Biological Aging Measures Based on Blood DNA Methylation and Risk of Cancer: A Prospective Study

Pierre‐Antoine Dugué, Julie K. Bassett, Ee Ming Wong, Jihoon E. Joo, Shuai Li, Chenglong Yu, Daniel F. Schmidt, Enes Makalic, Nicole Wong Doo, Daniel D. Buchanan, Allison Hodge, Dallas R. English, John L. Hopper, Graham G. Giles, Melissa C. Southey, Roger L. MilneJNCI 2020 · DOI

Abstract

While chronological age is a primary risk factor for cancer, it fails to account for individual variations in the aging process. This study evaluates the association between "second-generation" epigenetic aging measures—PhenoAge, GrimAge, and methylation-predicted telomere length—and the risk of seven different cancer types. Utilizing data from the Melbourne Collaborative Cohort Study (MCCS), researchers found that PhenoAge and GrimAge are significant predictors of cancer risk, with GrimAge showing a particularly strong association with lung cancer. These findings suggest that epigenetic clocks trained on clinical biomarkers and mortality are superior to first-generation clocks in forecasting oncological outcomes.

Introduction

The discrepancy between a person’s chronological age and their physiological state is known as biological aging. To quantify this, scientists utilize DNA methylation—chemical modifications to DNA that change predictably over time. Early "first-generation" epigenetic clocks were designed to predict chronological age but often lacked the sensitivity to predict disease onset or mortality.

This research focuses on "second-generation" clocks, which were developed using health-related biomarkers and mortality data rather than just chronological time. By assessing PhenoAge and GrimAge, this study aims to determine if these advanced measures can serve as effective tools for identifying individuals at elevated risk for various cancers.

Methodology

The researchers conducted a prospective, nested case-control study within the Melbourne Collaborative Cohort Study (MCCS), a longitudinal dataset of over 41,000 participants recruited in the early 1990s. The analysis included 3,117 individuals who developed cancer and matched healthy controls.

Three primary methylation-based measures were calculated from blood samples:

  1. PhenoAge: Trained on chronological age and nine clinical biomarkers (e.g., albumin, glucose, C-reactive protein).
  2. GrimAge: Trained on smoking history (pack-years) and seven plasma proteins associated with mortality.
  3. Methylation-predicted telomere length (DNAmTL): A DNA-based estimate of the protective caps on chromosomes.

Statistical analysis was performed using conditional logistic regression to estimate Rate Ratios (RRs) per standard deviation (SD) increase in biological age, adjusting for potential confounders such as BMI, alcohol consumption, physical activity, and detailed smoking history.

Key Results

General Cancer Risk

For the majority of cancers studied, every one SD increase in PhenoAge or GrimAge was associated with a 12% to 13% increase in overall cancer risk. This confirms that accelerated biological aging is a systemic indicator of cancer susceptibility.

The Lung Cancer Connection

GrimAge emerged as a remarkably potent predictor for lung cancer, showing a Rate Ratio of 1.82 per SD. Even after rigorous adjustment for self-reported smoking status and pack-years, the association remained strong. This suggests that GrimAge captures the cumulative biological damage caused by smoking and environmental exposures that traditional questionnaires might miss.

Broad Associations across Cancer Types

PhenoAge demonstrated significant associations (RRs ~1.2–1.3) across a wide spectrum of malignancies, including:

  • Colorectal cancer
  • Kidney cancer
  • Lung cancer
  • Mature B-cell neoplasms
  • Urothelial cancer

The Telomere and Prostate Paradoxes

Interestingly, methylation-predicted telomere length (DNAmTL) showed no significant association with cancer risk, suggesting that epigenetic aging and telomere attrition may represent distinct biological pathways. Additionally, GrimAge showed a slight inverse association with non-aggressive prostate cancer (RR=0.79), a finding that warrants further investigation into the specific metabolic drivers of prostate tissue.

Discussion and Insights

The "Second-Gen" Advantage

The study highlights that PhenoAge and GrimAge significantly outperform earlier models like the Horvath clock. By training algorithms on health outcomes and mortality-linked proteins, these clocks function as more accurate "biological odometers," reflecting the actual wear and tear on the body’s systems.

Clinical Utility and Screening

The association between biological age and cancer was most pronounced within the five years following the blood draw. This indicates that these measures could be particularly valuable in clinical settings to identify high-risk individuals who may benefit from more frequent or intensive screening protocols.

Limitations and Future Directions

While the study provides robust evidence, several limitations remain:

  • Tissue Specificity: The measures were derived from blood DNA. Future research should explore whether tissue-specific epigenetic aging (e.g., in the colon or lungs) provides even greater predictive power.
  • Diversity: The findings from the MCCS cohort need replication in more diverse global populations to ensure generalizability.
  • Causality: It is currently unclear whether accelerated epigenetic aging is a direct cause of cancer or a marker of the cellular degradation that precedes it.

In conclusion, second-generation epigenetic clocks like PhenoAge and GrimAge offer a promising new frontier for personalized medicine, providing a more nuanced understanding of cancer risk than chronological age alone.

Summary updated

Abstract

The methylation-based measures <i>PhenoAge</i> and <i>GrimAge</i> may provide insights into the relationship between biological aging and cancer and be useful to predict cancer risk, particularly for lung cancer.
BibTeX
@inproceedings{dugue2020biological,
  author = {Pierre‐Antoine Dugué and Julie K. Bassett and Ee Ming Wong and Jihoon E. Joo and Shuai Li and Chenglong Yu and Daniel F. Schmidt and Enes Makalic and Nicole Wong Doo and Daniel D. Buchanan and Allison Hodge and Dallas R. English and John L. Hopper and Graham G. Giles and Melissa C. Southey and Roger L. Milne},
  title = {{Biological Aging Measures Based on Blood DNA Methylation and Risk of Cancer: A Prospective Study}},
  booktitle = {JNCI},
  year = {2020},
  doi = {10.1093/jncics/pkaa109},
}
Figure