Your DNA May Appear Older Than You Think

The difference between chronologic age and inferred “biological age,”—also known as age acceleration—is related with breast cancer risk and represents progress for molecular measures that may enhance risk assessment approaches in the future. The application of epigenetic clocks to infer subject age using measures of DNA methylation has utility for investigating the biology of aging and aging as a risk factor for cancer. Epigenetic clocks estimate chronologic age using sets of CpG sites that have been previously defined by regression of DNA methylation on age in large numbers of subjects and in a tissue-specific or tissue-agnostic manner. Age clocks using DNA methylation are established as being more accurate age estimators than other molecular-based approaches, potentially due to the stability of both DNA and of cytosine methylation, which is a covalent modification. Alterations to DNA methylation occur early in neoplastic transformation and have been widely observed across tumor types, firmly establishing a premise for identifying and exploiting DNA methylation biomarkers of cancer risk.

A fundamental component—and utilitarian feature—of epigenetic clocks is that the residual between known subject chronologic age and the estimated age can be thought to represent biological age, or age acceleration, which as a new variable can then be tested for association with other variables or outcomes of interest. In this issue of the Journal, Kresovich et al. (1) employ multiple DNA methylation age clocks to subjects in the Sister Study prospective breast cancer cohort and report strong evidence that biological age acceleration increases the odds of breast cancer. The use of a time-to-event approach, adjustment for variation in leukocyte cell type proportions, and inclusion of analyses stratified on menopausal status and disease severity provide new information about how biological aging inferred from readily accessible system-level peripheral blood contributes to breast cancer risk.

In their work, Kresovich et al. (1) measured DNA methylation in blood and applied more than one epigenetic clock, which adds robustness and underscores both the still-evolving nature of age estimators and their subtle distinctions. Variation in estimated methylation age from different epigenetic clocks can be attributable to different input data used for calibration, whether single vs multiple tissue types, and/or different measurement platforms. Kresovich et al. used the Hannum clock (2) and the Levine clock (3), which were developed using DNA methylation data from peripheral blood—though the more recent Levine clock included both repeated measures and other clinical phenotypic aging measures from subjects in the National Health and Nutrition Examination (NHANES)—to select CpGs for age estimation. In addition, the authors tested methylation age from the Horvath clock, which was developed using a large number of data sets across many distinct tissue types (4). Specifically relevant to breast cancer, in Horvath’s clock, there was particularly poor calibration for breast tissue where age acceleration was observed across multiple data sets. Some of the breast tissue data sets were tumor-adjacent normal samples from surgical resection. Remarkably though, in breast tissue from disease-free women, results were consistent where an approximate 9-year age acceleration was observed on average. Age acceleration determined using Horvath’s DNA methylation age method in normal breast tissue from disease-free subjects has been replicated in Johnson et al. (5), where results also suggested even higher age acceleration among African American subjects.

Given the cell-type-specific nature of DNA methylation patterning, there is clear advantage to striving for breast-specific estimates for age acceleration in studies of breast cancer. However, obtaining breast-specific estimates of age acceleration from biopsy tissue is invasive and infeasible for routine use in epidemiologic cohorts and for eventual risk assessment approaches applied to populations. Nipple aspirate fluid is an alternative to biopsy for obtaining a breast-specific biospecimen, and measuring molecular features therein has an established premise (6,7), though challenges in collection, substrate yield, and representation of the whole organ are limitations to consider. Human milk is another noninvasive option to collect DNA from shed cells and provides ample substrate. Indeed, DNA methylation alterations relevant to breast cancer have been reported in human milk (8,9). Epidemiological studies that include collections of human milk (typically birth cohort studies) are underway and poised to investigate age acceleration in relation with established and putative breast cancer risk factors. Long-term follow-up of subjects in those cohorts for breast cancer outcomes, in conjunction with targeted recruitment of women with high breast cancer risk, are needed. Though there is a limited window in which to collect milk, its availability proximal to childbirth (of course a breast cancer risk modifier) may itself provide insight into mechanistic underpinnings of risk.

More generally, investigations of the relationship between age acceleration using epigenetic clocks and cancer risk in epidemiological studies with archival blood specimens present a multitude of exciting research opportunities. Questions regarding the biology of aging itself and the extent to which epigenetic age is determined by genetic vs environmental and/or behavioral factors remain open. The enrollment criterion in the Sister Study prospective cohort is that women do not have breast cancer but have a biological sister with a breast cancer diagnosis, where hypotheses on genetic determinants of epigenetic age acceleration may begin to be addressed. On environmental and behavioral factors, initial observations of associations between increased body mass index and epigenetic age acceleration in liver tissues (10) is suggestive of potential utility in predicting early onset liver cancer and may also be relevant to breast cancer. Both smaller, and large population and longitudinal studies with detailed data from diverse subjects that integrate epigenetic measures can address research opportunities in this space. For instance, repeated measures to estimate epigenetic age proximal to times when breast cancer-modifying life events occur such as before and after giving birth and beyond. Further, relationships of age acceleration with dietary intake, alcohol use, and physical exercise are of interest and are beginning to be investigated (11). The findings from Kresovich et al. (1) and emerging work on biological age acceleration suggest high potential impact for future work on biological aging and cancer risk assessment for breast and other cancer types. In-depth studies of both established and putative breast cancer risk factors for their relationship with epigenetic age acceleration can add to our understanding of the biology underlying disease risk factors and present new opportunities for primary and secondary prevention of breast cancer.

Notes

Affiliations of author: Department of Epidemiology (BC), Department of Molecular and Systems Biology (BC), and Department of Community and Family Medicine (BC), Geisel School of Medicine at Dartmouth, Lebanon, NH.

The author has no conflicts of interest to disclose.