A research team including Professor Seishi Ogawa (ASHBi Principal Investigator) and Assistant Professor Yotaro Ochi at the Graduate School of Medicine, Kyoto University, along with Professor Sören Lehmann at the Karolinska Institutet in Sweden, performed large-scale epigenomic analysis on more than 1,500 acute myeloid leukemia (AML) patient samples. As a result, they demonstrated that AML is classified into 16 distinct groups with different characteristics based on epigenomic properties. Furthermore, each group possessed different molecular mechanisms, clinical prognoses, and drug sensitivities, revealing AML diversity that cannot be fully captured by conventional genetic mutations alone.
Ogawa stated: "For example, in Group K, which has a poor prognosis with a cure rate of only 30%, we found high sensitivity to ABL kinase inhibitors. Confirming these findings in clinical trials may lead to the development of new treatments. Since we are making this data publicly available, we hope many researchers will utilize it to help decipher leukemia." The study was published in Nature.
AML is a hematologic malignancy characterized by the abnormal proliferation of undifferentiated myeloid cells derived from hematopoietic stem and progenitor cells. Genetic abnormalities occurring in hematopoietic stem and progenitor cells are thought to be the cause. Its diversity has also been classified by differences in genomic mutations. However, even among cases with identical genomic mutations, variations in treatment efficacy sometimes occurred.
In this study, using the ATAC-seq method, which enables comprehensive analysis of open chromatin regions, the research team analyzed the chromatin state of 1,563 AML samples. As a result, they clarified that AML can be classified into 16 characteristic subgroups (A to P) based on chromatin state. Furthermore, by performing single-cell analysis of gene expression and chromatin state targeting thousands of leukemia cells per sample, they confirmed that these subgroups are identified by characteristic chromatin states common to leukemia cells within each group.
These subgroups possessed distinct genetic mutations, differentiation states, gene expression patterns, DNA methylation patterns, and transcriptional regulatory networks. Notably, many subgroups did not fully align with existing classifications based on genomic abnormalities, indicating the presence of new AML diversity uncaptured by conventional genomic analysis alone.
Integrated analysis of epigenomic data revealed that different transcription factor networks and super-enhancer structures were formed in each subgroup. The presence of unique gene regulatory mechanisms defining the characteristics of leukemia cells is demonstrated.
This classification based on chromatin state was also strongly associated with patient prognosis. Adding chromatin information to the existing genome-based risk classification (ELN classification) improved prognostic accuracy. Furthermore, 200 drugs, including approved agents and candidate therapeutic compounds, were applied to 112 cell samples spanning 16 subtypes, and drug sensitivity was analyzed. In some subgroups, specific sensitivity was observed to molecularly targeted agents such as the ABL inhibitors bosutinib and imatinib and MEK inhibitors, none of which had been identified as candidate therapeutics under the ELN classification. This demonstrated potential application for novel stratified therapies.
This study is the world's first large-scale research showing that not only genetic mutations but also epigenomic changes in chromatin states are crucial for understanding the onset mechanisms and characteristics of leukemia. Such epigenomic information is expected to be applied in next-generation precision medicine for diagnosis, prognosis prediction, and drug selection in the future.
The constructed large-scale multi-omics database will serve as a key foundational resource in cancer epigenomics research, including AML, and is expected to aid in elucidating novel therapeutic targets and onset mechanisms.
On the other hand, because this study is primarily observational, further experimental validation is needed regarding the functional significance of the transcription factor networks, super-enhancer structures, and drug sensitivities characteristically observed in each subgroup. Technical and cost challenges remain for immediate application of chromatin state-based classification to routine clinical practice. It is necessary to proceed with the development of simple, low-cost diagnostic methods and the exploration of optimal treatments.
Journal Information
Publication: Nature
Title: Chromatin landscape and epigenetic heterogeneity of acute myeloid leukaemia
DOI: 10.1038/s41586-026-10703-4
This article has been translated by JST with permission from The Science News Ltd. (https://sci-news.co.jp/). Unauthorized reproduction of the article and photographs is prohibited.