RUNX1-205, a novel splice variant of the human RUNX1 gene, has blockage effect on mesoderm–hemogenesis transition and promotion effect during the late stage of hematopoiesis

Abstract Runt-related transcription factor 1 (RUNX1) is required for definitive hematopoiesis; however, the functions of most human RUNX1 isoforms are unclear. In particular, the effects of RUNX1-205 (a novel splice variant that lacks exon 6 in comparison with RUNX1b) on human hematopoiesis are not clear. In this study, a human embryonic stem cell (hESC) line with inducible RUNX1-205 overexpression was established. Analyses of these cells revealed that induction of RUNX1-205 overexpression at early stage did not influence the induction of mesoderm but blocked the emergence of CD34+ cells, and the production of hematopoietic stem/progenitor cells was significantly reduced. In addition, the expression of hematopoiesis-related factors was downregulated. However, these effects were abolished when RUNX1-205 overexpression was induced after Day 6 in co-cultures of hESCs and AGM-S3 cells, indicating that the inhibitory effect occurred prior to generation of hemogenic endothelial cells, while the promotive effect could be observed during the late stage of hematopoiesis. This is very similar to that of RUNX1b. Interestingly, the mRNA expression profile of RUNX1-205 during hematopoiesis was distinct from that of RUNX1b, and the protein stability of RUNX1-205 was much higher than that of RUNX1b. Thus, the function of RUNX1-205 in normal and diseased models should be further explored.


Introduction
Runt-related transcription factor 1 (RUNX1) is the key gene for human hematopoiesis, which plays a critical role in the development of hemogenic endothelium and hematopoietic stem cell (HSC) formation during embryogenesis (Chen et al., 2009). Embryonic stem cells of Runx1 knockout mouse are unable to undergo hematopoietic differentiation, which can be rescued by Runx1 re-expression (Nishimura et al., 2004).
RUNX1 gene is 216 kb long and located on human chromosome 21 (21q22.12) and has two promoters termed P1 (distal) and P2 (proximal). Seventeen transcripts of human RUNX1 have been identified, among which RUNX1a/b/c have been well studied (Shinobu et al., 2007;Challen and Goodell, 2010;Shinobu et al., 2012;Draper et al., 2016). Their expressions are regulated by P1 and P2 promoters with alternative splicing, respectively (Levanon et al., 2001). RUNX1b/c have the same DNAbinding region and transcriptional regulatory domains while a few differences at the amino terminus, which indicated similar functions. RUNX1a is regulated by the P2 promoter, just like RUNX1b, but lacks the transcriptional regulatory domains and has an antagonism against the RUNX1b/c. The RUNX1a could stimulate the hematopoiesis while RUNX1b/c show a repressor for hematopoiesis at the earliest stage (Tsuzuki et al., 2007;Ran et al., 2013;Chen et al., 2017). These RUNX1 varieties exhibit distinct expression patterns during hematopoiesis (Challen and Goodell, 2010).
RUNX1-205, a novel splice variant of human RUNX1, is the sole and longest protein encoded by the complete coding region other than RUNX1a/b/c. RUNX1-205 lacks exon 6 in comparison with RUNX1b due to alternative splicing and plays a key role in ovarian cancer (Nanjundan et al., 2007;Hong and Fritz, 2019); however, its function in human hematopoiesis is unclear, which might be the final blank field of human RUNX1 variants research. The homologous mouse gene has complex functions in hematopoiesis (Komeno et al., 2014), indicating that RUNX1-205 plays an important role in human hematopoiesis. However, the function of RUNX1-205 in human hematopoiesis has not been explored using an in vitro system.
Although RUNX1-205 lacks exon 6, it retains an intact Runt-related DNA-binding domain and is highly homologous to RUNX1b. At the stage of mesoderm−hemogenesis transition, overexpression of RUNX1b at early stage blocks human hematopoiesis (Chen et al., 2017). The mRNA expression profile of RUNX1-205 and its functional similarities and differences with RUNX1b during hematopoiesis need to be elucidated. The role of RUNX1-205 in human physiology and pathologies must be explored further.

Genome structure analysis and alignment of RUNX1 homologous genes
The human RUNX1 gene has 12 exons, which encompass a runt homology domain (exons 3-5), an mSin3A interaction domain (exon 6), and a transactivation domain (exons 7B and 8). Exon 6 is deleted in human RUNX1-205 in comparison with RUNX1b (Supplementary Figure S1A). The function of RUNX1-205 is poorly understood. A splice variant in mouse called Runx1-202 has been reported, which is highly homologous to human RUNX1-205 (Komeno et al., 2014;Supplementary Figure S1B). A BLAST search of higher vertebrates from fish to human revealed that splice variants homologous to RUNX1-205 have been highly conserved during evolution (Supplementary Figure S1C).

Induction of RUNX1-205 overexpression at early stage blocks human hematopoiesis, whereas the induction at late stage promotes hematopoiesis
The effects of RUNX1-205 overexpression on human hematopoiesis differed according to when DOX treatment was initiated. Fluorescence activated cell sorting (FACS) analysis of co-cultured RUNX1-205/hESCs and the aorta-gonad-mesonephros (AGM-S3) cells revealed that treatment with DOX from Day 0 (D0) did not influence mesoderm induction (measured by product of KDR + CD34 − ; Figure 2A; Supplementary Figure S3A), but severely blocked early hematopoiesis (measured by product of KDR − CD34 + ). However, this effect was attenuated or abolished when DOX treatment was initiated after D6, and the hematopoiesis could even be stimulated in some degree with DOX treatment from D10 ( Figure 2B and C; Supplementary Figure S3B and C). Generation of CD34 + cells was prevented by induction of RUNX1-205 overexpression at early stage (especially from D0), and consequently, both CD34 + CD43 + and CD34 + CD45 + populations were lost. However, the production of CD34 + populations was almost normal when RUNX1-205 overexpression was induced from D6 or later and even significantly enhanced with RUNX1-205 induction from D10. This indicates that the development of hemogenic endothelium was blocked  Figure S3A). However, qRT-PCR analysis demonstrated that hematopoiesis-related genes, such as GATA1, GATA2, GATA3, c-KIT , vWF , and PU.1, were downregulated at D4 (Figure 3Aii). During embryoid body (EB) formation by RUNX1-205/hESCs, qRT-PCR and FACS analyses at EB-D10 demonstrated that the production of CD34 + cells was severely reduced by induction of RUNX1-205 overexpression at early stage. qRT-PCR analysis also revealed that hematopoiesis-related genes, such as GATA1, GATA2, and GATA3, were downregulated ( Figure 3B). In all these experiments, the earlier DOX was added, the inhibitory effect on hematopoiesis was more severe ( Figure 3A and B). This indicates that overexpression of RUNX1-205 at early stage could block hematopoiesis of hESCs in co-cultures with AGM-S3 cells and during EB formation, which is very similar to that of RUNX1b.

Induction of RUNX1-205 overexpression at early stage blocks colony formation, whereas the induction at late stage mildly stimulates hematopoietic potentials
The hematopoietic colony-forming assay was performed with co-cultures of RUNX1-205/hESCs and AGM-S3 cells at D14. In comparison with the noninduced control, induction of RUNX1-205 overexpression from D0 significantly blocked formation of burst-forming unit-erythroid (BFU-E), colony-forming unit-erythroid (CFU-E), colony-forming unit-mixed (CFU-Mix), and colony-forming unit-granulocyte/macrophage (CFU-GM) colonies (P < 0.05); however, the colony numbers were normal with RUNX1-205 induction from D6 ( Figure   Induction of RUNX1b and RUNX1-205 overexpression at early stage in co-cultures with AGM-S3 cells or during EB formation downregulates hematopoiesis-related genes. (A) Co-cultured RUNX1b/hESCs or RUNX1-205/hESCs were treated without or with DOX from D0 and analyzed by qRT-PCR at D4. Expression of KDR, which is related to mesoderm induction, was stable (i), while important hematopoiesisrelated genes were downregulated (ii). (B) FACS analysis at EB-D10 showed that overexpression of RUNX1-205 at early stage of EB formation blocked production of CD34 + cells (i), while qRT-PCR analysis showed that hematopoietic markers, such as CD34, GATA1, GATA2, and GATA3, were downregulated (ii). P < 0.05 was considered significant.
Induction of RUNX1-205 overexpression does not affect the hematopoiesis stage after mesoderm-hemogenesis transition but even later stage CD34 + KDR − populations sorted from RUNX1-205/hESCs co-cultured with AGMS-3 at D6 were further cultured for 5 days. FACS results showed that the induction of RUNX1-205 overexpression at the hematopoiesis stage (D6) did not affect the production of CD34 + CD43 + and CD34 + CD45 + populations, but promoted CD34 − CD43 + and CD34 − CD45 + populations ( Figure 6). This indicated that RUNX1-205 induction after the Figure 5 Sorting and hematopoietic potential detection at early stage of co-culture with AGM-S3 cells. KDR + cells were sorted from noninduced or induced RUNX1-205/hESC co-cultured with AGMS-3 cells at D2. About 5 × 10 3 sorted cells were re-plated in the 24-well plate with irradiated AGMS-3 cells, treated with or without DOX, and subjected to FACS analysis at D10 using antibody combinations against CD34/KDR and CD34/CD43. mesoderm−hemogenesis transition could not influence the development of hematopoiesis stem/progenitor cells (HSPCs) but promoted the production of CD34 − blood cells at later stage ( Figure 6).

RUNX1-205 is much more stable than RUNX1b
Western blot analysis demonstrated that RUNX1-205 was relatively stable after removal of DOX and remained high protein level for one day, whereas RUNX1b gradually disappeared ( Figure 7A). Further detection showed that RUNX1-205 was still detectable at D4, whereas RUNX1b could not be detected at D2. The protein half-life of RUNX1-205 is significantly longer than that of RUNX1b ( Figure 7B).

mRNA expression profiles of RUNX1-205 and RUNX1b differ in co-cultured H1 hESCs
Although RUNX1-205 had a similar effect on hematopoiesis as RUNX1b (Figures 2-4), qRT-PCR analysis of H1 hESCs co-cultured with AGM-S3 cells revealed that their mRNA expression profiles differed. mRNA expression of RUNX1-205 was high at early stage (D2-D4), low at late stage (after D4), and upregulated at D14. By contrast, mRNA expression of RUNX1b was high at early stage (D2-D4), lowest at D6, and gradually increased at late stage (after D8). mRNA expression of RUNX1c was similar to that of RUNX1b but exhibited greater fluctuation (Figure 8). Notably, mRNA expression of RUNX1-205 was much lower than that of RUNX1b even at early stage, but was significantly higher than that of RUNX1c (Figure 8).

Figure 7
Protein stability of human RUNX1b and RUNX1-205. Protein samples were prepared from 293T cells with inducible RUNX1b or RUNX1-205 expression at certain time points after DOX was removed. RUNX1b and RUNX1-205 were detected by western blotting within one day (at 0, 4, 8, 12, and 24 h; A) or longer time (0, 2, and 4 days; B). GAPDH served as a loading control.
The function of RUNX1-205 during hematopoiesis was elucidated using RUNX1-205/hESCs and compared with that of RUNX1b, which was previously described (Chen et al., 2017). Induction of RUNX1-205 overexpression at early stage (especially from D0) did not block induction of mesoderm, but blocked mesoderm-hemogenesis transition and the production of CD34 + CD43 − and its progenitor CD34 + KDR + cells (especially CD34 high CD43 − and CD34 high KDR + subpopulations) at early stage ( Figure 5), which led to the loss of CD34 + CD43 + and CD34 + CD45 + populations at late stage. These effects were attenuated or even abolished by RUNX1-205 induction at late stage (D6) according to examination on co-culture cells or further hematopoiesis culture, which even indicated the promotion of hematopoietic cells in some degree (Figures 2  and 6). Hematopoietic colony-forming assay confirmed that the hematopoietic potential was blocked by RUNX1-205 overexpression at early stage, but not affected or even mildly promoted by RUNX1-205 induction from D6 or later (Figure 4). qRT-PCR analysis demonstrated that various hematopoiesisrelated genes were downregulated at D4, which might be related to such blockage effects. A similar effect was observed during EB formation. In general, the earlier RUNX1-205 overexpression was induced, the inhibitory effect on hematopoiesis was more severe in both co-cultures with AGM-S3 cells and during EB formation. This indicates that induction of RUNX1-205 overexpression blocks mesoderm−hemogenesis transition, but not later stages, similar to the effect of RUNX1b. However, RUNX1-205 showed a promotion effect during the late stage of hematopoiesis, indicating that compared with RUNX1b, it may play different roles at different stages of mesoderm induction and hematopoiesis.
The protein stabilities of RUNX1-205 and RUNX1b in inducible 293T cells were investigated by western blotting. RUNX1-205 was much more stable than RUNX1b. This difference may be due to the absence of exon 6 in RUNX1-205. RUNX1b contains nine lysine residues, while RUNX1-205 lacks two of these residues (K182 and K188 in RUNX1b) due to deletion of exon 6. This may mean that RUNX1b is more readily targeted for degradation via the ubiquitin-proteasome pathway, which would explain its lower stability (Tsuji and Noda, 2000;Huang et al., 2001;Biggs et al., 2006). In addition, exon 6 of RUNX1 is important for methylation of arginine residues. The R206 and R210 residues in the RTAMR region of exon 6 can be methylated by PRMT1, and this prevents the interaction between RUNX1 and SIN3A (Zhao et al., 2008;L'Abbate et al., 2015). By contrast, RUNX1-205 lacks such interactions. The difference in posttranslational modifications due to the absence of exon 6 in RUNX1-205 may contribute to the difference in protein stability between RUNX1-205 and RUNX1b. A possible hypothesis is that the alternative splicing mechanism could change the ratio between transcripts of RUNX1b and RUNX1-205 so as to control the ratio of two proteins and their average protein life-time, which might be important for hematopoiesis in early stage.
qRT-PCR analysis revealed that the mRNA expression profiles of RUNX1-205 and RUNX1b differed during hematopoiesis. mRNA expression of RUNX1-205 was higher at early stage (D2-D4) than at late stage (after D4), suggesting that this isoform mainly functions during mesoderm induction. By contrast, mRNA expression of RUNX1b was high at both early (D2-D4) and late (after D8) stages, indicating that this isoform plays important roles throughout hematopoiesis. mRNA expression of RUNX1-205 was much lower than that of RUNX1b even at early stage but was significantly higher than that of RUNX1c. This indicates that RUNX1-205 ought to be an important isoform related to hematopoiesis. Expression of RUNX1-205 and RUNX1b in co-cultures with AGM-S3 cells was high at early stage during generation of mesoderm, indicating that both isoforms are important at this stage, and decreased at the mesoderm−hemogenesis transition. Therefore, the ectopic expression of them at this point might block the transition from mesoderm to hematopoiesis. RUNX1b was also important at late stage of hematopoiesis, and consequently its expression gradually increased after D8. However, expression of RUNX1-205 remained low after D4 and was not upregulated until D14, which indicates that this isoform has a distinct role in comparison with RUNX1b at late stage.
Though RUNX1-205 showed a similar function to RUNX1b during hESC differentiation to hematopoiesis, detailed functions of RUNX1-205 in normal and diseased models still need to be further explored.

Co-culture of hESCs and AGM-S3 cells
This study was approved by the institutional ethics committee of the Institute of Blood Transfusion, Chinese Academy of Medical Sciences and Peking Union Medical College (CAMS & PUMC). H1 hESCs were induced to undergo hematopoietic differentiation by co-culture with the mouse stromal cell-derived line AGM-S3 as reported previously (Thomson et al., 1998;Chen et al., 2017). Briefly, undifferentiated RUNX1b/hESCs and RUNX1-205/hESCs were dissected into small squares containing 0.5 × 10 3 -1 × 10 3 cells, plated onto irradiated AGM-S3 cells in human pluripotent stem cell-maintaining medium (Chen et al., 2017), cultured for 3 days at 37 • C in 5% CO 2 , and then switched to hematopoiesisinducing medium (defined as Day 0, D0; Chen et al., 2017). The culture medium was changed every day. Cells co-cultured for various numbers of days were dissociated using 0.05%-0.25% trypsin/EDTA (Invitrogen) for further analysis.

qRT-PCR analysis
Total RNA was extracted from 0.5 × 10 6 -1 × 10 6 cells using 1 ml TRIzol (Life Technologies) and purified according to the manufacturer's manual. Complementary DNA was synthesized using a reverse transcription kit (Bio-Rad). Each 20 µl reaction contained 4 µl 5× Mixture, 1 µl reverse transcriptase, 1 µg total RNA, and nuclease-free water to 20 µl. The conditions used for reverse transcription were as follows: 25 • C for 5 min, 42 • C for 30 min, 85 • C for 5 min, and storage at 4 • C. qPCR was performed using Fast Start Universal SYBR Green Master (Roche) on a CFX96 TM real-time system (Bio-Rad). Each 15 µl reaction contained 7.5 µl 2× Mixture, 0.4 µl each primer (10 µM), 4.7 µl H 2 O, and 2 µl complementary DNA. The conditions were as follows: denaturation at 10 min at 95 • C, followed by 45 cycles of 95 • C for 15 sec, 58 • C for 30 sec, and 72 • C for 30 sec. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as an internal control. The primers are listed in Supplementary Table S1.

MGG staining
Cells in BFU-E colonies were harvested and spun onto glass slides using a Cytospin 4 Cytocentrifuge (Thermo Fisher Scientific). For morphological observation, cells were stained with MGG solution (MERCK) and then imaged using an Olympus BX53 microscope equipped with an oil objective.

Cell sorting and endothelium culture
Noninduced or induced RUNX1-205/hESC co-cultures at D2 were dissociated with 0.05% trypsin solution and stained with anti-KDR antibody. KDR + cells were sorted using a BD FACSJazz TM Cell Sorter, and their purity was confirmed by FACS. About 5 × 10 3 sorted cells were re-plated in 24-well plate on irradiated AGM-S3 for 10 days with or without DOX induction, refreshed with hematopoiesis-inducing medium every other day, and finally analyzed by FACS analysis.

Statistical analysis
All experimental data were described as mean ± SD. Statistical significance was evaluated using the Student's t-test. P < 0.05 was considered significant. Data were analyzed using FlowJo V10 and GraphPad Prism 5 software.

Supplementary material
Supplementary material is available at Journal of Molecular Cell Biology online.