Methylation dynamics in the decades preceding acute myeloid leukaemia

Adriana V.a. Fonseca, Christopher Boniface, Caroline J. Watson, Sean R. Hackett, Calum Gabbutt, Akemi D Ramos-Yamasaki, Ah‐Hwee Tan, José M. Montoya, Ruslan Strogantsev, L. E. Hanssen, Y Q Zhang, Sophia Apostolidou, Aleksandra Gentry‐Maharaj, Sadik Esener, Paul T. Spellman, Trevor A. Graham, Usha Menon, Hisham Mohammed, Jamie R. BlundellbioRxiv 2025 · DOI

TL;DR

This study reveals that widespread DNA methylation changes occur in blood cells years, even decades, before acute myeloid leukemia (AML) diagnosis. These early epigenetic alterations are linked to the expansion of pre-leukemic clones and could serve as powerful biomarkers for early cancer detection and risk prediction.

What Problem Does This Paper Address?

Cancer development is a complex, multi-step process involving both genetic mutations and epigenetic alterations. While genetic drivers of cancer are known to accumulate over many years, the timing and nature of early epigenetic changes, specifically DNA methylation, in the decades preceding cancer diagnosis have remained largely unknown. Understanding these early epigenetic dynamics is crucial for developing effective early detection and risk stratification strategies. This paper specifically investigates these pre-cancerous methylation changes in the context of acute myeloid leukemia (AML), a blood cancer where epigenetic regulators are frequently mutated early in its development.

Methods

The researchers employed a multi-pronged approach:

  • Whole-Genome Bisulfite Sequencing (WGBS): Used to identify differentially methylated regions (DMRs) by comparing DNA methylation patterns in purified AML blasts and healthy hematopoietic stem cells (HSCs).
  • Deep Targeted Methylation Sequencing: A custom panel was designed to deeply sequence approximately 5,000 specific regions (including identified DMRs, epigenetic aging markers, and high methylation entropy regions) in longitudinal blood samples. This allowed for high-resolution tracking of methylation dynamics over time.
  • Enzymatic Methylation Sequencing (EM-seq): A library preparation method that enables high DNA input and unique molecular coverage, crucial for detecting subtle longitudinal changes.
  • Somatic Mutation Sequencing: Highly sensitive duplex error-corrected DNA sequencing was used to identify somatic mutations in AML-associated genes from the same longitudinal samples.
  • Statistical Modeling: Novel statistical methods were developed to identify DMRs from low-depth WGBS data and to associate specific methylation changes with clonal expansion dynamics. This included developing an F-statistic to test the significance of methylation changes driven by somatic clones.
  • CpG Lineage Tracing (fCpGs): A method was adapted to identify "fluctuating CpGs" (fCpGs) that act as stable epigenetic markers to trace clonal expansions, particularly those driven by genetic alterations not covered by the targeted sequencing panel.

Datasets

  • BLUEPRINT Consortium Data: Whole-genome bisulfite sequencing (WGBS) data from purified AML blasts and healthy hematopoietic stem cells (HSCs).
  • UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) Cohort: Longitudinal blood samples collected annually for up to 15 years prior to AML diagnosis from 50 women who later developed AML, and 50 age-matched controls. DNA was derived from serum.
  • Independent AML Cohort: Data from 119 AML samples (Glass et al.) used for validation of identified DMRs.

Key Insights and Ideas

  • Early and Widespread Epigenetic Reprogramming: Significant DNA methylation changes occur in blood cells years before AML diagnosis, indicating that epigenetic reprogramming is an early event in leukemogenesis, not just a late consequence.
  • "Epigenetic Drivers" vs. "Passenger" CpGs: The study distinguishes between methylation changes that actively contribute to cancer development ("epigenetic drivers") and those that change passively due to clonal expansion ("passenger" CpGs).
  • Neurodevelopmental Pathways Implicated: A surprising finding was the enrichment of AML-specific DMRs in genes associated with neurodevelopmental processes, suggesting a novel role for these pathways in early AML development.
  • CpG Lineage Tracing for Undetected Clones: "Fluctuating CpGs" (fCpGs) can serve as powerful, unbiased markers to identify clonal expansions, even when the underlying genetic drivers are not detected by standard sequencing. This has implications for detecting "cryptic" clonal hematopoiesis.
  • Epimutation Rate is High: The rate of spontaneous methylation changes ("epimutation") in hematopoietic stem cells is orders of magnitude higher than somatic mutation rates, contributing significantly to epigenetic diversity.

Key Results

  • Thousands of differentially methylated regions (DMRs) were identified, with many showing altered methylation patterns up to 10 years before AML diagnosis.
  • Over 99% of identified DMRs showed consistent methylation directionality in an independent AML dataset.
  • Hypermethylated DMRs were highly specific to AML, while hypomethylated regions often reflected normal differentiation patterns.
  • Longitudinal analysis revealed widespread, progressive methylation changes in pre-AML cases, contrasting with the stable methylation patterns in controls.
  • Tens of thousands of CpGs were found to be statistically linked to specific somatic clones, suggesting that methylation changes are inherited from the founder cell of clonal expansions.
  • An estimated epimutation rate of 0.1%–1% per CpG per year was determined for hematopoietic stem cells.
  • Approximately 1,600 AML-specific regions were identified as candidate "epigenetic drivers," recurrently altered in pre-AML donors.
  • The fCpG lineage tracing method could reliably identify clonal expansions with cell fractions above approximately 20%.

Limitations and Future Work

  • Distinguishing Drivers from Drift: While the study identifies candidate "epigenetic driver" regions, definitively proving their functional role in AML initiation remains challenging due to the high intrinsic epimutation rate and the complexity of methylation patterns. Future larger-scale studies in control populations will be crucial for rigorous validation.
  • Targeted Panel Limitations: The study acknowledges that important genetic and epigenetic events might be missed if they fall outside the targeted sequencing panels.
  • Functional Validation: While the study identifies pathways and regions, further functional studies are needed to elucidate the precise mechanisms by which these epigenetic alterations contribute to AML development.
  • Broader Applicability: The authors suggest that the principles and methods developed in this study, particularly for methylation-based biomarkers and lineage tracing, are likely applicable to other cancer types.

Summary updated

Abstract

DNA methylation is emerging as a highly sensitive and specific marker of cancer initiation and progression. How these cancer-specific methylation changes are established in the decades before cancer, however, remains largely unknown. Here, we use a unique collection of longitudinal blood samples collected annually up to 15 years prior to a diagnosis of acute myeloid leukaemia to sensitively track the dynamics of DNA methylation changes at high temporal resolution. We identify thousands of differentially methylated regions (DMRs) that exhibit altered patterns of methylation up to 10 years before acute myeloid leukaemia (AML) diagnosis. Most of these DMRs are strongly associated with expanding clones carrying somatic driver mutations. We identify a subset of ‘epigenetic driver’ DMRs characterised by recurrent CpG alterations that are highly shared across pre-AML cases. These are likely to reflect early epigenetic reprogramming associated with AML development. We also reveal large numbers of stochastic ‘passenger’ CpGs whose differential methylation results from hitch-hiking with clonal expansions driven by somatically acquired genetic events. These passenger CpGs can be exploited for lineage tracing to discover clonal expansions driven by missing driver mutations. Our findings show widespread changes in methylation patterns during the early stages of cancer development which could be utilised for risk prediction and therapeutic intervention.
BibTeX
@misc{fonseca2025methylation,
  author = {Adriana V.A. Fonseca and Christopher Boniface and Caroline J. Watson and Sean R. Hackett and Calum Gabbutt and Akemi D Ramos-Yamasaki and Ah‐Hwee Tan and José M. Montoya and Ruslan Strogantsev and L. E. Hanssen and Y Q Zhang and Sophia Apostolidou and Aleksandra Gentry‐Maharaj and Sadik Esener and Paul T. Spellman and Trevor A. Graham and Usha Menon and Hisham Mohammed and Jamie R. Blundell},
  title = {{Methylation dynamics in the decades preceding acute myeloid leukaemia}},
  year = {2025},
  doi = {10.1101/2025.06.26.661643},
  howpublished = {bioRxiv},
}
Figure