P695: SINGLE-CELL MULTIOMICS ANALYSIS OF MYELODYSPLASTIC SYNDROME PREDICTS CLINICAL RESPONSE TO TREATMENT WITH DNA DEMETHYLATION AGENT

Background: Epigenetic defects, such as DNA methylation alterations, are a common feature of human malignancies. This finding has supported the use of hypomethylating agents (HMAs) as potential agents to treat cancer patients, being accepted the clinical approval of these drugs in the treatment of myeloid malignancies as one of the great accomplishments of this research area. The best case is myelodysplastic syndrome (MDS), a pre-leukemia disease that can progress to acute myeloid leukemia (AML), where limited treatment repertoire was available until the emergence of HMAs. However, it has not been elucidated how we can predict if an MDS patient will be a responder or non-responder to the epigenetic compound and the duration of the effect. Using the awesome analytical resolution of the recently developed single-cell methodologies, we have herein addressed this point. Aims: To employ novel single-cell technologies to characterize mutational and immunophenotypic profiles that predict the response to HMAs and the duration of their effect in MDS patients. Methods: To evaluate the evolution of the clonal cellular and molecular composition of MDS upon the therapy with HMAs, we carried out single-cell DNA sequencing (scDNA-seq) of an amplicon panel for 53 genes frequently mutated in myeloid malignancies and single-cell protein sequencing (scProt-seq) of 45 cell-surface proteins to yield a simultaneous profile of the genetic context and immunophenotype. We sequenced hundreds of thousands of cells from these MDS samples where paired bone marrow samples were analyzed at the time of diagnosis and at the end of at least six cycles of azacitidine use. Results: The aforementioned single-cell multiomics approach led us to identify not only genetic mutations present in bone marrow samples and their clonal evolution throughout the course of azacitidine treatment, but also the cell types that harbor such mutations, which were defined by a unique combination of surface markers. More precisely, we proved the co-occurrence of particular missense, nonsense and frameshift mutations in the same myeloid progenitor cell and the linear or multibranched development of the mutant clones depending on the patient. Furthermore, we observed the persistence of cell lineages with specific mutations (such as non-classical monocytes and CD8+ effector T-cells) after the treatment with this DNA demethylation agent. Summary/Conclusion: Overall, our work shows that the co-occurrence of particular truncating/stop codon mutations in the same myeloid progenitor cell; the linear or multibranched evolution of the mutant clones; and the persistence of unique cell lineages with specific mutational profiles can predict if HMA treatment will provoke a clinical response in MDS patients. Keywords: Immunophenotype, Hypomethylating agents, Myelodysplastic syndrome, Somatic mutation