Integrative Multi-Omics Analyses Identify a High-Risk Subgroup of Patients with Poor Prognosis in t(8;21) Acute Myeloid Leukemia

Objectives: The t(8;21)(q22;q22) translocation is a common chromosomal abnormality in acute myeloid leukemia (AML). Although t(8;21) AML has a favorable prognosis, about 40% of the patients will eventually relapse. Therefore, it is crucial to improve current risk stratification and identify those with high risk of relapse as early as possible for improved treatment strategies and/or allo-HSCT in first remission. In this study, we tried to address this issue and provided novel insights into the classification strategy for prognosis evaluation in t(8;21) AML.

Methods: First, we applied non-negative matrix factorization to conduct clustering analysis on the gene expression profiles of 42 t(8;21) AML patients at our center and validated this on additional 46 t(8;21) AML samples from the HOVON cohort. Besides, we correlated the transcriptomic subtypes with treatment outcomes of patients and integrated targeted-sequencing data and flow cytometry results for combined analysis to delineate features of different subgroups. Furthermore, we performed single-cell RNA-sequencing and elucidated the regulatory patterns of the high-risk subgroup and revealed potential treatment targets.

Results: Three distinct transcriptional subgroups were identified based on transcriptomic profiling of 42 t(8;21) AML patients in our center. Cluster 1 (C1) was characterized by abnormal expression of leukemia stem and progenitor signatures. Cluster 2 (C2) displayed aberrant expression of genes associated with metabolic pathways and cluster 3 (C3) was highly enriched for interferon-related signaling. We identified similar three clusters from 46 t(8;21) AML patients of the HOVON cohort, suggesting the accuracy of our clustering strategy.

These subgroups exhibited significant differences in prognosis. Patients in C1 showed the poorest prognosis (p = 0.002), with 10 (62.5%) patients eventually experiencing relapse or remaining refractory throughout the treatment course, while only 3 (21.4%) patients had the similar situation in C2. All patients in C3 achieved complete remission, indicating that the activation of interferon related pathway may enhance treatment efficacy. Major molecular response (MMR) is defined as at least a 3-log reduction in RUNX1::RUNX1T1 transcript level, while complete molecular response (CMR) denotes the absence of detectable RUNX1::RUNX1T1. Here, our study revealed that only 1 (12.5%) patient in C1 achieved MMR compared with that of 4 patients in C2 (33.3%) and 4 patients in C3 (40%) after the initial induction therapy. And CMR was only observed in C2 (8.3%) and C3 (50%) within two courses of chemotherapy.

Additionally, we incorporated targeted sequencing data and identified that both KIT-D816 mutation (p < 0.001) and the occurrence of FLT3-ITD (p = 0.031) were enriched in C1 samples. And the reported unfavorable mutations ASXL2 and ZBTB7A also harbored higher proportions in C1 group. Besides, we included flow cytometry data and revealed a potential link between the CD7 expression and unfavorable prognosis of t(8;21) AML (p = 0.0289). Notably, the presence of CD7 was only enriched within C1, suggesting poorer treatment outcome of C1 patients.

Furthermore, we performed single-cell RNA-seq on patients classified as C1 and identified 9 cell clusters. Marker genes of C1 were predominantly present in the HSC-like subset, which was characterized by differentiation blockage and cell cycle arrest. Deconvolution analysis indicated that patients in C1 had a higher abundance of HSC-like subset, and patients with higher expression of HSC-like signatures conferred unfavorable prognosis. These suggested the crucial role of differentiation stage especially the most primitive HSC-like subset for transcriptome-based classification and prognosis prediction of t(8;21) AML.

Conclusions: We proposed a novel classification strategy based on multi-omics profiling that elucidated both the molecular and clinical variations observed in t(8;21) AML. Through this, we identified a high-risk subgroup characterized by a more primitive cellular hierarchy, higher KIT-D816 or FLT3-ITD mutation rates, poorer response to chemotherapy and worse prognosis, and elucidated the underlying regulatory mechanisms. Early identification at diagnosis and implementing enhanced treatment strategies might improve the outcomes for these high-risk patients.