Abstract
This study investigated epigenetic changes, specifically epigenetic ageing, in an adult cohort of healthy individuals using five validated epigenetic clock algorithms and a DNA methylation-based estimator of telomere length. Our study demonstrated significant biologic/epigenetic age dysregulation in sun-exposed vs. sun-protected dermal and epidermal skin, with a strong correlation to the validated Helfrich skin photoageing scale, occurring in patterns that overlap with those seen in basal cell carcinoma. This work highlights the power of novel epigenetic analyses in studying photoageing and skin cancer predisposition.
Dear Editor, Sun exposure is an established risk factor for skin cancer, known to induce widespread epigenetic changes in healthy skin that overlap with signatures found in squamous cell carcinoma.1 Epigenetic ‘clock’ algorithms can predict age and telomere length (TL) based on the tissue DNA methylation (DNAm) state. Discordance between predicted and chronological age (epigenetic age acceleration/deceleration) reflects genomic dysregulation and is associated with increased all-cause mortality and predisposition to cancer.2 Despite the observed global hypomethylation in sun-exposed skin, the effects of sun exposure on epigenetic ageing signatures and the overlap of such signatures with those in skin cancer have not been assessed.1 Thus, we investigated the role of DNAm-based ageing algorithms in predicting the deleterious impact of sun exposure in a sex- and tissue-specific manner.
Our analyses incorporated publicly available DNAm datasets of punch-biopsied sun-exposed and sun-protected healthy skin (GSE52980), and basal cell carcinoma (BCC) (GSE197723, GSE128784). The sun-exposure cohort included younger [n = 10 (five females, five males); mean (SD) age 26.6 (4.5) years] and older individuals [n = 10 (six females, four males); mean age 73.9 (8.8) years], from whom dermal and epidermal skin samples (n = 78) derived from sun-exposed and sun-protected body regions were used. The disease cohort included older individuals with BCC [n = 21 (9 females, 12 males); mean (SD) age 70.3 (11.4) years] and nonmalignant adjacent skin [n = 16 (4 females, 12 males); mean age 68.6 (12.2) years]. Samples were interrogated with five DNAm-based algorithms that collectively estimated age-adjusted epigenetic age (‘Horvath’ and ‘Skin & Blood’), biologic age (‘PhenoAge’ and ‘PCPhenoAge’) and TL (‘DNAmTL’) using the ‘methylCIPHER’ package for R.3 Analyses were adjusted for multiple testing using the false discovery rate method (q).4
We observed a strong positive Pearson correlation between chronological and epigenetic/biologic age (R = 0.88–0.98), and strong negative correlation between chronological age and TL (R = –0.77 to –0.59) in dermal and epidermal samples collectively; the former correlation was validated in the nonmalignant skin cohort (R = 0.65–0.94), demonstrating the accuracy of these algorithms in predicting age and TL in healthy skin.
We next assessed for epigenetic and biologic age, and TL dysregulation (defined as mean groupwise differences in residuals derived from the linear regression of predicted chronological age), using paired t-tests, and multinomial linear regression models adjusted for age, sex, skin group and tissue, accounting for pairing in both cohorts via random effects. Sun-exposed skin exhibited decreased epigenetic age [Horvath: –5.5 to –10.6 years, 95% confidence interval (CI) –3.5 to –15.7 (P = 7.2 × 10–3 to 4.5 × 10–2), t = –6.0 (P = 1.4 × 10–7); Skin & Blood: –4.6 to –5.5 years, 95% CI –2.0 to –8.7 (P = 7.2 × 10–3 to 1.1 × 10–2), t = –0.15 (P > 0.05)] and biologic age [PhenoAge: –2.8 to –8.3 years, 95% CI –0.8 to –12.7 (P = 7.2 × 10–3 to 4.5 × 10–2), t = –3.1 (P = 3.3 × 10–3); PCPhenoAge: –13.7 to 4.5 years, 95% CI –20.8 to 7.5 (P = 9.4 × 10–4 to 1.3 × 10–2), t = –2.6 (P = 1.3 × 10–2)], as well as shorter telomere length [DNAmTL: –0.3 to –0.4 kb, 95% CI 0.1 to –0.6 (P = 1.2 × 10–4 to 1.2 × 10–2), t = –5.2 (P = 3.3 × 10–6)] vs. sun-protected skin across both age groups and tissue types. This trend was observed in all tested groups except in the epidermal skin of younger individuals (17 total instances) and was more frequent among older (11-fold) vs. younger (6-fold) participants and among female (6-fold) vs. male (2-fold) participants. These findings were validated in BCC, which – compared with nonmalignant skin – exhibited decreased epigenetic and biologic age with all four age prediction algorithms [Horvath: t = –2.8 (P = 8.2 × 10–3); Skin & Blood: t = –2.6 (P = 1.5 × 10–2); PhenoAge: –16.5 years, 95% CI –26.0 to –7.1 (P = 5.0 × 10–3), t = –3.1 (P = 4.2 × 10–3); PCPhenoAge: –8.0 years, 95% CI –13.6 to –2.5 (P = 7.4 × 10–3), t = –2.7 (P = 1.2 × 10–2)], with no impact on TL (Table 1). Moreover, biologic age (PCPhenoAge) was negatively correlated with the Helfrich photo-protected skin ageing scale in both sun-exposed dermis and epidermis [Spearman’s ρ = –0.68 to 0.69, 95% CI –0.92 to –0.16 (P = 0.03 to 0.05)] but not in sun-protected skin, further supporting the link between biologic age deceleration and sun exposure.5 These trends were concordant with observations of decelerated epigenetic age in excess of 20 years in BCC assessed using overlapping clock algorithms, and may be indicative of increased skin cancer predisposition following lifetime sun exposure.6
Table 1.
A detailed summary of significant findings of epigenetic age dysregulation due to sun exposure and basal cell carcinoma (BCC) based on five algorithms
| DNA methylation algorithm (DOI) | Horvath (10.1186/gb-2013-14-10-r115) | Skin & Blood (10.18632/aging.101508) | PhenoAge (10.18632/aging.101414) | PCPhenoAge (10.1038/s43587-022-00248-2) | DNAmTL (10.18632/aging.102173) |
|---|---|---|---|---|---|
| Dermis, oldera | –8.7 (–4.2 to –13.2) [2.4 × 10–2] | –4.6 (–2.0 to –7.3) [2.5 × 10–2] | –7.1 (–2.5 to –11.3) [2.5 × 10–2] | ||
| Female | –10.6 (–5.6 to –15.7) [2.4 × 10–2] | –5.5 (–2.4 to –8.7) [3.2 × 10–2] | –8.3 (–3.9 to –12.7) [2.5 × 10–2] | ||
| Dermis, youngerb | –5.5 (–3.5 to –7.5) [4.4 × 10–3] | –4.3 (–1.4 to –7.1) [3.3 × 10–2] | 4.5 (1.5–7.5) [3.3 × 10–2] | ||
| Male | –4.6 (–1.7 to –7.5) [4.1 × 10–2] | ||||
| Female | –6.5 (–2.6 to –10.4) [4.0 × 10–2] | ||||
| Epidermis, olderc | –11.3 (–6.7 to –15.9) [7.5 × 10–3] | –0.3 (–0.2 to –0.4) [1.4 × 10–3] | |||
| Male | –0.4 (–0.3 to –0.5) [2.4 × 10–2] | ||||
| Female | –13.7 (–7.6 to –19.8) [2.5 × 10–2] | –0.3 (–0.1 to –0.4) [2.4 × 10–2] | |||
| Epidermis, youngerd | –0.3 (–0.09 to –0.5) [4.1 × 10–2] | ||||
| BCCe | –16.3 (–4.3 to –28.3) [4.4 × 10–2] | –8.0 (–3.3 to –15.2) [3.6 × 10–2] | |||
| Male | –20.4 (–7.6 to –33.1) [3.6 × 10–2] | –10.0 (–2.7 to –17.3) [4.4 × 10–2] |
Data are presented as paired t-test value (95% confidence interval) [adjusted false discovery rate q-value]. Blank fields represent nonsignificant findings. Sun exposure was associated with epigenetic and biologic age (in years) deceleration and reduced telomere length, while BCC was associated with biological age deceleration alone. The effects of sun exposure were more frequently observed in older females, while those of BCC were found more often in males. Effect sizes are expressed in years (epigenetic and biologic age) and kb (telomere length). a–dDysregulation in sun-exposed vs. sun-protected tissue, as well as in eBCC vs. healthy skin.
Lower-than-expected epigenetic age was linked to several cellular phenomena, tested specifically on the Skin & Blood clock, including increased mitochondrial activity, greater stem cell proliferation and reduced nutrient sensing. These effects are associated with upregulated cellular maintenance and repair mechanisms in response to various stressors, including DNA damage. Moreover, telomere shortening, measured here as lower than expected DNAmTL, is a known protective mechanism against cancer that can induce cellular senescence and genomic instability potentially reflective of a premalignant state in sun-exposed skin.6,7
Our observations support the hypothesis that epigenetic age dysregulation is a maladaptive response against the carcinogenic effect of sun exposure, which may persist throughout the malignant transformation of skin.8 Moreover, the link between epigenetic clock-based age dysregulation and the Helfrich photoageing scale, validated using patient photographs and correlated strongly with increased age and smoking status, supports a precision-medicine approach of integrating molecular and clinical scales to improve skin cancer risk stratification and to help guide photoageing treatment on an individual basis.
Contributor Information
Richie Jeremian, Faculty of Medicine and Health Sciences; Research Institute of the McGill University Health Centre, Montréal, QC.
Alexandra Malinowski, Department of Pharmacology and Toxicology, Temerty Faculty of Medicine.
Yuliya Lytvyn, Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON.
Jorge R Georgakopoulos, Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON.
Anastasiya Muntyanu, Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON.
Asfandyar Mufti, Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON.
Philippe Lefrançois, Division of Dermatology, Department of Medicine, McGill University, Montréal, QC; Lady Davis Institute, Jewish General Hospital, Montréal, QC, Canada.
Jensen Yeung, Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON.
Ivan V Litvinov, Research Institute of the McGill University Health Centre, Montréal, QC; Division of Dermatology, Department of Medicine, McGill University, Montréal, QC.
Funding sources
This work was supported by the Canadian Institutes of Health Research (CIHR) Project Scheme Grant #426655 to I.V.L.; a CIHR Catalyst Grant (#428712) to I.V.L.; a Cancer Research Society (CRS)–CIHR Partnership Grant (#25343) to I.V.L.; a Canadian Dermatology Foundation research grant to I.V.L.; by the Fonds de la recherche du Québec – Santé (#296643) to I.V.L.; and a Skin Investigator Network (SkIN) of Canada Trainee Travel Award to R.J.
Conflicts of interest
The authors declare no conflicts of interest.
Data availability
All available data are provided in the article.
Ethics statement
Not applicable.
References
- 1. Vandiver AR, Irizarry RA, Hansen KD. et al. Age and sun exposure-related widespread genomic blocks of hypomethylation in nonmalignant skin. Genome Biol 2015; 16:80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Yu M, Hazelton WD, Luebeck GE, Grady WM. Epigenetic aging: more than just a clock when it comes to cancer. Cancer Res 2020; 80:367–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Thrush KL, Higgins-Chen AT, Liu Z, Levine ME. R methylCIPHER: A methylation clock investigational package for hypothesis-driven evaluation & research. Available at: https://www.biorxiv.org/content/10.1101/2022.07.13.499978v1 (last accessed 11 January 2024).
- 4. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Stat Soc Ser B 1995; 57:289–300. [Google Scholar]
- 5. Helfrich YR, Yu L, Ofori A. et al. Effect of smoking on aging of photoprotected skin: evidence gathered using a new photonumeric scale. Arch Dermatol 2007; 143:397–402. [DOI] [PubMed] [Google Scholar]
- 6. Kabacik S, Lowe D, Fransen L. et al. The relationship between epigenetic age and the hallmarks of aging in human cells. Nat Aging 2022; 2:484–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Nassour J, Schmidt TT, Karlseder J.. Telomeres and cancer: resolving the paradox. Annu Rev Cancer Biol 2021; 5:59–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ogrodnik M, Gladyshev VN. The meaning of adaptation in aging: insights from cellular senescence, epigenetic clocks and stem cell alterations. Nat Aging 2023; 3:766–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All available data are provided in the article.