;36:3146–53. https://doi.org/10.2337/dc12-2661.
Monnier VM, Bautista O, Kenny D, Sell DR, Fogarty J, Dahms W, et al. Skin collagen glycation, glycoxidation, and crosslinking are lower in subjects with long-term intensive versus conventional therapy of type 1 diabetes: Relevance of glycated collagen products versus HbA(1c) as markers of diabetic complications. vol. 48. 1999. https://doi.org/10.2337/diabetes.48.4.870.
Conway BN, Aroda VR, Maynard JD, Matter N, Fernandez S, Ratner RE, et al. Skin Intrinsic Fluorescence Is Associated With Coronary Artery Disease in Individuals With Long Duration of Type 1 Diabetes. Diabetes Care. 2012;35:2331–6. https://doi.org/10.2337/dc12-0053.
Klein BEK, Horak KL, Maynard JD, Lee KE, Klein R. Association of Skin Intrinsic Fluorescence with Retinal Microvascular Complications of Long Term Type 1 Diabetes in the Wisconsin Epidemiologic Study of Diabetic Retinopathy. Ophthalmic Epidemiol. 2017;24:211–6. https://doi.org/10.1080/09286586.2016.1269934.
Van Waateringe RP, Mook-Kanamori MJ, Slagter SN, Van Der Klauw MM, Van Vliet-Ostaptchouk JV, Graaff R, et al. The association between various smoking behaviors, cotinine biomarkers and skin autofluorescence, a marker for advanced glycation end product accumulation. PLoS ONE. 2017;12: e0179330. https://doi.org/10.1371/journal.pone.0179330.
Larroumet A, Foussard N, Majchrzak C, Monlun M, Blanco L, Mohammedi K, et al. Skin autofluorescence predicts mortality in type 1 diabetes. Diabetes Metab 2021:101273. https://doi.org/10.1016/j.diabet.2021.101273.
van Waateringe RP, Fokkens BT, Slagter SN, van der Klauw MM, van Vliet-Ostaptchouk JV, Graaff R, et al. Skin autofluorescence predicts incident type 2 diabetes, cardiovascular disease and mortality in the general population. Diabetologia. 2019;62:269–80. https://doi.org/10.1007/s00125-018-4769-x.
Foussard N, Larroumet A, Rigo M, Mohammedi K, Baillet-Blanco L, Poupon P, et al. Skin autofluorescence predicts cancer in subjects with type 2 diabetes. BMJ Open Diabetes Res Care. 2021;9: e001312. https://doi.org/10.1136/BMJDRC-2020-001312.
Landi MT, Bishop DT, MacGregor S, Machiela MJ, Stratigos AJ, Ghiorzo P, et al. Genome-wide association meta-analyses combining multiple risk phenotypes provide insights into the genetic architecture of cutaneous melanoma susceptibility. Nat Genet 2020 525 2020;52:494–504. https://doi.org/10.1038/s41588-020-0611-8.
Leslie RDG, Beyan H, Sawtell P, Boehm BO, Spector TD, Snieder H. Level of an Advanced Glycated End Product Is Genetically Determined. Diabetes. 2003;52:2441–4. https://doi.org/10.2337/DIABETES.52.9.2441.
Adams JN, Raffield LM, Martelle SE, Freedman BI, Langefeld CD, Carr JJ, et al. Genetic Analysis of Advanced Glycation End Products in the DHS MIND Study. Gene. 2016;584:173. https://doi.org/10.1016/J.GENE.2016.02.029.
Eny KM, Lutgers HL, Maynard J, Klein BEK, Lee KE, Atzmon G, et al. GWAS identifies an NAT2 acetylator status tag single nucleotide polymorphism to be a major locus for skin fluorescence. Diabetologia. 2014;57:1623–34. https://doi.org/10.1007/s00125-014-3286-9.
Roshandel D, Klein R, Klein BEK, Wolffenbuttel BHR, Van Der Klauw MM, Van Vliet-Ostaptchouk JV, et al. New Locus for Skin Intrinsic Fluorescence in Type 1 Diabetes Also Associated With Blood and Skin Glycated Proteins. Diabetes. 2016;65:2366–76. https://doi.org/10.2337/db15-1484.
Ardlie KG, DeLuca DS, Segrè A V., Sullivan TJ, Young TR, Gelfand ET, et al. The Genotype-Tissue Expression (GTEx) pilot analysis: Multitissue gene regulation in humans. Science (80- ) 2015;348:648–60. https://doi.org/10.1126/science.1262110.
Scholtens S, Smidt N, Swertz MA, Bakker SJ, Dotinga A, Vonk JM, et al. Cohort Profile: LifeLines, a three-generation cohort study and biobank. Int J Epidemiol. 2015;44:1172–80. https://doi.org/10.1093/IJE/DYU229.
McCarthy S, Das S, Kretzschmar W, Delaneau O, Wood AR, Teumer A, et al. A reference panel of 64,976 haplotypes for genotype imputation. Nat Genet. 2016;48:1279–83. https://doi.org/10.1038/ng.3643.
Pruim RJ, Welch RP, Sanna S, Teslovich TM, Chines PS, Gliedt TP, et al. LocusZoom: regional visualization of genome-wide association scan results. Bioinformatics. 2010;26:2336. https://doi.org/10.1093/BIOINFORMATICS/BTQ419.
Koetsier M, Lutgers HL, De Jonge C, Links TP, Smit AJ, Graaff R. Reference values of skin autofluorescence. Diabetes Technol Ther. 2010;12:399–403. https://doi.org/10.1089/dia.2009.0113.
Koetsier M, Nur E, Chunmao H, Lutgers HL, Links TP, Smit AJ, et al. Skin color independent assessment of aging using skin autofluorescence. Opt Express. 2010;18:14416. https://doi.org/10.1364/oe.18.014416.
Levey AS, Stevens LA, Schmid CH, Zhang Y, Castro AF, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–12. https://doi.org/10.7326/0003-4819-150-9-200905050-00006.
Loh PR, Tucker G, Bulik-Sullivan BK, Vilhjálmsson BJ, Finucane HK, Salem RM, et al. Efficient Bayesian mixed-model analysis increases association power in large cohorts. Nat Genet. 2015;47:284–90. https://doi.org/10.1038/ng.3190.
Willer CJ, Li Y, Abecasis GR. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinforma Appl NOTE. 2010;26:2190–1. https://doi.org/10.1093/bioinformatics/btq340.
R Core Team. R: A Language and Environment for Statistical Computing 2020.
Turner SD. qqman: an R package for visualizing GWAS results using Q-Q and manhattan plots. BioRxiv. 2014;3: 005165. https://doi.org/10.1101/005165.
Yang J, Weedon MN, Purcell S, Lettre G, Estrada K, Willer CJ, et al. Genomic inflation factors under polygenic inheritance. Eur J Hum Genet. 2011;19:807. https://doi.org/10.1038/EJHG.2011.39.
LocusZoom is a suite of tools to provide fast visualization of GWAS results for research and publication n.d. http://locuszoom.sph.umich.edu/locuszoom/. (Accessed 21 Aug 2020)
Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7. https://doi.org/10.1186/S13742-015-0047-8.
van Waateringe RP, Slagter SN, van der Klauw MM, van Vliet-Ostaptchouk JV, Graaff R, Paterson AD, et al. Lifestyle and clinical determinants of skin autofluorescence in a population-based cohort study. Eur J Clin Invest. 2016;46:481. https://doi.org/10.1111/ECI.12627.
Cornelis MC, Byrne EM, Esko T, Nalls MA, Ganna A, Paynter N, et al. Genome-wide meta-analysis identifies six novel loci associated with habitual coffee consumption. Mol Psychiatry. 2015;20:647–56. https://doi.org/10.1038/mp.2014.107.
Eny KM, Orchard TJ, Miller RG, Maynard J, Grant DM, Costacou T, et al. Caffeine Consumption Contributes to Skin Intrinsic Fluorescence in Type 1 Diabetes. Diabetes Technol Ther. 2015;17:726–34. https://doi.org/10.1089/dia.2015.0017.
Yasuda M, Shimura M, Kunikata H, Kanazawa H, Yasuda K, Tanaka Y, et al. Relationship of skin autofluorescence to severity of retinopathy in type 2 diabetes. Curr Eye Res. 2015;40:338–45. https://doi.org/10.3109/02713683.2014.918152.
Van Waateringe RP, Slagter SN, Van Beek AP, Van der Klauw MM, Van Vliet-Ostaptchouk J V, Graaff R, et al. Skin autofluorescence, a non-invasive biomarker for advanced glycation end products, is associated with the metabolic syndrome and its individual components. Diabetol Metab Syndr 2017;9. https://doi.org/10.1186/s13098-017-0241-1.
Boersma HE, Waateringe RP van, Klauw MM van der, Graaff R, Paterson AD, Smit AJ, et al. Skin autofluorescence predicts new cardiovascular disease and mortality in people with type 2 diabetes. BMC Endocr Disord 2021;21. https://doi.org/10.1186/S12902-020-00676-4.
Verzijl N, DeGroot J, Thorpe SR, Bank RA, Shaw JN, Lyons TJ, et al. Effect of collagen turnover on the accumulation of advanced glycation end products. J Biol Chem. 2000;275:39027–31. https://doi.org/10.1074/JBC.M006700200.
Wolffenbuttel BHR, Giordano D, Founds HW, Bucala R. Long-term assessment of glucose control by haemoglobin-AGE measurement. Lancet. 1996;347:513–5. https://doi.org/10.1016/S0140-6736(96)91141-1.
Flanagan N, Healy E, Ray A, Philips S, Todd C, Jackson IJ, et al. Pleiotropic effects of the melanocortin 1 receptor (MC1R) gene on human pigmentation. Hum Mol Genet. 2000;9:2531–7. https://doi.org/10.1093/hmg/9.17.2531.
Visconti A, Duffy DL, Liu F, Zhu G, Wu W, Chen Y, et al. Genome-wide association study in 176,678 Europeans reveals genetic loci for tanning response to sun exposure. Nat Commun 2018 91 2018;9:1–7. https://doi.org/10.1038/s41467-018-04086-y.
Rittié L, Berton A, Monboisse JC, Hornebeck W, Gillery P. Decreased contraction of glycated collagen lattices coincides with impaired matrix metalloproteinase production. Biochem Biophys Res Commun. 1999;264:488–92. https://doi.org/10.1006/BBRC.1999.1519.
Molinari J, Ruszova E, Velebny V, Robert L. Effect of advanced glycation endproducts on gene expression profiles of human dermal fibroblasts. Biogerontology. 2008;9:177–82. https://doi.org/10.1007/S10522-008-9129-7.
Okano Y, Masaki H, Sakurai H. Dysfunction of dermal fibroblasts induced by advanced glycation end-products (AGEs) and the contribution of a nonspecific interaction with cell membrane and AGEs. J Dermatol Sci. 2002;29:171–80. https://doi.org/10.1016/S0923-1811(02)00021-X.
Taliun D, Harris DN, Kessler MD, Carlson J, Szpiech ZA, Torres R, et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed Program. BioRxiv. 2019;2: 563866. https://doi.org/10.1101/563866.
UKBiobank TOPMed-imputed PheWeb n.d. https://pheweb.org/UKB-TOPMed/variant/11:102705627-T-C. (Accessed 14 Jan 2022)
Rufián-Henares JA, Pastoriza S. Melanoidins in Coffee. Coffee Heal Dis Prev. 2015:183–8. https://doi.org/10.1016/B978-0-12-409517-5.00020-6.
Knowles JW, Xie W, Zhang Z, Chennemsetty I, Assimes TL, Paananen J, et al. Identification and validation of N-acetyltransferase 2 as an insulinsensitivity gene. J Clin Invest. 2015;125:1739. https://doi.org/10.1172/JCI74692.
Aguet F, Barbeira AN, Bonazzola R, Brown A, Castel SE, Jo B, et al. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science (80- ) 2020;369:1318–30. https://doi.org/10.1126/SCIENCE.AAZ1776.
Liu F, Hamer MA, Deelen J, Lall JS, Jacobs L, van Heemst D, et al. The MC1R Gene and Youthful Looks. Curr Biol. 2016;26:1213–20. https://doi.org/10.1016/J.CUB.2016.03.008.
Bishop DT, Demenais F, Iles MM, Harland M, Taylor JC, Corda E, et al. Genome-wide association study identifies three loci associated with melanoma risk. Nat Genet 2009 418 2009;41:920–5. https://doi.org/10.1038/ng.411.
The authors would like to acknowledge all the study participants, services of the Lifelines Cohort Study and the contributing research centers delivering data to Lifelines.
The Lifelines Biobank initiative has been made possible by subsidy from the Dutch Ministry of Health, Welfare and Sport, the Dutch Ministry of Economic Affairs, the University Medical Center Groningen (UMCG the Netherlands), University Groningen and the Northern Provinces of the Netherlands. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Ethics approval and consent to participate
This study was approved by the medical ethical review committee of the University Medical Center Groningen. All participants provided written informed consent before participating in the study. (METC 2007/152).
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Additional file 3.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
About this article
Cite this article
Vollenbrock, C.E., Roshandel, D., van der Klauw, M.M. et al. Genome-wide association study identifies novel loci associated with skin autofluorescence in individuals without diabetes. BMC Genomics 23, 840 (2022). https://doi.org/10.1186/s12864-022-09062-x
- Skin Autofluorescence
- Genome-wide association study
- Single nucleotide polymorphism
- Skin Pigmentation