Skip to main content

Hidden Circular DNA Fragments Linked to Relapse in Childhood Leukaemia

Date

How data analytics helped uncover a surprising factor in cancer relapse

Summary

Researchers at the University of Leeds, working with national and international partners, have made a groundbreaking discovery about the hidden role of tiny circular pieces of DNA in childhood leukaemia. These fragments, known as excised signal circles (ESCs), were once believed to be harmless by-products of normal immune cell development. However, through a combination of advanced laboratory work and data-driven analysis, the team found that ESCs can persist and even replicate within leukaemia cells. Their presence was linked to a higher risk of relapse in children with leukaemia, revealing a potential new biomarker to identify vulnerable patients and guide future treatments.

Background

Childhood B-cell precursor acute lymphoblastic leukaemia (BCP-ALL) is one of the most common childhood cancers, and while treatment outcomes have improved dramatically over recent decades, relapse remains a major challenge. When the disease returns, it is often more resistant to therapy and more difficult to cure.

In healthy immune cells, a natural process called V(D)J recombination rearranges segments of DNA to produce the vast diversity of antibodies needed to fight infection. This process leaves behind small circular DNA fragments known as excised signal circles (ESCs). For years, scientists assumed these circles quickly disappeared, serving no biological purpose once created.

Recent research into other cancers, however, revealed that similar circular DNA fragments — known as extrachromosomal DNA (ecDNA) — can play active roles in tumour growth and therapy resistance. This prompted the Leeds team to explore whether ESCs might also persist in leukaemia cells and influence the likelihood of relapse.

Aim

The project set out to investigate whether ESCs can replicate and persist in immune and cancer cells, and whether their presence is linked to relapse in childhood leukaemia. The researchers also sought to understand how ESCs might interact with the genetic machinery of the cell to drive disease recurrence. In doing so, the team aimed to uncover new biological mechanisms that could eventually inform diagnostic and therapeutic advances.

Method

To answer these questions, the researchers combined detailed laboratory experiments with powerful computational data analysis. They tracked ESCs during normal B-cell development in mice and examined samples from children with BCP-ALL, using advanced DNA sequencing to detect the molecular traces of ESC formation and replication.

Bioinformatics and computational modelling were then employed to quantify ESC abundance across different samples, explore associations between ESC levels, genetic mutations, and patient outcomes, and determine how ESC activity affects key cellular pathways. Gene expression analysis helped reveal whether cells rich in ESCs showed distinct patterns of DNA repair or replication activity.

This multidisciplinary approach — uniting molecular biology, genomics, and data science — allowed the team to detect patterns that would otherwise have remained invisible. Expertise from the Leeds Institute for Data Analytics (LIDA), led by Professor David Westhead, played a crucial role in analysing the large, complex datasets that underpinned these findings.

Challenges

Detecting ESCs proved technically challenging because of their extremely small size and fleeting nature. They can easily be overlooked in conventional sequencing data, so the team developed custom analytical tools to identify and quantify them accurately. Demonstrating that ESCs could persist and replicate required the creation of novel computational models and rigorous testing to rule out artefacts.

Interpreting these findings in a clinical context was another major hurdle. Establishing a robust link between ESC levels and relapse risk meant analysing vast amounts of genomic and clinical data from multiple patient cohorts. The team’s success in overcoming these challenges highlights the value of integrating biological expertise with advanced data analytics to uncover hidden drivers of disease.

Findings

The study revealed that ESCs are far from inert remnants of normal immune processes. Instead, they can persist within cells and even replicate over successive generations. This persistence suggests they may have an active biological role, rather than being mere by-products.

Crucially, higher levels of ESCs at diagnosis were strongly associated with relapse in children with BCP-ALL. Cells with abundant ESCs also showed increased expression of genes involved in DNA repair and replication, suggesting that ESCs may interfere with or exploit the cell’s maintenance machinery. Over time, this could promote genetic instability, allowing mutations to accumulate and fuelling the re-emergence of the disease.

Together, these findings provide the first direct evidence that ESCs contribute to the biological processes driving leukaemia relapse.

Conclusion

This research overturns long-standing assumptions about excised signal circles, demonstrating that they can persist, replicate, and influence the progression of childhood leukaemia. Clinically, ESC levels could serve as an early biomarker to identify patients most at risk of relapse, enabling doctors to tailor treatment and monitoring accordingly. For researchers, these DNA circles represent a new and potentially targetable feature of cancer biology.

From a data science perspective, the study showcases how advanced computational analysis can uncover subtle but crucial signals within complex genomic data. It underscores the Leeds Institute for Data Analytics’ mission to transform raw data into insights that improve patient outcomes. By bridging computational and biological research, the team has opened a new chapter in understanding — and ultimately preventing — leukaemia relapse in children.

Reference:

Gao, Z., Scott, J.N.F., Edwards, M.P. et al. Excised DNA circles from V(D)J recombination promote relapsed leukaemia. Nature 645, 774–783 (2025).
Read the full article