Targeting Quiescent Leukemic Stem Cells and Overcoming Therapy Resistance in NUP98-NSD1 Fusion-Positive AML

NUP98-NSD1 fusion-positive acute myeloid leukemia (AML) is a subtype associated with poor prognosis, often diagnosed in pediatric patients with cytogenetically normal AML. This fusion involves the nuclear pore complex member NUP98 on chromosome 11 and the histone methyltransferase NSD1 on chromosome 5. Individuals with concurrent NUP98-NSD1 and transcription factor WT1 mutations exhibit chemotherapy resistance and poor outcomes, with less than 10% 4-year event-free survival. However, the molecular characteristics, developmental dependencies, and potential therapeutic targets for NUP98-NSD1-driven AML remain unclear.

To address these gaps, we established a humanized AML model by using CRISPR/Cas9 to induce endogenous NUP98-NSD1 fusion and WT1 knockout in primary hematopoietic stem cells (HSCs) from human fetal liver, postnatal umbilical cord blood, pediatric, and adult bone marrow. Notably, NUP98-NSD1 displayed a developmental dependency, where its expression transformed fetal liver HSCs, to a lesser extent cord blood, but, remarkably, not pediatric and adult bone marrow HSCs. In vitro, endogenous NUP98-NSD1 oncoprotein conferred clonal selection and proliferation advantages in HSCs. In vivo, xenotransplantation into immunodeficient mice recapitulated myeloid bias and leukemogenesis. Single-cell RNA and ATAC sequencing (scRNA-seq and scATAC-seq) revealed that NUP98-NSD1 maintains a unique self-renewal program characterized by aberrant expression of HOX genes, heightened inflammatory pathways, and enhanced epigenetic activity of EGR1, MYC, and AP-1 complex. Importantly, WT1 loss-of-function mutations further exacerbated these effects, leading to a shallower bipotential leukemic hierarchy enriched with quiescent leukemic stem cells (LSCs) that were resistant to both standard chemotherapy and small molecule inhibition against Menin.

We next sought to uncover vulnerabilities within the NUP98-NSD1 subtype that could be exploited therapeutically, irrespective of WT1 mutation status. We identified PRDM16 overexpression as a potential therapeutic target in NUP98-NSD1 AML. PRDM16, a regulator of stem cell self-renewal and oxidative metabolism, was shown to be crucial for leukemia progression. PRDM16 knockout impaired leukemia progression, particularly targeting LSCs, promoting differentiation, and enhancing sensitivity to chemotherapy and Menin inhibition in vivo. Additionally, PRDM16 loss reduced mitochondrial respiration, highlighting its role in the metabolic regulation of quiescent LSCs.

Lastly, we observed downregulation of estrogen signaling in primitive LSCs and heightened estrogen response in more differentiated cells in NUP98-NSD1 fusion-positive AML. Estrogen, acting primarily through estrogen receptor alpha (ESR1), has been shown to promote HSC proliferation and differentiation. The higher incidence of NUP98-NSD1-rearranged AML in males and its pediatric onset provided a rationale to investigate estrogen signaling as a therapeutic target, particularly for inducing differentiation and apoptosis in quiescent LSCs. Using tamoxifen, a selective estrogen receptor modulator, we aimed to activate ESR1 in LSCs. Tamoxifen treatment in NUP98-NSD1/WT1ko LSCs led to differentiation, increased apoptosis, and reduced mitochondrial respiration by targeting OXPHOS pathways. Thus, dual actions of tamoxifen disrupted the protective quiescent state and metabolic program of LSCs simultaneously.

This study elucidates the developmental and molecular complexities of NUP98-NSD1 fusion-positive AML, particularly the role of WT1 mutations in therapy resistance. It identifies PRDM16 as a key therapeutic target and highlights tamoxifen's potential to target quiescent LSCs. These findings offer promising new avenues for treating this challenging AML subtype.