The role of myeloid transcription factor networks in the response of normal and malignant stem cells to LSD1 inhibition

Abstract

Acute Myeloid Leukemia (AML) is a blood cancer arising in the bone marrow and is the most common form of leukemia in adults, with a survival rate below 30%. AML is driven by combinations of mutations and comprises a heterogeneous cell population with diverse differentiation patterns and leukemic stem cells (LSCs), which contribute to disease recurrence. Current treatments are poorly tolerated by older adults, the population most affected by AML, highlighting the need for more targeted, better-tolerated therapies. Epigenetic modulation is a promising strategy and understanding how LSD1 inhibitors, which target histone demethylation, affect normal and malignant cells is crucial for clinical translation. This thesis is based on the hypothesis that LSD1 inhibition has distinct effects in normal versus malignant hematopoietic cells. In normal hematopoiesis, LSD1 inhibition (LSD1i) enhances proliferation and progenitor maintenance by blocking differentiation. In leukemia, LSD1i reduces LSC content and promotes differentiation through activation of myeloid transcription factor (TF) networks. These effects were investigated using syngeneic murine AML models, primary murine stem cells, and human material in vitro and in vivo. First, I examined LSD1i effects on normal hematopoiesis using short-term in vitro assays. Although LSD1i expanded phenotypic stem cells, they lacked engraftment potential. RNA sequencing revealed loss of key stem cell markers and increased activity of monocytic TF networks. Second, I studied LSD1i in murine AML models with differential sensitivity to treatment. Differences in monocytic TF programs explained treatment responses. Using an IRF8 knockout model, we showed it modulates responses to LSD1i, all-trans retinoic acid (ATRA), and their combination. Finally, I investigated KAT6A inhibition, which could induce differentiation in the more resistant model. Our central finding is that AML sensitivity to ATRA and LSD1i depends on monocytic TF networks, which may serve as predictive biomarkers for these differentiation-inducing therapies.