MultiOmic characteriZation of Acute Myeloid Leukemia Evolving From myelopRoliferative Neoplasm to Identify New Targeted Therapeutic Strategies (MOZART)

MultiOmic characteriZation of Acute Myeloid Leukemia Evolving From myelopRoliferative Neoplasm to Identify New Targeted Therapeutic Strategies

MultiOmic characteriZation of Acute Myeloid Leukemia Evolving From myelopRoliferative Neoplasm to Identify New Targeted Therapeutic Strategies

University Hospital, Brest, Institut Paoli-Calmettes, Saint-Louis Hospital, Paris, France, Gustave Roussy, Cancer Campus, Grand Paris

Myeloproliferative neoplasms (MPN) are chronic myeloid malignancies characterized by a risk of evolution to acute myeloid leukemia (AML). This unpredictable complication is associated with a grim outcome with median overall survival ranging between 2 to 10 months. To date, even allogeneic transplantation fails to significantly improve the prognosis. Biological and molecular mechanisms driving leukemic transformation are complex, ill-defined, and heterogeneous between patients. The investigator hypothesize that deciphering the molecular heterogeneity of post-MPN AML may lead identifying efficient drugs targeting of the most relevant leukemogenic pathways. Our main objective is to identify new targeted therapeutic approaches in post-MPN AML through in-depth characterization of the dysregulated pathways. The investigator will first characterize in an already annotated cohort of 120 post-MPN AML homogeneous patients subgroups using comprehensive multiomic analyses. Dysregulated pathways will be identified in each subgroup using the omics data and single-cell RNA-sequencing will be performed in a subset of patients in each subgroup. A customised drug-panel will be derived from the dysregulated pathway for an ex vivo drug screening, which will use a flow-cytometry read-out enabling to identity drug effect on cells survival, differentiation, and stemness. The 3 most promising drugs will be validated in a preclinical in vivo model of patient's derived xenograft (PDX) and their impact on clonal architecture will be studied in primary cell cultures using single-cell DNA-sequencing. Overall, this proposal may provide a better understanding of MPN leukemic transformation mechanisms and provide a path for personalized therapies. Our findings may therefore pave the way to drugs development in post-MPN AML that would provide a rationale for implementation of early clinical trials in these dreadful diseases.

Patients samples and clinical data: The investigator will study samples from 120 patients with a post-MPN acute myeloid leukemia. These samples and the corresponding clinical data are available through FIMBANK, a national network of biological resources for myeloproliferative neoplasms (grant INCa, BCB 2013, Pr Valérie Ugo) and through the prospective phase II clinical trial CPX351-TA-SMP testing CPX351 monotherapy in post-MPN AML (NCT04992949, inclusions started in 01-2022). WP1: Deciphering the heterogeneity of post-MPN AML (primary objective) To answer these objectives, the investigator will conduct a multi-omics approach including targeted-NGS with a 400-genes panel, RNA-seq and methylome in a total of 120 post-MPN AML samples. All the genomic libraries will be constructed at the genomic facility of Angers University Hospital and the sequencing will be performed on a NovaSeq6000 in the GenoBIRD Platform in Nantes. Bioinformatic analysis will be performed by teams #1 and #3 and will derive for each sample: SNV/Indel and CNV from DNA sequencing, expression of mRNA and lncRNA, genes fusion and splicing events from RNA-seq, and methylation beta-values from methylome. In order to identify homogeneous subgroups from the genomic data, the investigator will perform unsupervised clustering analyses of each layer of genomic data. Then, all layers will be combined for integration of clusters using the Cluster Of Clusters Analysis (COCA) method (Wilkerson and Hayes, 2010). WP2: Identify the mechanisms of transformation and putative targets for therapy For this purpose, the investigator will analyze omics data generated in WP1 to identify the main molecular mechanisms driving the leukemic transformation of MPN. The investigator will perform a 2-step procedure: first by analyzing each genomic dataset separately and then, by analyzing all datasets together in an integrated multiblock analysis using the MOGSA method (Integrative Single Sample Gene-set). A total of 60 samples originating from a subset of patients classified in WP1 will be tested for ex vivo drug screening. The investigator will design a custom-made drug panel including standards of care, several drugs in clinical development in AML and, more importantly, a selection of drugs specifically targeting potential leukemic vulnerabilities identified. WP3: Confirm the efficacy of selected best drugs and their impact on clonal architecture To further validate the translational relevance of post-MPN AML deregulated pathways, the three most promising drug candidates will then be evaluated in a set of five post-MPN PDX models including at least 2 TP53-mutated post-MPN AML. The investigator will also evaluate how the drugs identified in WP2 may impact clonal evolution of the disease which is a key step towards understanding and improving the treatment of post-MPN AML. The 3 best candidate drugs or combinations identified in WP2 will be studied in cells from 5 selected patients with a complex molecular profile to evaluate the response of various subclones.

Patients will be recruited in the French biobank of MPN (BCB INCa FIMBANK) and the FILO's national biobank of acute leukemias (FILOthèque). All clinical and biological data are anonymized.

Multiomics analysis include targeted-NGS with a 400-genes panel, RNA-seq and methylome. All the genomic libraries will be constructed at the genomic facility of Angers University Hospital and the sequencing will be performed on a NovaSeq6000 in the GenoBIRD Platform in Nantes. Bioinformatic analysis will be performed and will derive for each sample: SNV/Indel and CNV from DNA sequencing, expression of mRNA and lncRNA, genes fusion and splicing events from RNA-seq, and methylation beta-values from methylome.

The drug screening will be performed on the 'NEXT-AML' platform at St-Louis Hospital, Paris. This platform uses a multiparametric screening strategy based on flow cytometry measurements of cell viability, cell differentiation and stem cell compartment. Primary patient cells will be cultured in a specific niche-like medium with amino-acids, cytokines and stromal cells (Dal Bello et al. 2022). Twenty-five drugs at 6 concentrations covering a 1000-fold concentration range will be studied for each sample.

GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways identified among the different subgroups by integration of omics data, cell type and trajectories inferred using scRNA-Seq

Confirm the efficacy of the most promising newly identified candidate drugs or therapeutic combinations in vivo in PDX models

Confirm the efficacy of the most promising newly identified candidate drugs or therapeutic combinations in vivo in PDX models

Identify the effect of the best candidate drugs/therapeutic combinations at the single-cell level and assess how clonal architecture evolves after treatment

Absolute change of the proportion of each leukemic clone in single-cell DNA-sequencing of in vitro culture

Inclusion Criteria: Patients with a prior diagnosis of MPN: polycythemia vera, essential thrombocythemia or primary myelofibrosis according to the WHO criteria Acute myeloid leukemia evolution defined by ≥ 20% of blasts cells Available material from bone marrow sampling at the time of leukemic transformation (i.e. ≥ 20% of blasts cells): DNA (1µg), RNA (500ng) +/- frozen mononuclear cells in DMSO for a subset of 60 patients (2 vials of at least 8 millions cells). Informed consent (or requalification procedure) Exclusion Criteria: - Patient not affiliated to the French health insurance