The Influence of FLI1 Expression Levels on Megakaryopoiesis: Studies Using iPSCs

Jacobsen syndrome is a rare, inherited hemizygous deletion of chromosome 11q that is often associated with a dysmegakaryopoiesis and macrothrombocytopenia termed Paris Trousseau syndrome (PTSx). Among the genes involved in the chromosomal deletion are FLI1 and ETS1, both of which belong to the ETS family of transcription factors and have been associated with megakaryopoiesis. One prior study using primary human hematopoietic stem cells suggested that the defect in PTSx was due to FLI1 allelic exclusion resulting in the generation of two distinct megakaryocyte (MK) populations, one immature and one mature. More consistent with the clinical course of this disorder, we hypothesize that PTSx is caused by FLI1 haploinsufficiency, where all MKs are affected and do not mature properly. The goal of our studies was to better understand MK development by investigating the role of FLI1 during megakaryopoiesis, including in PTSx. However, Fli1 deletions in mice did not replicate the defect observed in humans, so we used genome-engineered human induced pluripotent stem cell (iPSC) lines. We have established an iPSC line from a PTSx patient and derived from this a FLI1 overexpressing (OE) line in which FLI1 cDNA was cloned into the AAVS1 "safe harbor" locus with MK-specific expression driven by the GP1bα promoter. In parallel, we have a healthy control line, a control-derived FLI1 OE line, and a homozygous FLI1+/- line. In the control- and PTSx-FLI1 OE lines, FLI1 mRNA levels in MKs were 2X higher than control levels. The FLI1+/- line was generated using TALENs and expressed RNA at levels comparable to the PTSx line. To analyze MK progenitor potential, the iPSC lines were differentiated to hematopoietic progenitor cells (HPCs) and analyzed using Megacult colony assays. The PTSx line generated 4- to 6-fold less CFU-MK colonies per 1000 plated CD41+CD235+ cells compared to control (P=0.1) and PTSx-FLI1 OE (P=0.002). Likewise, the FLI1+/- line had less colonies compared to control (P=0.2) and control-FLI1 OE (P=0.005). The control-FLI1 OE line generated 70% more colonies (P=0.22) than the control line. To analyze MK generation, identical numbers of HPCs were expanded in liquid culture containing MK-specific cytokines and the numbers of CD41+CD42a+ cells were quantitated. The PTSx line had <20% the number of MKs generated in the control. The numbers of MKs generated from the PTSx-FLI1 OE and FLI1+/- HPCs were about half that of control line, while the number of MKs generated from the control FLI1 OE line was 40% higher than control. Platelet function studies show that CD42b+ PTSx and FLI1+/- platelet-like particles (PLPs) were unresponsive to convulxin stimulation compared to CD42b+ control, control-FLI1 OE, and PTSx-FLI1 OE PLPs. In addition, the PTSx and FLI1+/- MKs began to lose CD42b after only 3 days in culture while the PTSx-FLI1 OE and control MKs began to lose CD42b after 6 days in culture. The control-FLI1 OE MKs still retained CD42b expression after 8 days in culture. Overall, these data support our hypothesis that FLI1 haploinsufficiency underlies PTSx, as two distinct MK populations were not observed. Furthermore, the FLI1+/- MKs had similar characteristics to the PTSx-derived MKs, which pinpoints FLI1 deletion as the cause of PTSx MK deficiency. More importantly, we show that MK commitment of HPCs, MK expansion and maturation, their ability to retain CD42b expression and response to agonist stimulation correlate with FLI1 expression levels. Our findings have implications for production of functional MKs and platelets for future clinical application.