Introduction: Loss-of-function mutations in Runt-related transcription factor 1 (RUNX1) are commonly found in both germline and somatic hematopoietic malignancies and confer particularly poor prognosis in AML. However, it remains unclear how RUNX1 functions during hematopoietic and leukemic development, particularly because RUNX1 mutations alone are not sufficient to cause myeloid malignancy and some models show that RUNX1 mutations confer hematopoietic stem cell defects. Recently, mouse models have shown that RUNX1-deficient neutrophils upregulate NFκB activity, and hematopoietic stem and progenitor cells (HSPCs) with overactive inflammatory pathways gain competitive advantage under chronic inflammation. Thus, we hypothesized that while RUNX1 mutations impair normal HSPC function, inflammation may select for or rescue RUNX1 mutant HSPCs.

Methods: To interrogate the effect of RUNX1 loss in human CD34+ HSPCs, we disrupted the RUNX1 locus using CRISPR/Cas9 and AAV6-mediated homology directed repair. Importantly, by using an AAV6 vector that carries arms of homology flanking a fluorescent reporter expression cassette, we are able to track and isolate cells edited at the RUNX1 locus for in vitro and in vivo functional analyses and for molecular characterization using RNA-seq and ATAC-seq.

Results: First, we used this system to evaluate the functional consequences of RUNX1 knockout (KO) in human CD34+ HSPCs. Loss of RUNX1 caused early erythroid-megakaryocytic differentiation arrest and skewing toward monocytic differentiation. RUNX1 KO cells demonstrated decreased proliferation, cell cycle arrest, and reduction in serial replating potential in vitro. In competitive transplantation experiments in NSG mice, RUNX1 KO engraftment decreased over time in both primary and secondary transplant, revealing a competitive disadvantage. Second, ATAC-seq peak motif analysis showed that PU.1 and NFκB motifs are more accessible upon RUNX1 KO whereas GATA, TAL1, and RUNX motifs were less accessible. Similarly, gene set enrichment analysis of transcriptional data confirmed the broad upregulation of NFκB-mediated inflammatory programs; downregulation of GATA1-dependent heme metabolism and platelet development pathways; and downregulation of MYC- and E2F-dependent cell cycle programs. These observations imply that RUNX1 directs cell fate decisions by recruiting and activating lineage-specific hematopoietic transcription factors and augmenting stem cell proliferation programs.

We next sought to determine which cytokines are sufficient to drive RUNX1 KO cell expansion. RUNX1 KO cells not only expanded preferentially in NSG mice expressing human SCF, GM-CSF, and IL-3 (NSGS mice), but also were no longer defective in competitive transplants in these mice. Further, treatment with IL-3 was sufficient to significantly expand RUNX1 KO cells in vitro. Flow cytometry revealed that the IL-3 receptor CD123 is upregulated in RUNX1 KO cells compared to control. Similarly, RUNX1-mutant AML patient samples express higher levels of CD123 than RUNX1-wildtype AML patient samples. Finally, evaluation of publicly available RUNX1 ChIP-seq of bone marrow CD34+ HSPCs revealed that RUNX1 directly binds the promoter of CD123. Ongoing efforts are aimed at determining whether targeting CD123 and IL-3 signaling may be a viable therapeutic approach for the prevention or treatment of RUNX1-mutant myeloid malignancies.

Conclusion: In summary, we established a RUNX1-deficient human HSPC model not only to evaluate the role of RUNX1 in hematopoiesis, but also to characterize intrinsic and extrinsic factors involved in RUNX1-deficient clonal expansion and leukemic transformation. We show that RUNX1 KO causes monocytic skew at the expense of erythro-megakaryocytic potential and severely limits HSC engraftment and expansion in vivo. Molecular profiling reveals that these effects are associated with dysregulation of both transcription factor activity and cytokine signaling. However, exposure to IL-3 rescues RUNX1-deficient cell proliferative defects in vitro and competitive engraftment defects in vivo. This hypersensitivity to IL-3 signaling is mediated in part by increased expression of the IL-3 receptor CD123. These findings reveal how RUNX1 mutations may initially behave in a deleterious manner but can ultimately confer an advantage to HSPCs under certain environmental conditions.


Majeti:CD47 Inc.: Divested equity in a private or publicly-traded company in the past 24 months; Gilead Sciences: Divested equity in a private or publicly-traded company in the past 24 months, Patents & Royalties; Kodikaz Therapeutic Solutions Inc: Membership on an entity's Board of Directors or advisory committees.

Author notes


Asterisk with author names denotes non-ASH members.