Variants in the TERT Gene Increase the Occurrence of Solar Lentigines by Modifying Telomerase Expression Exclusively in the Skin

Solar lentigines (SL) are a characteristic of the aging skin. There is a huge interindividual variability in the number of SL that are determined by a combination of genetic and environmental factors. While sun exposure represents the most important environmental factor (Bastiaens et al., 2004Bastiaens M. Hoefnagel J. Westendorp R. Vermeer B.-J. Bouwes Bavinck J.N. Solar Lentigines are Strongly Related to Sun Exposure in Contrast to Ephelides.Pigment Cell Research. 2004; 17: 225-229https://doi.org/10.1111/j.1600-0749.2004.00131.xCrossref PubMed Scopus (105) Google Scholar), the search for genetic susceptibility variants is still ongoing. Recently, a genome-wide association study (GWAS) in two Chinese cohorts (n=5,918) reported TERT (telomerase reverse transcriptase; top SNP rs2853672, p=5.68E-42) and OCA2 (oculocutaneous albinism II; rs1800414, p=4.67E-8) as two previously unreported genome-wide significant loci associated with a composite SL phenotype score based on the forearm and hands. However, only the TERT signal was validated in a smaller European cohort (n=462), albeit with sub-genome-wide significance (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar). TERT is a compelling candidate for skin aging as its (only known) function is the maintenance of telomere lengths, which shorten with age. Surprisingly, the TERT variants did not seem to act via alteration of telomere lengths despite multiple lines of investigation such as Mendelian randomization (MR) using GWAS on telomere length in blood leukocytes (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar). In this study, we performed comprehensive GWAS analyses of SL on the hands of 1,137 elderly European participants from the Berlin Aging Study II (BASE-II) that were independent from (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar). Furthermore, we applied extensive fine-mapping including colocalization (Giambartolomei et al., 2014Giambartolomei C. Vukcevic D. Schadt E.E. Franke L. Hingorani A.D. Wallace C. Plagnol V. Bayesian Test for Colocalisation between Pairs of Genetic Association Studies Using Summary Statistics.PLoS Genetics. 2014; 10e1004383https://doi.org/10.1371/journal.pgen.1004383Crossref Scopus (1167) Google Scholar) and MR (Broadbent et al., 2020Broadbent J.R. Foley C.N. Grant A.J. Mason A.M. Staley J.R. Burgess S. MendelianRandomization v0.5.0: updates to an R package for performing Mendelian randomization analyses using summarized data.Wellcome Open Research. 2020; 5: 252https://doi.org/10.12688/wellcomeopenres.16374.2Crossref PubMed Scopus (39) Google Scholar) analyses by taking into account tissue-specific effects, which were not considered previously (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar). Taken together, our data deliver strong support for the role of TERT in SL development and provide further insides into possible functional mechanisms. Detailed methods can be found in the Supplementary Methods. Briefly, the number and size of SL in BASE-II was quantified based on photographs of both hands using a categorical score (Supplementary Figure 1). All BASE-II participants provided written informed consent and the study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Charité-Universitätsmedizin Berlin – approval number EA2/029/09. Genome-wide SNP data were processed as described previously (Hong et al., 2020Hong S. Prokopenko D. Dobricic V. Kilpert F. Bos I. Vos S.J.B. Tijms B.M. Andreasson U. Blennow K. Vandenberghe R. Cleynen I. Gabel S. Schaeverbeke J. Scheltens P. Teunissen C.E. Niemantsverdriet E. Engelborghs S. Frisoni G. Blin O. Bertram L. Genome-wide association study of Alzheimer's disease CSF biomarkers in the EMIF-AD Multimodal Biomarker Discovery dataset.Translational Psychiatry. 2020; 10https://doi.org/10.1038/s41398-020-01074-zCrossref Scopus (25) Google Scholar). SNP-based heritability was calculated using GREML (Yang et al., 2010Yang J. Benyamin B. McEvoy B.P. Gordon S. Henders A.K. Nyholt D.R. Madden P.A. Heath A.C. Martin N.G. Montgomery G.W. Goddard M.E. Visscher P.M. Common SNPs explain a large proportion of the heritability for human height.Nature Genetics. 2010; 42: 565-569https://doi.org/10.1038/ng.608Crossref PubMed Scopus (2942) Google Scholar). We performed linear regression analyses of the SL score on each SNP (7,171,623 SNPs total) adjusting for genetic ancestry, sex, and (as sensitivity analyses) age. Our European-descent study yielded a SNP-based heritability estimate of 0.59 (SE=0.29), which is considerably higher than that reported for the Asian cohorts (0.32 to 0.42) (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar). Our GWAS identified variants in TERT as the only genome-wide significant association signal for SL (Figure 1, Supplementary Figure 2, Supplementary Tables 1-2). The most significant signal was elicited by SNP rs2735940 (β [G allele]=0.33, p=1.35E-16, allele frequency[G]=0.52, Supplementary Figure 3) and explained 4.9% of the phenotypic variance (compared to 2.95% in the Asian cohorts (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar)). Conditional analyses did not reveal additional independent significant TERT variants (FDR=0.01). The top SNP of (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar) is in perfect linkage disequilibrium (LD) with the top SNP of this study (r2=1), thus both represent the same signal despite the fact that they are based on different ethnic groups. In contrast to TERT, our analyses provided no noteworthy evidence for association with variants in OCA2 (Figure 1). We note that the C allele frequency of the lead OCA2 SNP in (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar) (rs1800414) was 0.001 in our European dataset, leading to its exclusion during quality control. In line with this, the C allele is nearly absent in all non-Asian ethnicities (GnomAD allele frequency in non-Asians<<0.002, allele frequency in Europeans ∼0.0002), but it represents the major allele in East Asians (frequency=0.6). Further work needs to clarify whether rs1800414 represents a genuine susceptibility variant in Asian populations. To conclusively assess OCA's role for SL in European populations, it would be necessary to investigate rs1800414 in substantially larger European cohorts and/or to apply different statistical approaches (e.g., testing for an enrichment of rare missense mutations). Given that the top TERT SNP rs2735940 is also significantly associated with telomere length in blood leukocytes (p=4.4E-114), we tested for a causative effect of telomere length on SL risk applying MR using the same data. In agreement with similar-minded MR results reported in (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar), we did not observe evidence for a causal effect (Supplementary Figure 4, Supplementary Table 3). We note that our MR analyses are based on the newest and largest telomere length GWAS (n=472,174) (Codd et al., 2021Codd V. Wang Q. Allara E. Musicha C. Kaptoge S. Stoma S. Jiang T. Hamby S.E. Braund P.S. Bountziouka V. Budgeon C.A. Denniff M. Swinfield C. Papakonstantinou M. Sheth S. Nanus D.E. Warner S.C. Wang M. Khera A.V. Samani N.J. Polygenic basis and biomedical consequences of telomere length variation.Nature Genetics. 2021; 53: 1425-1433https://doi.org/10.1038/s41588-021-00944-6Crossref PubMed Scopus (93) Google Scholar) derived from the same descent group as our primary SL GWAS (i.e. European), increasing the conclusive strength of the (negative) results presented here. Furthermore, using the same datasets, colocalization analyses showed a negligible posterior probability of the presence of a shared causal TERT variant for telomere length in blood and SL (PP.H4<1E-5), but instead yielded a high posterior probability for two distinct TERT signals across these phenotypes (PP.H3>0.99). The TERT signal comprises at least three SNPs that are in nearly perfect LD. Thus, each of them may represent the relevant functional SNP underlying the association signal, a distinction that cannot be made by statistical analyses: SNP rs2735940 maps into a proximal enhancer-like signature <1.4kb upstream of TERT (Encode-ID: EH38E2352730), while SNP rs2853672 (r2=1), reported by (Peng et al., 2023Peng Q. Liu Y. Huels A. Zhang C. Yu Y. Qiu W. Cai X. Zhao Y. Schikowski T. Merches K. Liu Y. Yang Y. Wang J. Zhao Y. Jin L. Zhang L. Krutmann J. Wang S. Genetic Variants in Telomerase Reverse Transcriptase Contribute to Solar Lentigines.Journal of Investigative Dermatology. 2023; 143: 1062-1072.e25https://doi.org/10.1016/j.jid.2022.11.016Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar), is located intronically (∼1kb from the intron-exon boundary of intron 2). The third SNP, rs33951612 (r2=0.98 to the other two SNPs), which was not directly tested in our study due to the fact that it is a 1-bp indel (alleles AAA vs AAAA), is also located in intron 2 (∼0.8kb from the intron-exon boundary). Interestingly, rs2735940 (as well as the other SNPs mentioned above) represents a strong cis expression quantitative trait locus (eQTL) in GTEx skin tissue (sun-exposed skin: p=5.5E-6, unexposed: p=1.7E-4, with the SL risk-increasing allele associated with increased TERT expression). Importantly, rs2735940 did not show eQTL effects in any of the other 15 GTEx tissues analyzed (FDR=0.01; Figure 1). This tissue specificity of the TERT eQTL signal for skin may explain the missing causal link between telomere length and SL using the MR and colocalization analyses described above as these were based on telomere length GWAS data in blood. In agreement with this notion, we detected evidence for shared causal variants for SL and TERT expression in skin but not blood: Specifically, the top SL SNP rs2735940 colocalized with SL occurrence and TERT expression in both GTEx sun-exposed (n=414, PP.H4=0.99) and unexposed skin (n=335, PP.H4=0.92), but not in blood (n=369, PP.H4=0.03). Analyses did not show evidence for colocalization for any of the other genes (n=43) located within the TERT locus and expressed in skin (all PP.H4<0.27), emphasizing the direct involvement of TERT in SL development. In summary, this study provides an independent validation for the role of TERT in the development of non-facial SL in a large European-descent dataset and explains the missing replication of the probably genuine OCA signal by large differences in the allele frequency between ethnicities. We show that TERT variants increase SL occurrence by modifying TERT expression exclusively in the skin. This skin-specific effect was not reported previously and may provide the hitherto missing functional link to the occurrence of SLs via TERT expression and telomere length alterations. Future functional studies are needed to decipher the molecular mechanisms underlying this association which can now be considered as "established". Summary statistics of the SL GWAS are publicly available at the GWAS Catalog (https://www.ebi.ac.uk/gwas/home, accession number: GCST90281274) as well as at the Zenodo platform (https://zenodo.org, accession number: 8337955). Subject-level data can be obtained by qualified investigators upon request to the authors. All GTEx data analyzed are publicly available at www.gtexportal.org/home/datasets. The summary statistics of the telomere length GWAS by Codd et al. are publicly available at https://figshare.com/s/caa99dc0f76d62990195 None of the authors reports any conflicts of interest. Conceptualization: C.M.L., L.B., data generation: V.H., J.S., V.D., E. S.-T., I. D., L.B. Formal analyses and statistical advice: L.D., O.O., J.H., J.S., K.B., C.M.L, writing of the initial manuscript draft: L.D., C.M.L., critical revision of the manuscript: all co-authors Legend. a. Genome-wide association results of the occurrence of non-facial solar lentigines (SL) in 1,137 participants of the Berlin Aging Study II (BASE-II). The red dotted line indicates the genome-wide significance threshold of α=5E-8 and the gray dotted line the sub genome-wide significance threshold of α=1E-5. Genes are annotated according to the closest coding gene if located +/-200kb around the corresponding SNP, otherwise the nearest non-coding gene in +/-1Mb is annotated. b. Locus zoom plot of the genome-wide significant index SNP rs2735940 in the SL GWAS. This plot was generated using FUMA. SNP rs2735940 is visualized in dark blue. All other SNPs are color-coded according to the linkage disequilibrium (r2) to the index SNP. SNPs that are not in LD with the top SNP (r2<0.1) are color-coded in gray. c. Multi-tissue expression quantitative trait locus (eQTL) plot for the top SNP rs2735940 and TERT expression. This figure was generated using the GTEx portal (Broad Institute, 2021Broad Institute. (2021). GTEx Portal. Broad Institute of MIT and Harvard. https://www.gtexportal.org/home/Google Scholar). NES=normalized effect size, m-value=posterior probability that an eQTL effect exists in each tissue tested in the cross-tissue meta-analysis, CI=confidence interval. d. Box plot of TERT expression data for rs2735940 genotypes in sun-exposed skin tissue. This figure was generated using the GTEx portal (Broad Institute, 2021Broad Institute. (2021). GTEx Portal. Broad Institute of MIT and Harvard. https://www.gtexportal.org/home/Google Scholar). This study used data from the BASE-II, which has been supported by the German Federal Ministry of Education and Research under grant numbers 16SV5538 (to L.B.) and 16SV5536K (to. E.S.-T.). The responsibility for the contents of this publication lies with its authors. C.M.L. was supported by the Heisenberg grant of the German Research Foundation (DFG; LI 2654/4–1). We thank Prof. Sebastian-Edgar Baumeister for helpful discussions and critical review of the manuscript. We thank the two reviewers for their thoughtful and constructive feedback on our manuscript. Download .pdf (.13 MB) Help with pdf files Download .pdf (.07 MB) Help with pdf files Download .pdf (.0 MB) Help with pdf files Download .pdf (.07 MB) Help with pdf files Download .xlsx (.02 MB) Help with xlsx files