The Roles of RUNX1 in Human Hematopoiesis and Megakaryopoiesis Revealed By Genome-Targeted Human iPSCs and an Improved Hematopoietic Differentiation Model

The RUNX1 gene (also called AML1), one of the most mutated genes in acute myeloid leukemia (AML), was first identified decades ago that encodes a key regulatory transcriptional factor. Numerous studies using mouse and zebrafish models show that RUNX1 is essential for definitive hematopoiesis.  In mice, its homozygous knock-out (KO) in hematopoietic stem/progenitor cells also causes defects in lymphoid and megakaryocytic (MK) development.  However, heterozygous Runx1 gene mutations in laboratory mouse and zebrafish had little effects on development of hematopoietic stem/progenitor cells (HSPCs) or the MK cell lineage.  In contrast, heterozygous germline mutations in RUNX1 were found in patients with familial platelet disorder (FPD) with predisposition to AML and MDS.  The mechanisms underlying the observed differences between humans and small animal models remain unclear. 

In the past decade, we and others have utilized human embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) to investigate human hematopoiesis including the roles of the RUNX1 gene.  Using a feeder-free culture system, we generated human CD34+CD45+ HSPCs cells from human ESCs and iPSCs around 11-14 days after embryoid body (EB) formation.  The CD34+CD45+ HSPCs were capable to form multiple types of hematopoietic cells such as myeloid, erythroid, and polyploid MK cells (Connelly et al., 2014; Liu et al., 2015).  We also reported that human iPSCs derived from FPD patients containing a heterozygous RUNX1 mutation were defective in MK formation, and targeted correction of the mutated RUNX1 allele by genome editing restored the MK potential (Connelly et al., 2014).

Since then, we have extended our studies by precise genomic targeting in human wildtype iPSCs to ablate exon 5 that is common in all 3 isoforms of the RUNX1 gene, or exon 1B that is unique to the RUNX1c isoform.  Bi-allelic KO of RUNX1 at exon 5 completely abolished the formation of hematopoietic cells at days 11-14 after EB formation.  Complete disruption of exon 1B showed little effect, indicating that the RUNX1c isoform is dispensable for definitive hematopoiesis in the presence of the RUNX1a and RUNX1b isoforms transcribed from the downstream P2 promoter. Detailed analysis of EBs at days 6-8 revealed that bi-allelic RUNX1 KO at exon 5 (ablating all 3 isoforms) did not affect the formation of CD34+/CD31+/CD144+ endothelial–like cells.  However, the endothelial-to-hematopoietic transition (EHT) was completely blocked and no CD45+ hematopoietic cells emerged from the EHT culture supplemented with hematopoietic cytokines.

To elucidate the functions of different RUNX1 isoforms in early steps of human hematopoiesis, we adapted the EHT culture system that uses CD34+ cells isolated from earlier stages of EB formation (day 6-8), before definite hematopoiesis was observable (after day 8, Bai et al., 2015).  Two different iPSCs clones with homozygous or bi-allelic RUNX1 KO at exon 5 both failed to form CD45+ hematopoietic cells after EHT culture. Because constitutive transgene expression of RUNX1b or RUNX1c (but not RUNX1a) cDNA in human iPSCs inhibits hematopoietic differentiation, we transduced the day 6 EB cells at the beginning of the EHT culture with RUNX1-expressing vectors.  Lentiviral vectors constitutively express RUNX1b or RUNX1c (but not RUNX1a) cDNA partially restored the EHT and hematopoietic (CD45+) cell formation after 4-5 days in EHT culture. We further used a lentiviral vector in which RUNX1c (as an ER fusion protein) can be conditionally activated by 4-HT induction. Induction starting at day 4 and lasting for 3 days rendered the maximal effect of hematopoietic cell formation in the EHT culture using CD34+ cells isolated at day 6 of EB formation.

Our data corroborate with limited in vivo data using human fetal tissues on the possible roles of RUNX1 in definitive hematopoiesis.  At present we are analyzing iPSC clones with mono-allelic disruption of exon 5 in a wild type iPSC line, comparing to iPSCs derived from FPD patients. The current study of using isogenic human iPSCs will help to understand the roles of RUNX1 in human hematopoiesis and megakaryopoiesis, and offer an amenable system to study the RUNX1 gene functions and downstream target genes.