Impaired Hematopoietic Differentiation of iPSCs Derived From Patients with FPD/AML

Abstract 767

Somatic mutation of RUNX1 has been implicated in a variety of hematopoietic malignancies including myelodysplastic syndrome and acute myeloid leukemia, and previous studies using mouse models disclosed its critical roles in hematopoiesis. During embryonic development, Runx1 is absolutely essential in the emergence of hematopoietic stem and progenitor cells through hemogenic endothelium. In contrast, conditional disruption of Runx1 in adult hematopoietic system revealed that it was critical in the differentiation of megakaryocytes and lymphocytes as well as in the function of hematopoietic stem cells (HSCs). However, these results were derived from gene-disruption studies in mouse models, and the role of RUNX1 in human hematopoiesis has never been tested in experimental settings.

Familial platelet disorder/ acute myeloid leukemia (FPD/AML) is a rare autosomal dominant disorder caused by germline mutation of RUNX1, marked by thrombocytopenia and propensity to acute leukemia. To investigate the physiological function of RUNX1 in human hematopoiesis and the pathophysiology of FPD/AML, we derived induced pluripotent stem cells (iPSCs) from three distinct FPD/AML pedigrees (FPD-iPSCs) and examined their defects in hematopoietic differentiation. These pedigrees have distinct heterozygous mutations in RUNX1 gene, two in the N-terminal RUNT domain affecting its DNA-binding activity and one in the C-terminal region affecting its transactivation capacity.

After obtaining informed consent from the affected patients, we established iPSCs from their peripheral T cells by infecting Sendai viruses expressing four reprogramming factors (OCT3/4, SOX2, KLF4 and c-MYC). FPD-iPSCs could be established in comparable frequency as the one from normal individuals (WT-iPSCs). Initial characterization of FPD-iPSCs revealed that the established clones retained typical characteristics of pluripotent stem cells such as the expression of Nanog, Oct3/4, SSEA-3, SSEA-4, Tra-1-60 or Tra-1-81, and the teratoma formation in immunodeficient mice.

Next we examined the hematopoietic differentiation capacity of FPD-iPSCs by co-culturing on AGMS-3 cells, a stromal cell line established from aorta-gonad-mesonephros (AGM) region. FPD-iPSCs and WT-iPSCs were dispersed and plated on inactivated AGM-S3 cells and were co-cultured in the presence of vascular endothelial growth factor. On day 10 through day 14 of co-culture, cells were collected and analyzed for the emergence of hematopoietic progenitors (HPCs) by flow cytometry. Interestingly, FPD-iPSCs generated CD34⁺ cells or CD45⁺ cells in significantly lower frequencies as compared to WT-iPSCs. To evaluate the differentiation capacity of HPCs generated from iPSCs, CD34⁺ cells were sorted by flow cytometry and subjected to colony forming assays. This revealed that CD34⁺ cells derived from FPD-iPSCs generated significantly fewer colonies as compared to those from WT-iPSCs in all colony types examined, showing that differentiation capacity of HPCs were impaired by RUNX1 mutation. Furthermore, CD34⁺ cells from FPD-iPSCs generated CD41a⁺CD42b⁺ megakaryocytes (MgK) in significantly lower frequencies as compared to WT in in vitro liquid culture with stem cell factor (SCF) and thrombopoietin (TPO). Of note, MgKs differentiated from FPD-iPSCs are smaller in size as evidenced by mean-FSC by flow cytometry. These results indicate that differentiation of MgKs is impaired both quantitatively and qualitatively. Importantly, all three FPD-iPSC lines share the same phenotype in the above-described assays, suggesting that N-terminal and C-terminal RUNX1 mutations impose similar defects in hematopoietic differentiation of FPD-iPSCs.

Taken together, this study, for the first time, demonstrated that mutation of RUNX1 leads to the defective differentiation of hematopoietic cells in human settings. The phenotype observed in this study, at least in part, recapitulates the ones previously reported in Runx1-homozygously deficient mice, suggesting that the mutations of RUNX1 seen in FPD/AML indeed act in dominant negative manner.