Abstract
Introduction: AML is a heterogeneous disease characterized by an abnormal clonal architecture and is classified into subtypes based on cytogenetics, gene mutations, and rearrangements.Characterization of single cell landscape and highlighting heterogeneity in each AML subtype can be used for novel risk stratification strategies to guide specialized therapy decisions and improve clinical outcomes for pediatric AML patients. Therefore, in this study genetic variations based on AML subtypes' were leveraged to explore transcriptional heterogeneity of AML-blast and immunological landscape at the single cell level.
Methods: We analyzed single-cell transcriptome of 15 AML bone marrow samples collected at the time of diseases diagnosis consisting of samples from six AML genetic subtypes: RUNX1-RUNX1T1 (t(8;21)), MLL rearrangement, CEBPA, NPM, 7q (7/add(7q)/del(7q), and inv(16)(p13q22single (Ulukay et al., 2021). We analyzed 30,000 high-quality cells for transcriptome alterations after quality control, processing, and normalization using the Seurat package (Butler, A. et al., 2018). Differential expression analysis was performed using the Wilcoxon rank test (FC>1.2, P<0.05) to identify blast and immunological cells gene signatures for each subtype. The survival genie platform (Dwivedi, B. et al., 2022) was implemented for exploring the survival association of AML subtype-specific gene signatures.
Results: The supervised analysis comparing subtype-specific blast with other blast and immune cells resulted in the identification of gene signatures uniquely enriched in each AML genetic subtype. MLL (KMT2A) rearrangement is one of the most common recurrent cytogenetic aberrations in pediatric AML. MLL rearranged samples (n=4) depicted overexpression of specific genes including SOCS2, KCNE5, AZU1, TRPM4, and C4orf48. Previously SOCS2 has been found to contribute to the growth and maintenance of AML leukemic stem cells (Nguyen, C.H. et al., 2021). TRPM4 codes for a calcium-activated ion channel and has been shown to be overexpressed in MLL-rearranged AML (Wang et al., 2020). RUNX mutated samples (n=4) showed overexpression of TRH, ELANE, POU4F1, NPW, and C1QTNF4 in the blast cells. Similar analysis on inv(16) samples (n=3) identified significant overexpression of ITM2A, KLF2, and GPX1 genes. KLF2 expression induces differentiation and reduced proliferation in AML cell lines (Humbert, M et al., 2011). The blast-specific gene signatures identified from scRNA-seq were independently validated on the TARGET AML bulk RNA sequencing (RNA-seq) dataset after categorizing samples according to their mutational profile. Comparative analysis revealed that most of the AML subtype blasts-specific gene signatures identified in scRNA-seq data were also uniquely overexpressed in the corresponding AML subtypes from the bulk RNA-seq data. Further survival association of the AML subtype-specific gene signatures was explored using the survival genie platform. The MLL-associated gene signature showed association with poor outcome survival (Fig. 1A) whereas the (Fig. 1B) and inv (16)- associated gene signatures showed significant association with favorable outcomes in the TARGET AML dataset. Gene set enrichment analysis on the T cells in the scRNA-seq data revealed that MLL rearranged samples were enriched with exhausted T cells while RUNX mutated samples depicted higher levels of naïve T cell expression.
Conclusions: In this study, the hypothesis that mutation-associated transcriptional profiles drive different outcomes in pediatric AML was investigated at the single cell level. We found genetic subtype-specific transcriptome differences in AML-blasts as well as differences in T cells might be contributing toward diverse clinical outcomes.
Disclosures
Bhasin:Anxomics LLC: Current Employment, Current equity holder in private company. Bhasin:Canomiks INC: Other: Equity Ownership.
