Cancer research has come a long way from killing malignant tumours or stopping the advance of the disease to entirely reversing the state of the cell into normal ones, as illustrated by a breakthrough in a Korean lab ast year.
Building on this work on cancer cell reversal by Korean scientist Kwang-Hyun Cho, a research team led by him has now identified a "switch", a more precise point for such an intervention, according to their new report published in the journal Advanced Science.
Cell fate changes involve abrupt transitions, or “critical transitions”, particularly in tumorigenesis. Understanding the molecular regulatory networks underlying these transitions may enable cancer reversion — reprogramming cancer cells to normal states, the researchers from Korea Advanced Institute of Science and Technology stated in the report.
Early studies relied on static analyses, limiting insights into dynamic transitions. Advances in single-cell transcriptomics and pseudotime analysis now help identify transition genes and model regulatory networks.
However, tumorigenesis remains difficult to model due to genetic alterations. Focusing on ‘tipping points’ rather than entire trajectories may better capture key transitions. Recent research highlights attractor states and gene expression heterogeneity, crucial for understanding tumorigenesis and potential therapeutic interventions.
The team developed REVERT (REVERse Transition), a novel systems framework designed to identify reversion switches in tumorigenesis by analysing the ‘attractor landscape’ of dynamic gene regulatory networks (GRN). Using single-cell transcriptomic data from cancer patients, REVERT employs Boolean network modeling to map the transition from normal to cancer cell states, the authors shared in the report published January 22, 2025.
Attractor landscape analysis is a computational technique used to model and visualise the stable states, or attractors, that a biological system can adopt. In cancer research, this method helps scientists understand critical transitions between normal and malignant cellular states. By mapping these transitions, researchers can pinpoint molecular events that drive tumour progression or, potentially, tumour reversion.
REVERT enables the identification of key transcription factors (TF) that can reverse cancer by minimising a malignancy score. It reconstructs dynamic GRN models within tumour transition states, capturing the stochastic nature of cell state changes.
It aims to predict optimal intervention targets for reversing cancer progression and can be applied broadly to cell fate transitions in differentiation, development and cancer cell reprogramming.
The REVERT framework consists of four key steps:
• Identification of the transition state: REVERT clusters tumour and normal cells to define a transition state. This analysis helps distinguish normal-like, tumour-like and transitional subclones.
• Dynamic network model reconstruction: A pseudotime analysis tracks differentially expressed genes along the transition from normal to tumour states. By integrating these temporal changes with known gene regulatory interactions, Boolean logic functions are constructed to model transcription factor dynamics within strongly connected components.
• Cancer score (CS) computation: REVERT introduces a cancer score to quantify attractor landscape stability, measuring the relative dominance of normal and tumour cell states. A lower CS indicates a state closer to a normal phenotype.
• Target identification for cancer reversion: REVERT predicts key transcription factors whose modulation reduces CS, guiding therapeutic interventions.
By integrating computational modeling with experimental validation, REVERT provides a powerful tool for identifying cancer reversion targets and advancing precision oncology.
Developing a systems framework for controlling cell states based on omics data (compilation of data resulting from the study of various ‘omes’ of an organism, such as genome, transcriptome, proteome, metabolome and epigenome) has been a significant challenge in various biological fields, including drug discovery, stem cell engineering and regenerative medicine. This study successfully leverages this dataset.
Future directions for REVERT include integrating multiomics single-cell sequencing data to refine tumour transition state identification. By combining genomic and transcriptomic information, REVERT can offer even greater insights into molecular mechanisms underlying tumour progression and potential therapeutic interventions for cancer reversion.