Abstract Acute myeloid leukemia (AML) is a cancer of the myeloid cell lineage characterized by acquired gene mutations and chromosomal abnormalities. We have previously shown that AML initiation and progression can be predicted by applying a state-transition theory to the analysis of time-series messenger RNA (mRNA) and microRNA (miRNA) sequencing data from peripheral blood mononuclear cells (PBMCs) collected during disease development. We showed that AML initiation and progression in mouse models can be tracked via blood mononuclear cell transcriptomes and modeled as particle undergoing Brownian motion in a double-well potential with critical points for health, transition, and leukemia. Building on this framework, our current study evaluates chemotherapy response using time-series sequencing data from an AML mouse model. We plotted mRNA and miRNA transcriptome trajectories over time and found that chemotherapy initially shifted both trajectories away from the leukemic state toward the healthy state, followed by relapse back toward leukemic state. Notably, miRNA responses showed a delay of 2 weeks behind mRNA. To investigate the cause of delayed miRNA transcriptomic response to chemotherapy, we performed Weighted Gene Co-expression Network Analysis (WGCNA) on miRNA expression profiles to identify co-expressed clusters. By projecting each miRNA cluster onto the leukemia state-spaceYK1 , we identified a cluster that exhibit strong contribution to the AML state transition. This is a cluster of 30 miRNAs with coordinated expression patterns, significantly upregulated and associated with delayed transcriptomic response post-chemotherapy.Notably, 80% of these miRNAs are located within the DLK1-DIO3 imprinted region (chromosome 14q32 in humans, 12qF1 in mice), a locus previously implicated in stress response, acute promyelocytic leukemia (APL), solid tumors, and type 2 diabetes. However, their involvement in standard-of-care chemotherapy responses in AML has not been previously reported.Further mechanistic analysis revealed a regulatory mechanism in which high expression of these miRNAs suppresses the PI3K/mTOR signaling pathway. Inhibiting the PI3K/mTOR pathway lowers reactive oxygen species (ROS) production, which creates an environment for leukemia stem cells (LSC) to survive. Since chemotherapy primarily eliminates leukemic blast cells, the persistence of LSCs may significantly contribute to relapse.Our findings highlight the power of state-space modeling to uncover dynamic regulatory mechanisms and identify miRNA-driven factors that may lead to AML relapse after chemotherapy. Citation Format: Ziang Chen, Jennifer Rangel Ambriz, Yu-Hsuan Fu, David Eugene Frankhouser, Denis O'Meally, Lianjun Zhang, Ying-Chieh Chen, Sergio Branciamore, Jihyun Irizarry, Bin Zhang, Guido Marcucci, Russell Rockne, Ya-Huei Kuo. Identification of microRNA clusters as predictors of chemotherapy response in acute myeloid leukemia abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 5454.
Chen et al. (Fri,) studied this question.