Regulation of RUNX1 dosage is crucial for efficient blood formation from hemogenic endothelium

Abstract

During ontogeny, hematopoietic stem and progenitor cells arise from hemogenic endothelium through an endothelial-to-hematopoietic transition that is strictly dependent on the transcription factor RUNX1. Although it is well established that RUNX1 is essential for the onset of hematopoiesis, little is known about the role of RUNX1 dosage specifically in hemogenic endothelium and during the endothelial-to-hematopoietic transition. Here, we used the mouse embryonic stem cell differentiation system to determine if and how RUNX1 dosage affects hemogenic endothelium differentiation. The use of inducible Runx1 expression combined with alterations in the expression of the RUNX1 co-factor CBFβ allowed us to evaluate a wide range of RUNX1 levels. We demonstrate that low RUNX1 levels are sufficient and necessary to initiate an effective endothelial-to-hematopoietic transition. Subsequently, RUNX1 is also required to complete the endothelial-to-hematopoietic transition and to generate functional hematopoietic precursors. In contrast, elevated levels of RUNX1 are able to drive an accelerated endothelial-to-hematopoietic transition, but the resulting cells are unable to generate mature hematopoietic cells. Together, our results suggest that RUNX1 dosage plays a pivotal role in hemogenic endothelium maturation and the establishment of the hematopoietic system.

Keywords: CBFβ; Dosage; EHT; Hemogenic endothelium; RUNX1; SOX7.

Fig. 1.

Biphasic Runx1 expression during hematopoietic differentiation of mESCs. (A) Runx1 RNA expression determined by RNA-seq across five stages of differentiation. Fpkm, fragments/kilobase of transcript/million mapped reads. (B) Mouse Runx1 locus depicting Runx1b (P2-promoter driven) and Runx1c (P1 promoter driven) with mapped RNA-seq data. (C) qPCR across five stages of differentiation. Runt domain and exon 8 primers detect both Runx1b and Runx1c transcripts. Mean of n=5 or n=6 technical replicates is shown. Cross indicates below detection limit.

Fig. 2.

High Runx1 induction in iRunx1ko induces accelerated EHT. (A) Left: schematic of hemogenic endothelium differentiation in mESC-derived EHT cultures. HE1 (TIE2+/c-KIT+/CD41−) cells transition via the HE2 stage (TIE2+/c-KIT+/CD41+) into TIE2−/CD41+ hematopoietic progenitors and TIE2−/CD45+ committed myeloid cells. Right: flow cytometry of HE1 and HE2 populations in day 2 wild-type mESC-derived EHT cultures. (B) Schematic of hematopoietic differentiation experiments. FLK1⁺ cells, isolated from mESCs differentiated as EBs, were induced with doxycycline 1 day after re-plating in EHT-culture conditions. Day 2 and 3 EHT cultures were subjected to hematopoietic CFU assays with or without additional doxycycline. (C) Flow cytometry data of a 4-day time course depicting the transition through HE1 and HE2 in iRunx1ko cells after induction by low (0.06 µg/ml) and high (0.3 µg/ml) levels of doxycycline. Top left: percentage HE cells. Top right: median fluorescence intensity of TIE2+ cells in HE. Bottom left: percentage of HE2 cells within HE. Bottom right: median fluorescence intensity of CD41+ cells within HE2. Individual biological experiments and mean±s.d. are shown. Two-way ANOVA. N=biological replicates.

Fig. 3.

High Runx1 induction in iRunx1ko hampers hematopoietic potential. (A) Flow cytometry time course depicting the percentage of CD41+ (top) and CD45+ (bottom) cells in iRunx1ko upon doxycycline induction. Individual biological experiments and mean±s.d. are shown. Two-way ANOVA. N=biological replicates. (B) Myeloid hematopoietic colony-formation assay with iRunx1ko re-plated after 2 (top) or 3 (bottom) days of EHT culture. Individual biological experiments and mean±s.d. are shown. Two-way ANOVA. N=biological replicates. (C) iRunx1ko day 3 EHT-cultures cultured for 5 days in liquid hematopoietic mix. Right-hand panel shows a rare colony of budding cells in non-induced iRunx1ko cells. (D) Schematic of mast cell differentiation. FLK1+ cells were re-plated in mast cell media under ultra-low-attachment conditions. One day after re-plating, the cells were induced with doxycycline. Mast assays were analyzed after 22 days of culture. (E) Flow cytometry analysis of three independent iRunx1ko mast assays with or without doxycycline. In two out of three experiments, CD45+ cells could be generated in the absence of doxycycline induction. (F) RUNX1 western blot on day 3 non-induced iRunx1ko mast and EHT culture HA-tag immunoprecipitations. Input represents 5% of the total lysates. HA antibody raised in mouse, RUNX1 antibody raised in rabbit.

Fig. 4.

Cbfb^−/− have reduced RUNX1 protein levels and cannot initiate EHT. (A) Western blot for CBFβ on CRISPR/Cas9-generated Cbfb^−/− mESC clones. Lane 3 shows a wild-type (wt) mESC positive control. (B) Representative flow cytometry plots for Cbfb^−/− day 2 and day 7 EHT cultures. (C-E) Four Cbfb^−/− clones were differentiated across two independent experiments. Individual clones and mean are shown. (C,D) Flow cytometry data for four Cbfb^−/− clones analyzed on day 2-4 (C) and day 7 (D) of EHT culture. (E) Time-course qPCR analysis of Cbfb^−/− EHT cultures. Data are normalized to the wild-type line. (F) Time-course western blot analysis of EHT cultures for RUNX1, CBFβ and β-actin on two Cbfb^−/− clones (lanes 2, 5, 8, and 3, 6, 9) and wild type (lanes 1, 4, 7). β-actin-normalized relative densities for RUNX1 and CBFβ are shown. (G,H) Flow cytometry analysis of four Cbfb^−/− clones across five independent mast cell differentiations.

Fig. 5.

Low RUNX1 levels can induce EHT-like changes. (A,B) Representative flow cytometry plots (A) and flow cytometry data (B) of day 3 iRunx1ko and iRunx1ko^Cbfb−/− EHT cultures. (A) Left: HE (TIE2+/c-KIT+). Right: TIE2+/c-KIT+ gated histograms depicting the percentage CD41+ cells within HE. Horizontal bars indicate the CD41+ population. (B) Individual biological experiments and mean±s.d. are shown. Two-way ANOVA. N=biological replicates.

Fig. 6.

Persistence of HE core-like structures in doxycycline-induced iRunx1ko^Cbfb−/− cultures. (A) Representative pictures of day 5 iRunx1ko (top) and iRunx1ko^Cbfb−/− (bottom) EHT cultures. HE core-like structures (arrowheads) can be observed in doxycycline-induced iRunx1ko^Cbfb−/−. (B) Flow cytometry analysis of day 5 iRunx1ko and iRunx1ko^Cbfb−/− EHT cultures with or without doxycycline. Individual biological experiments and mean±s.d. (only for N>2) are shown. Two-way ANOVA (only for N>2). (C) Myeloid hematopoietic colony-formation assay with iRunx1ko and iRunx1ko^Cbfb−/− re-plated after 3 days of EHT culture. Individual biological experiments and mean±s.d. (only for N>2) are shown. Two-way ANOVA (only for N>2). N=biological replicates.