Combinatorial Genetics Uncovers Novel Targets for the Treatment of Npm1/Cohesin Mutated AML

Acute myeloid leukemia (AML) is a cancer that results in an accumulation of abnormal myeloid cells in the bone marrow. While some targeted therapies are currently in development, the high level of genetic heterogeneity and low 5-year survival rate of AML underscore the need to uncover further therapeutic targets. 30 somatic mutations are recurrent in adult patients, with the average patient harboring 5-15 mutations (1). It is unknown how each genetic lesion interacts with others to change the course of the disease. Determining this necessitates the use of combinatorial genetics, which has the power to uncover novel targets for the treatment of each patient's unique disease.

The most common mutation found in adults with AML (25-30% of patients) is in the NPM1 gene, and is referred to as NPM1c (1,2). This mutation results in a mislocalization of the Npm1 protein exclusively to the cytoplasm. While NPM1c is considered to be a driver of AML, mice harboring this mutation only develop disease after a prolonged latency (18 months) and with incomplete penetrance, suggesting additional mutations are acquired during transformation (3,4). Approximately 50% of patients that have an NPM1 mutation also contain a heterozygous mutation in a cohesin complex gene (STAG2, SMC1A, SMC3, or RAD21), resulting in haploinsufficiency (1). Although cohesin mutations alone are insufficient for transformation to AML, enhanced self-renewal of HSPCs and significant changes in gene expression are observed in several different models (5,6,7). Given the clinical data, we hypothesized that cohesin mutations would cooperate with NPM1c to enhance transformation to AML through changes in gene expression. To address this hypothesis, we crossed the inducible NPM1^cflox/+and SMC3^flox/+mouse models.

Mice harboring both mutations developed AML with increased penetrance compared to NPM1^c/+ mice (65% vs. 55%). Enhanced self-renewal was also observed in double vs. single SMC3^-/+ or NPM1^c/+ HSPCs in vitro. RNA sequencing revealed that double mutant HSPCs exhibit deregulation of a unique set of genes compared to single mutant HSPCs. In our analysis, we observed the Rac1/2 guanine nucleotide exchange factor DOCK1 to be overexpressed in the double, but not the NPM1^c/+single, cells. High expression of DOCK1 has been correlated with decreased overall and disease-free survival in AML patients (8). Therefore, we hypothesized that Dock1 may be an attractive therapeutic target in COHESIN^Mut;NPM1^c AML. We found that shRNA-induced knockdown of DOCK1 or the use of a small molecule Dock inhibitor decreased cell growth and enhanced apoptosis in SMC3^-/+;NPM1^c/+ double mutant leukemias. Likewise, inhibition of Rac1/2 reduced cell growth, enhanced apoptosis, and decreased self-renewal in double mutant leukemic cell lines, suggesting that Dock1 may drive leukemogenesis through Rac1/2 activation. Collectively, our data suggest that Dock1/Rac represents a unique target for the treatment of patients harboring the Npm1/Cohesin mutational combination. Our results provide validity to the concept that combinatorial genetics can be used to target the unique genetic landscape of an individual patient.