Myelodysplastic neoplasms (MDS) are heterogenous blood disorders that arise from dysfunctional hematopoietic stem cells (HSCs) and progenitor cells (HSPCs). According to the cancer stem cell (CSC) model, MDS is organized as a cellular hierarchy that arises from the malignant transformation of HSPCs into rare CSCs. MDS-CSCs are thought to persist during treatment and regenerate disease during relapse. Prior studies have linked MDS-CSCs to MDS-HSPC (Woll et al. Cancer Cell 2014, Pang et al, PNAS 2013, Will et al, Blood 2012), however, a specific cell type has not been isolated and purified as the MDS-CSC. Prognostic gene expression signatures in MDS have also been linked to immature HSPCs (Shiozawa et al, Blood 2017), however, a cell specific signature has not been identified. There is a need to characterize a cell specific gene signature for MDS-CSCs in order to study these cells. To address this need, we performed iterative statistical analyses on MDS gene expression data in order to identify a candidate CSC signature. We hypothesized that by analyzing genes specifically up-regulated in MDS-HSPCs, we can derive a CSC specific gene signature that is not only associated with poor outcomes in MDS, but also marks a subset of cells in MDS with stem cell programs. Up-regulated genes in MDS-HSPCs compared to healthy controls were derived by re-analyzing 73 sorted samples (Woll et al, Cancer Cell 2014) using the limma (Ritchie et al, Nucleic Acids Res 2015). Using these genes, we subsequently analyzed their association with survival in a cohort of 244 MDS patients (Shiozawa et al, Blood 2017, Gerstung et al, Nat Commun 2015, Tyner et al, Nature 2018). We performed iterative Cox proportional hazard models on a training data (n=146), using single and multiple gene combinations. A 2 gene score (i.e., MDS2), comprising IDS and SETBP1, was identified as the most significantly associated feature with decreased survival in MDS compared to Age, Sex, Cytogenetic Risk, and an established MDS score (Shiozawa et al, Blood 2017). Single cell expression of MDS2 was evaluated in MDS scRNA-seq samples (Dussiau et al, BMC Biol2022), and rare cells were identified expressing high levels of MDS2, i.e., MDS2 cells. These cells were integrated with healthy HSPCs. MDS2 cells reside primarily between HSCs and MPPs on a diffusion map, following a differentiation trajectory towards GMPs and monocyte precursors. These observations are consistent with prior studies, as MDS-CSCs were shown to be enriched in HSCs and GMPs (Pang et al, PNAS 2013, Woll et al, Cancer Cell 2014). Analysis of upregulated genes in MDS2 cells revealed that antigen processing, assembly and presentation were the most enriched processes. This analysis supports our approach for identifying a cell specific gene signature, and future work will focus on further single cell analyses and evaluation of CSC content and function of MDS2 cells.

Citation Format: Asiri Ediriwickrema, Andrew Gentles, Ravindra Majeti. IDS and SETBP1 is highly prognostic in myelodysplastic neoplasms and is a candidate stem cell signature [abstract]. In: Proceedings of the AACR Special Conference: Acute Myeloid Leukemia and Myelodysplastic Syndrome; 2023 Jan 23-25; Austin, TX. Philadelphia (PA): AACR; Blood Cancer Discov 2023;4(3_Suppl):Abstract nr A52.