Background

Acute Myeloid Leukemia (AML) is a life-threatening hematologic malignancy characterized by uncontrolled proliferation of malignant myeloid blood cells. The medical community has been tracking genetic mutations in the deoxyribonucleic acid (DNA) sequence of cancer cells to predict patient prognosis and determine treatment strategies. The current European LeukemiaNet (ELN) guidelines also classify patients into risk groups based on the presence of specific gene mutations.

However, it is impossible to perfectly predict the varying drug responses and prognoses among patients based solely on genetic mutations. Even patients with the same mutations showed significantly different responses to chemotherapy or targeted therapies, which has been pointed out as a clear limitation of existing genome decoding technologies. The diversity of the epigenome, a higher-level system that regulates gene expression, which does not change the sequence itself, has been identified as a key variable determining the actual properties of leukemia cells. Therefore, the academic community has continued to make efforts to analyze the three-dimensional structural changes of chromatin, which represents how DNA is folded and unfolded within cells, on a large scale, beyond simple changes in gene sequences.

Key Findings

A multinational collaborative research team, including Yotaro Ochi, Assistant Professor, and Seishi Ogawa, Professor, Department of Medicine, Kyoto University, and Sören Lehmann, Professor, Karolinska Institutet, utilized clinical samples from 1,563 AML patients for analysis. This research team established the 'eCHROMA AML (Encyclopedia of Chromatin in AML),' the largest-ever database of cancer epigenomes.

They mapped the active regions of gene regulation within cells using an Assay for Transposase-Accessible Chromatin with sequencing (ATAC-seq) to measure the chromatin opening state of leukemia cells. Analysis using an artificial intelligence-based clustering algorithm revealed that AML patients can be classified into 16 distinct subgroups with different epigenetic characteristics. These 16 groups showed unique transcriptional factor networks and DNA methylation patterns according to the opening and closing patterns of chromatin structure, and the distribution of super-enhancers, which are strong gene regulatory regions, also showed different patterns.

To validate the reliability of the identified epigenetic subgroups, the research team simultaneously applied single-cell RNA sequencing (scRNA-seq) and single-cell ATAC sequencing (scATAC-seq) to more than 280,000 leukemia cells obtained from 36 patients. The analysis confirmed that the unique chromatin states of the 16 groups are consistently maintained across individual cancer cell populations. At the same time, the research team also developed a simple diagnostic tool to aid in clinical application. They successfully extracted a 30-gene expression signature that can identify high-risk epigenetic subgroups using only standard RNA sequencing (RNA-seq) data, without having to go through complex and expensive ATAC-seq experiments each time.

Significance and Prospects

This discovery marks a turning point in redefining AML not simply as a result of genetic mutations, but as an epigenetic disease dominated by structural abnormalities in chromatin. The newly defined 16 subgroups serve as prognostic indicators that accurately distinguish high-risk patient groups that could not be differentiated by conventional genomic sequence analysis alone. It is now possible to explain the differences in drug sensitivity among patients at a more three-dimensional epigenetic level, beyond the gene level.

However, there are still hurdles to overcome before this diagnostic method can be established in standard clinical practice. The diagnostic accuracy of the 30-gene signature needs to be further validated in actual clinical settings, and clinical trials of customized therapies optimized for each epigenetic subgroup should follow. Even though the data of more than 1,500 multinational patients were used, additional multinational validation studies are also essential to fully encompass the world's racial diversity.