DNA methylation-estimated phenotypes, telomere length and risk of ischemic stroke: epigenetic age acceleration of screening and a Mendelian randomization study

Aierpati Maimaiti, Jianhua Ma, Chenguang Hao, Dengfeng Han, Yongxin Wang, Zengliang Wang, Rena AbudusalamuAging 2024 · DOI

TL;DR

This study investigates the link between biological aging (as measured by DNA methylation clocks and telomere length) and the risk of ischemic stroke. While it found a significant association between accelerated epigenetic aging and ischemic stroke, particularly cardioembolic stroke, Mendelian randomization analyses did not establish a direct causal link between age phenotypes and stroke risk.

What Problem Does This Paper Address?

Chronological age is a well-established risk factor for ischemic stroke, but the underlying biological mechanisms are not fully understood. This paper delves into the concept of "biological age," specifically as estimated by DNA methylation (DNAmAge) and telomere length, to explore its relationship with ischemic stroke. The authors aim to determine if accelerated biological aging, as opposed to chronological aging, is a significant contributor to stroke risk and if there's a bidirectional causal relationship between these factors. Understanding these connections could lead to better stroke prevention and treatment strategies.

Methods

The study employed two primary analytical approaches:

  • Methylation Clock Analysis: They utilized established epigenetic clocks (specifically mentioning Horvath's clock) to calculate DNAmAge for each patient. Age acceleration was then quantified by comparing DNAmAge to chronological age, both as a simple difference (Age Acceleration Diff) and as residuals from a regression model (Age Acceleration Residual) to account for chronological age's effect.
  • Mendelian Randomization (MR): This epidemiological technique uses genetic variants (SNPs) as instrumental variables to infer causal relationships between an exposure (e.g., age phenotypes, stroke) and an outcome (e.g., stroke, telomere length). They performed two-sample bidirectional MR analyses, investigating both the effect of age phenotypes on stroke and the effect of stroke on age phenotypes. Various MR methods (Inverse Variance Weighted, MR-Egger, Weighted Median, etc.) were used to assess robustness and detect pleiotropy.

Datasets

The study utilized several datasets:

  • Ischemic Stroke Cohort: Data from 185 patients diagnosed with ischemic stroke, who underwent whole-genome methylation sequencing of whole blood. This cohort was categorized into atherothrombotic stroke (109 cases), cardioembolic stroke (18 cases), and small-vessel stroke (58 cases).
  • In Silico DNA Methylation Data: Methylation status of 353 CpG sites related to age was analyzed using Illumina BeadChip450K methylation array data.
  • GWAS Summary Statistics:
    • For epigenetic age acceleration: Data from a meta-analysis of GWAS on biological aging, including 34,710 European descent individuals and 6,195 African American descent individuals.
  • For telomere length: Genetic association estimates for epigenetic telomere length from the UK Biobank (472,174 participants).
  • For stroke: Data from a multiancestry genome-wide association meta-analysis of 521,612 individuals (67,162 cases and 454,450 controls) for stroke and its common etiological subtypes (large-artery atherosclerotic stroke, cardioembolic stroke, and small-vessel disease stroke).

Key Insights and Ideas

  • Cardioembolic Stroke Stands Out: The most striking finding was the significant age acceleration observed in patients with cardioembolic stroke, showing a median difference of 9 years between their DNAmAge and chronological age. This suggests a particularly strong link between this stroke subtype and accelerated biological aging.
  • Gender Differences in Aging: Females exhibited a statistically significant greater degree of age acceleration compared to males, primarily driven by cardioembolic stroke. This highlights potential sex-specific pathways in the aging-stroke relationship.
  • Stroke's Impact on Telomeres: The MR analysis indicated a negative causal effect of stroke (and ischemic stroke specifically) on telomere length. This implies that experiencing a stroke might contribute to the shortening of telomeres, a marker of cellular aging.
  • No Direct Causal Link from Age to Stroke: Despite the associations, the MR analyses did not find evidence for a direct causal effect of age phenotypes (DNAmAge, telomere length) on the incidence of stroke or its subtypes. This suggests that while biological aging markers are associated with stroke, they might not be direct drivers in a causal sense, or the current MR methods and data are insufficient to detect such links.
  • Potential Causal Link from Stroke to PhenoAge: There was some indication of a potential causal effect of ischemic stroke on PhenoAge acceleration, suggesting that stroke might, in turn, hasten a specific measure of biological aging.

Key Results

  • Age Acceleration in Stroke Patients: DNAmAge was observed to be higher than chronological age in the ischemic stroke cohort.
  • Stroke Subtype Differences: Cardioembolic stroke patients showed the most significant age acceleration, followed by small-vessel and atherothrombotic stroke.
  • Gender Disparity: Females showed higher age acceleration, particularly in cardioembolic stroke cases.
  • MR Results:
    • Stroke had a negative causal effect on telomere length (OR = 0.927 for all stroke, P = 0.08; OR = 0.934 for ischemic stroke, P = 0.019).
  • No significant causal relationship was found between age phenotypes and stroke risk.
  • A modest positive causal effect of ischemic stroke on PhenoAge acceleration was observed using the Weighted Median method (OR = 1.838, P = 0.048).

Limitations and Future Work

The authors acknowledge several limitations:

  • Sample Size: The potential insufficiency of sample sizes in available GWAS data may have hindered the detection of weak causal effects.
  • Survivorship Bias: The study design might have introduced survivorship bias, as it only included individuals who survived until stroke ascertainment.
  • Unmeasured Confounding: Despite MR techniques, unmeasured or residual confounding could still influence the findings.
  • Limited Stroke Subtypes: The analysis focused on prevalent stroke subtypes, potentially missing causal associations in more specific or rarer categories.
  • Generalizability: The limited representation of non-European samples in the study might restrict the generalizability of the results.

Future work should focus on:

  • Investigating Biological Mechanisms: Further research is needed to understand the biological mechanisms underlying the associations, particularly the gender disparities and the link between cardioembolic stroke and accelerated aging.
  • Larger Sample Sizes: Utilizing larger sample sizes and more precise biological age assessments in MR studies is crucial to confirm or refute potential causal relationships.
  • Lifestyle and Intervention Studies: Exploring the potential ameliorative impact of lifestyle modifications and pharmacological interventions on the relationship between accelerated aging and stroke risk is warranted.
  • Developing Screening Tools: The identified DNA methylation sites and telomere length associated with stroke risk could be valuable for developing new screening tools for early detection of high-risk individuals.

Summary updated

Abstract

The study provides insight into the relationship between DNAmAge and ischemic stroke, particularly cardioembolic stroke, and suggests possible gender differences. These insights carry profound clinical significance and set stage for future investigations into the entwined pathways of stroke and accelerated aging.
BibTeX
@inproceedings{maimaiti2024dna,
  author = {Aierpati Maimaiti and Jianhua Ma and Chenguang Hao and Dengfeng Han and Yongxin Wang and Zengliang Wang and Rena Abudusalamu},
  title = {{DNA methylation-estimated phenotypes, telomere length and risk of ischemic stroke: epigenetic age acceleration of screening and a Mendelian randomization study}},
  booktitle = {Aging},
  year = {2024},
  doi = {10.18632/aging.206072},
}
Figure