Acute myeloid leukemia (AML) is an aggressive blood cancer representing ~20% of childhood and ~80% of adult leukemias. Despite advances, treatment has remained the same for decades: cytotoxic chemotherapy with cytarabine combined with an anthracycline. Unfortunately, prognosis remains quite poor, and there is a clear and urgent need for better combination therapies to improve outcomes for AML patients.

While a molecularly heterogenous disease, all AML's are characterized by a block in differentiation resulting in the uncontrolled proliferation of immature myeloid elements. “Differentiation therapy,” which causes leukemic cells to overcome this block and differentiate into mature myeloid cells lacking leukemia-initiating activity, is an appealing therapeutic approach. It is successfully used in the subset of acute promyelocytic leukemia but has not yet shown success in other leukemia sub-types. Developing differentiation therapy for other leukemias would be an important advance - both agnostic to the genetic background of the leukemia, as well as more effective and better-tolerated than standard chemotherapy.

To identify potential compounds which promote myeloid differentiation, our lab previously performed a small molecule screen in an ER-HoxA9 conditionally immortalized murine GMP cell line. One of the identified compounds from the screen which promoted differentiation was an inhibitor of dihydroorotate dehydrogenase (DHODH), an enzyme which converts dihydroorotate to orotate as part of the de novo synthesis of the first pyrimidine building block uridine. DHODH is ubiquitously-expressed, and inhibition of DHODH (DHODHi) leads to rapid depletion of uridine, which impacts all metabolites which require uridine as a carrier (e.g. UDP-GlcNAc). The cell's ability to tolerate periods of pyrimidine starvation are dependent on a variety of alternative salvage pathways and are not understood. Several inhibitors of DHODH are clinically available.

In a follow up study, we combined the small molecule DHODH inhibitor Brequinar (BRQ) with a genome-wide CRISPR/Cas9 genetic perturbation screen to identify targets and pathways that sensitized cells to the differentiating effects of BRQ (Figure A). This screen was performed both in conditionally immortalized ER-HoxB8 murine GMPs, and in a murine stromal cell line. Across all doses and dosing schedules, unique to the GMP cell line, the most strikingly depleted gene in this screen has been FLT3, indicating that when FLT3 was lost, cells were more sensitized to differentiation by BRQ.

The mechanism by which this combination is effective is not currently understood. Preliminary data reveals that the combination of FLT3 inhibition (specifically AC220) and DHODH inhibition in vitro appears to be additive, and possibly synergistic. FLT3 is already an attractive therapeutic target given the frequency of activating mutations and overexpression in AML. In addition, FLT3 signals downstream through STAT5, which is known to function as an important “switch” for malignant transformation in hematopoietic cells in the setting of metabolic stress (Figure B). We hypothesized that the loss of FLT3 signaling and DHODH inhibition independently lead to decreased O-GlcNAcylation and inactivation of STAT3/5, depriving cells of a critical survival signal and sensitizing cells to the metabolic stress of nucleotide deprivation (Figure C). Global metabolic profiling of GMPs and AML cells (both FLT3-WT and FLT3-KO) treated with FLT3i, DHODHi, and the combination is currently underway to understand the impact of these changes on pyrimidine nucleotides, TCA cycle metabolites, and uridine-dependent metabolites …