Type: Oral
Session: 618. Acute Myeloid Leukemias: Biomarkers and Molecular Markers in Diagnosis and Prognosis: Functional Genomics in Prognosis and Novel Therapies
Hematology Disease Topics & Pathways:
Research, Lymphoid Leukemias, ALL, Acute Myeloid Malignancies, AML, Genomics, Diseases, Lymphoid Malignancies, Myeloid Malignancies, Biological Processes
NUP98 rearrangements (NUP98r) are found in a variety of pediatric hematologic malignancies, and pediatric acute myeloid leukemia (AML) patients with NUP98r are known to have a poor prognosis. NUP98 fusion partners are associated with specific leukemia subtypes, such as NUP98::NSD1 with myeloblastic AML and NUP98::KDM5A with acute megakaryocytic leukemia (AMKL). Various other NUP98r are also found in other malignancies such as T-cell acute lymphoblastic leukemia (T-ALL). Although previous studies revealed recurrent FLT3 internal tandem duplications (FLT3-ITD) and WT1 mutations in NUP98::NSD1 AML or RB1 loss in NUP98::KDM5A AMKL (PMID:36815378), the genome-wide mutational patterns of NUP98r leukemias and its association with transcriptional status remains to be elucidated.
Methods
We established a pediatric NUP98r leukemia cohort consisting of 185 samples (177 cases) from our institute, public databases, or collaborators, including AML (n=154), T-ALL (n=18), therapy-related myeloid neoplasms (t-MN, n=4), and myelodysplastic syndrome (MDS, n=1). We obtained RNA sequencing (RNA-seq, n=185), whole-genome/exome sequencing (WGS/WES, n=94), and single cell RNA-seq (scRNAseq, n=6). Fusion genes were called from RNA-seq, and somatic alterations were called from both WGS/WES and RNA-seq. Functional assays were performed using CRISPR/Cas9 in cord blood CD34+ cells transformed with NUP98::KDM5A (PMID:34903620).
Results
Among 177 cases, we identified 21 unique NUP98 fusions, among which NUP98::NSD1 (n=92) and NUP98::KDM5A (n=45) were the most frequent. We also observed NUP98::RAP1GDS1 (n=11), enriched in T-ALL and rarely reported in previous studies. The remaining 18 fusions were found in a range of 1-4 patients, many of which are epigenetic regulators or homeobox genes. Genomic profiling confirmed known associations between NUP98::NSD1 AML and FLT3-ITD or WT1, along with novel associations like JAK2 mutations in NUP98::KDM5A AMKL, and IL7R and NOTCH1mutations with NUP98r T-ALL. Transcriptional analyses with other subtypes of AML (n=816) and T-ALL (n=1,319) showed NUP98r leukemias distributed in six large clusters, where fusions are associated with specific clusters but not exclusively. For example, NUP98::KDM5A with RB1 loss was found exclusively in an AMKL cluster, but NUP98::KDM5A without RB1 loss was found both in myelocytic AML and T-ALL clusters. Similarly, NUP98::RAP1GDS1 cases were found in immature AML and T-ALL clusters, with co-occurring NOTCH1 mutations exclusively in T-ALL clusters, suggesting that fusion partners and cooperating alterations contribute to the disease phenotypes.
ScRNAseq showed AMKL cells with NUP98::KDM5A and RB1 loss at the megakaryocyte–erythroid progenitor (MEP) stage and T-ALL cells with NUP98::KDM5A and NOTCH1 mutations at the pre-T cell stage. A diagnosis-relapse pair sample of NUP98::NSD1 AML showed enrichment at the monocytic stage while transitioning to the granulocyte/monocyte progenitor (GMP) stage at relapse with additional WT1 mutations, indicating changes in cellular hierarchy with cooperating alterations.
Induction of RB1 deletion or WT1 frameshift mutations with CRISPR/Cas9 in cord blood CD34+ cells transformed with NUP98::KDM5A both showed growth advantage over non-targeting controls. Importantly, RNA-seq and flow cytometric analyses revealed significant upregulation of megakaryocytic lineage marker, CD41a, and downregulation of MPO or CEBPA with RB1 deletion, consistent with patient data of AMKL with RB1 loss. In contrast, WT1 frameshift mutations decreased GFI1B and ITGA2B (encoding CD41a) expression while maintaining MPO and CEBPA expression, suggesting loss of megakaryocytic phenotypes.
Conclusion
Our data suggests that NUP98 fusion oncoproteins and co-occurring somatic alterations cooperatively contribute to the disease phenotypes. Further studies of drug sensitivities in using this NUP98r cord blood CD34+ model will guide fusion and disease-dependent treatment strategies for NUP98r leukemia to overcome its refractory nature.
Disclosures: Iacobucci: Arima Genomics: Consultancy; Mission Bio: Other: Travel expenses .
See more of: Oral and Poster Abstracts