A novel human pluripotent stem cell-based gene activation system identifies IGFBP2 as a mediator in the production of hematopoietic progenitors in vitro

A major challenge in the stem cell biology field is the ability to produce fully functional cells from induced pluripotent stem cells (iPSCs) that are a valuable resource for cell therapy, drug screening and disease modelling. Here we developed a novel inducible CRISPR-mediated activation strategy (iCRISPRa) to drive the expression of multiple endogenous transcription factors important for in vitro cell fate and differentiation of iPSCs to haematopoietic progenitor cells and used this to identify a key paracrine role for IGFBP2 in the development of hematopoietic progenitors. We first identified nine candidate transcription factors that we predicted to be involved in blood cell emergence during development, then generated novel gRNAs directed to the transcriptional start site of these transcription factors that could also be detected during scRNAseq. iCRISPRa activation of these endogenous transcription factors resulted in a significant expansion of arterial-fated endothelial cells expressing high levels of IGFBP2 and showed that IGFBP2 remodels the metabolic activity during in vitro endothelial to hematopoietic transition. As well as providing fundamental new insights into the mechanisms of haematopoietic cell fate and differentiation, the broader applicability of iCRISPRa provides a valuable tool for studying dynamic processes controlling developmental events as well as for recapitulating abnormal phenotypes characterised by ectopic activation of specific endogenous gene expression in a wide range of systems.


Introduction
The haematopoietic system develops early during gestation through so called "waves" of hematopoietic progenitors that arise in different anatomical regions and result in the production of various progenitor and stem cells [1][2][3][4][5] .The precise signalling pathways leading to the production of hematopoietic stem and progenitor cells (HSPCs) during embryonic development are yet to be completely defined, posing a limitation on how to recapitulate the process in vitro from pluripotent stem cells.
During development, HSPCs are generated by a subset of endothelial cells, known as hemogenic endothelium [6][7][8][9] , via the endothelial to hematopoietic transition (EHT) 10 .During the EHT, endothelial cells undergo profound transcriptional remodelling whereby the expression of endothelial genes is gradually downregulated, and the transcription of the hematopoietic program is initiated 11 while the cells round up and eventually detach to enter the circulation 12,13 .
To explore the molecular control on the development of the hematopoietic system and to address the differences with the in vitro differentiation of induced pluripotent stem cells (iPSCs), we compared our single cell transcriptomics analysis of in vitro derived hemogenic endothelium and early progenitors 14 to that of the in vivo HSC-primed human hemogenic endothelium 15 .We then developed a novel DOX-inducible CRISPR gene activation system to assess the role of the genes that were expressed at a lower level within in vitro-derived cells compared to their in vivo counterparts.We employed single-cell RNA sequencing to track the presence of guide RNAs and we monitored the phenotypic effects of gene activation.With this experimental pipeline, we identified and functionally validated a novel role for IGFBP2, IGF Binding protein 2, in the development of in vitro progenitors and showed that IGBP2 remodels the metabolic activity during in vitro EHT.

Comparison of iPSC-derived endothelial cells with AGM AGM endothelial cells dataset identifies 9 differentially expressed transcription factors.
We, and others, have shown that in vitro differentiation of human iPSCs is a powerful tool to model intraembryonic hematopoiesis 14,[16][17][18] .To understand the molecular basis underlying the challenges associated with the in vitro production of blood stem and progenitor cells in vitro from differentiating iPSCs, we compared our scRNAseq dataset from differentiating iPSCs 14 to that of cells derived in vivo from the human aorta-gonad-mesonephros (AGM) region 15 .We integrated the transcriptomic data of in vitro derived endothelial (IVD_Endo)     and hematopoietic cells (IVD_HPC) with that of arterial endothelial cells (aEC), arterial hemogenic cells (aHEC) and venous endothelial cells (vEC) derived from human embryos collected between Carnegie stage 12 and 14 (Figure 1A).To identify possible target genes that could be manipulated in vitro to improve iPSCs differentiation, we first determined the transcription factors marking the aHEC in vivo (Table 1).Then, we filtered the transcription factor genes that were expressed at lower level in the IVD_Endo and IVD_HPC.This strategy identified 9 target transcription factors RUNX1T1, NR4A1, GATA2, SMAD7, ZNF124, SOX6, ZNF33A, NFAT5 and TFDP2 (Figure 1B), whose expression was detected across other AGM datasets in the hemogenic endothelium 17 (Supplementary Figure 1).

Development of a DOX-inducible dCAS9-SAM activation system in human iPSCs
We previously developed an all-in-one Synergistic Activator Mediator system, UniSAM, that mediates the transcriptional activation of endogenous gene expression 19 .To activate the nine target genes identified in this study, we developed a novel doxycycline-inducible SAM (iSAM) cassette targeted into the AAVS1 locus of human iPSCs (Figure 2A).We demonstrated the activation of RUNX1C expression in HeLa cells and that this was correlated with the doxycycline (DOX) concentration in a linear manner (Figure S2A-B).To verify the activation in human PSCs at single-cell resolution, we employed a RUNX1C-GFP human embryonic stem cell (hESC) reporter cell line (Figure S2C-G).As predicted, the level of expression of the mCherry tag, the fluorescent reporter within the iSAM cassette, was proportional to the concentration of DOX (Figure S2D-E) and to the number of RUNX1C-GFP+ cells (Figure S2 D-F).Furthermore, the RUNX1C expression level, measured by the mean fluorescence intensity (MFI) of the RUNX1C-GFP reporter, also correlated with the concentration of DOX (Figure S2G).We then tested the iSAM cassette in the iPSC line (SFCi55) (Figure 2B-D).Only when iPSCs were transfected with iSAM and the gRNA for RUNX1C, in the presence of DOX, was the expression of the RUNX1C gene and RUNX1 protein detected (Figure 2B-C).
We then targeted the iSAM cassette into the AAVS1 locus using a Zinc Finger Nuclease (ZFN) strategy 14,20,21 .iPSC clones that had specifically integrated the iSAM cassette into the AAVS1 locus were validated by genomic PCR and sequencing (Figure S2A-B).The AAVS1 locus has been reported to be a "safe harbour" site that is resistant to epigenetic silencing and indeed, we had previously demonstrated that transgenes inserted into the AAVS1 locus under the control of the constitutively active CAG promoter were efficiently expressed both in undifferentiated and in differentiated iPSCs 20,[22][23][24] .However, after the iSAM line had been established and cultured under self-renewal conditions, we noted a dramatic reduction in the number of mCherry+ cells in undifferentiated iSAM iPSCs upon DOX induction, that we predicted to be due to transgene-silencing of the rTTA DOX-inducible cassette (Figure S3C).
To overcome this problem, we treated the iSAM iPSC line with an inhibitor of histone deacetylases (HDACs), sodium butyrate (SB), reported to have no adverse effect on iPSC maintenance 25,26 and we also confirmed that the treatment with SB had no effect on viability and cell proliferation in our culture conditions (Figure S3D-E).A short 48-hour treatment significantly increased the number of mCherry+ cells upon DOX induction, proportional to the SB concentration (Figure S3F).We, therefore, maintained the iSAM iPSCs in the presence of SB and this fully restored the inducibility of the transgene with virtually all cells expressing mCherry in the presence of DOX (Figure S3G).
To test the effect of activating the 9 target genes on the transcriptomes of differentiating iPSC cells, we engineered the gRNAs to allow their detection within the single cell RNA sequencing pipeline 27 .We inserted a capture sequence prior to the termination signal to avoid any alteration in the secondary structure of the loops thus preserving the binding of the synergistic activators of the SAM system to the stem loops of the gRNAs.Of the two capture sequences available 27 , we decided to use the one that was predicted to result in fewer secondary structure alterations and this new gRNA was named 2.1 (Figure 2E).We compared the activation level achieved with the new 2.1 gRNA to that of the original 2.0 backbone using various gRNAs targeting RUNX1C (Figure 2F).These results convincingly demonstrated that the addition of the capture sequence in the gRNA 2.1 does not alter the level of endogenous gene activation that could be achieved (Figure 2G).Altogether, these results show that the iSAM system is able to induce gene expression that translates into an increased protein expression and thus provides a platform to steer phenotypical changes in cell identity.
To activate our target genes, we designed 5-7 gRNAs in the 200bp upstream of the transcriptional start sites of each of the 9 target genes.We subcloned a total of 49 gRNAs (Table S1) into the gRNA 2.1 backbone and packaged them into lentiviral particles, (herein referred to as the AGM library) as well as a non-targeting (NT) gRNA that was used as control.The iSAM iPSC line was transduced with the targeting gRNAs or the control nontargeting gRNA to generate the iSAM_AGM and iSAM_NT iPSCs line, respectively.After puromycin selection, their integration in the genome was also confirmed by PCR and sequencing (Figure S3H-I).

Single Cell RNA sequencing in combination with CRISPR activation identifies arterial cell type expansion in association with higher hematopoietic progenitor potential
To assess the transcriptional changes in response to the activation of the target genes, we differentiated the iSAM iPSCs, induced with DOX and subjected them to single-cell RNA sequencing using the 10X pipeline.After 10 days of differentiation, in the presence or absence of DOX, we FAC-sorted live CD34+ cells from iSAM_AGM and the iSAM_NT iPSCs.Following data filtering, we selected cells in which the gRNAs expression was detected and showed that our approach activated all the target genes, except for ZFN124.A higher level of expression of RUNX1T1, NR4A1, GATA2, SMAD7, SOX6, ZNF33A, NFAT5, TFDP2 was observed following DOX treatment of cells of iSAM_AGM cells compared to the iSAM_NT cells (Figure 3A).To study the effect of gene activation on transcriptional and cellular phenotype, we performed clustering analysis and detected a total of 7 clusters (Figure 3B).A high level of GJA4 and DLL4 expression was used to annotate the arterial-like cell cluster, while high levels of hemogenic markers, such as RUNX1 and CD44, were used to annotate the hemogenic clusters, EHT_1 and EHT_2 (Figure 3B, C-D).To assess the effect of the activation on cell identity, we analysed the proportion of the cell clusters between the libraries and detected a significant expansion of the arterial cluster in the DOXinduced iSAM-AGM compared to the iSAM_NT cells (Figure 3E).To validate the effect of the activation on the expansion of arterial cell population, we analysed their prevalence by flow cytometry.Although DOX treatment resulted in an average 1.63 ± 0.13 fold increase of CD34 + DLL4 + cells in the control iSAM_NT sample (Figure S3A), the increase observed in the iSAM-AGM cells was significantly greater, with an expansion of 3.33 ± 0.73 fold increase of CD34 + DLL4 + immunophenotypic arterial cells identified by flow cytometry (Figure 3F).To assess the effect of this predicted arterial cell expansion on the emergence of functional progenitors, we isolated CD34 + cells using magnetic beads and cocultured 20,000 cells on OP9 supportive stromal cells for 7 days, in the presence of hematopoietic differentiation cytokines.After one week, the progenitor cells were assessed by colony-forming unit (CFU) assays and scored 14 days later.We detected an increased number of CFU-E (colonyforming unit erythroid) and CFU-GM (colony-forming unit granulocyte/macrophage) and a reduction of CFU-M (colony-forming unit macrophage) in iSAM-AGM samples cultured in the presence of DOX compared to the absence of DOX but no significant effect of DOX in SAM_NT samples (Figure 3G).These data indicate that activation of target transcription factors using our novel CRISPR strategy results in transcriptional remodelling and a steer in cell identity that we detected as a functional difference in the hematopoietic progenitor profile.

The addition of IGFBP2 to the in vitro differentiation leads to a higher number of functional hematopoietic progenitor cells.
To better understand the molecular mechanism behind the increased progenitor development, we compared the expression profile of the arterial cells between the different activation libraries, and we obtained a list of genes upregulated upon activation of the targets (Table 1).The most upregulated gene, IGFBP2, was expressed at significantly higher levels in the iSAM_AGM library in the presence of DOX compared to the others (Figure 4A, Table 1).We then compared the gRNA distribution in these arterial cells from the iSAM_AGM treated with DOX to that of arterial cells without activation.We observed a significant enrichment of the RUNX1T1-specific gRNAs (Table1), indicating that the increased IGFBP2 expression could be downstream of RUNX1T1 activation.IGF Binding Protein 2 is a member of the family of IGF binding proteins and is thought to be secreted from cells where it then binds IGF1, IGF2, and other extracellular matrix proteins, modulating their function.To test if the increased frequency of functional hematopoietic progenitors was due to IGFBP2 paracrine signalling, presumably derived from the arterial cells, we supplemented the media with IGFBP2 at 100ng/ml after the induction of endothelial cell differentiation.To explore the role of IGFBP2 we employed the parental iPSCs line, SFCi55 from which the iSAM line was derived.We isolated CD34+ cells at day 8 and co-cultured them on OP9 cells in presence of IGFBP2 for one week, then tested for their hematopoietic clonogenic potential using CFU assays (Figure 4B).Cells treated with IGFBP2 showed a significant increase in the total number of haematopoietic CFU colonies compared to cells in control cultures.To assess whether IGFBP2 also had a paracrine effect on the production of arterial cells themselves, we analyzed the proportion of DLL4+ cells, but no difference was detected in the presence of IGFBP2 (Figure 4C).This observation supports our hypothesis that the change in hematopoietic progenitor production following activation in the iSAM_AGM line is associated with differential gene expression within the arterial cells rather than expansion of that cell population.We then focused on the characterization of the cells derived from the CD34+ cells after coculture with the OP9 in the presence of IGFBP2.Our results showed a significant expansion of the CD34+ and CD43+ cell populations further supporting our hypothesis (Figure 4D).To address the potential role of IGFBP2 in vivo, we analysed single cell sequencing data from the human AGM at Carnegie stages 14 and 15.
IGFBP2 is highly expressed within the AGM niche by stromal and epithelial cells and, most importantly, highest in endothelial cells (Figure 4E).Together these data show that IGFBP2 addition results in increased hematopoietic blood production in vitro and that the endothelial compartment is the most likely source of IGFBP2 both in vitro and in vivo.

IGFBP2 remodels the metabolic profile of endothelial cells.
Following the observation that IGFBP2 supports haematopoietic progenitor differentiation in vitro from human iPSCs we performed a time course single-cell RNA experiment of SFCi55 iPSCs differentiated in its presence.We isolated CD34 + cells and plated them in EHT culture on laminin 28 , rather than on OP9 support, to assess exclusively the specific effects of IGFBP2-on differentiating iPSCs.FAC-sorted single/live cells from day 10 and 13 in the presence and absence of IGFBP2 were subjected to single-cell RNAseq.Our time course transcriptomic analyses showed that IGFBP2 induced a change in the transcriptome, specifically on day 13 (Figure 5A).A cluster of endothelial cells enriched for genes associated with the KEGG pathway of growth factor binding, was detected almost exclusively at day 13; these cells displayed a different transcriptional signature (as indicated by a shift in their position in the UMAP embedding) in the presence of IGFBP2.. Interestingly, this cluster showed expression of GJA4 across the cells, indicating their broad arterial identity, with a specific reduction of DLL4 expression level concomitantly to an upregulation of RUNX1 in the region of the cluster induced by IGFBP2 (Figure 5B).This increase in hemogenic identity within the endothelial cell compartment is consistent with the observation of increased functional progenitor production.We compared the transcriptome of cells at day 13 in the presence and absence of IGFBP2 and, using KEGG enrichment analysis, we observed that the genes that were upregulated by IGFBP2 were highly enriched in the oxidative phosphorylation term (Figure 5C-D).Since the metabolic switch between glycolytic to mitochondrial metabolism has been previously reported in definitive hematopoiesis 29,30 , we tested whether the addition of IGFBP2 could result in a different ATP production profile.We analysed the ATP production at day 13, when RUNX1 expression was induced by IGFBP2 addition, by using specific inhibitors of the complex I and II and complex V to quantify the intracellular ATP and mitochondrial synthesis, respectively.We detected a reduction in glycolytic-derived ATP (Figure 5E-F), which translates to a higher contribution of mitochondria metabolism for IGFBP2-treated cells (Figure 5G).We then looked at the expression levels of genes encoding the enzymes of the glycolytic pathway and its checkpoints, glucose and lactate transporters, monocarboxylate transporters and the enzymes associated with hexokinase and phosphofructose reactions.We observed a general downregulation of the glycolytic enzymes and its checkpoints, with the exception of PFKM and SLC16A1, (Figure S4), which were expressed at lower levels in the IGFBP2treated cells at day 13 compared to the control on the same day.These differences were not observed on day 10 cells, which is consistent with our previous observations that the IGFBP2-induced remodelling of the transcriptome happens exclusively on day 13.These results indicate that the presence of IGFBP2 induces the upregulation of RUNX1 and OxPhos genes and the metabolic remodelling of endothelial cells.

Discussion
The complexity and dynamism of developmental hematopoiesis in vivo have imposed challenges in accurately reproducing the process in vitro.We hypothesized that this could, in part, be due to the inability to recapitulate the appropriate transcriptional programme in cultured cells.Here, we developed a novel CRISPR-activation system to induce the expression of genes that are expressed at low levels in cells that are generated in vitro, to explore the downstream consequences of their activation and to assess the effects on emerging hematopoietic progenitor cells.
When we compared endothelial cells derived in vitro from hiPSCs to those in the AGM region of the human embryo, at the time point when early hematopoietic commitment takes place, we identified nine transcription factors that were expressed at a lower level in the cultured cells.Some of these had been associated with blood cell development, including GATA2 [31][32][33] , SMAD7 34 , NR4A1 34 , SOX6 35 and RUNX1T1 15,17 .Other genes such as, ZNF124, ZNF33A, NFAT5 and TFDP2 had not been previously associated with hematopoiesis and could provide early evidence of their possible role in hematopoiesis that will require further studies.We were particularly interested in RUNX1T1 (also known as ETO) which is associated with the leukemic fusion protein, AML1/ETO resulting from the t(8;21) chromosomal translocation 36 .Furthermore, RUNX1T1 expression has been recently detected in transcriptomic analyses of the human AGM region 15,17 , but its precise role during the ontogeny of the blood system has not been elucidated.The addition of a capture sequence to the gRNA backbone enabled their detection coincidentally with the single-cell transcriptome, and this allowed us to demonstrate that RUNX1T1 gRNAs were significantly enriched in cells within an expanded arterial cluster and their presence was associated with a high level of expression of IGFBP2.IGFBP2 KO mice show increased expression of cellcycle inhibitors and HSC apoptosis, implicating IGFBP2 as a modulator of HSCs cell cycle and survival 37 .More recently IGFBP2 was reported as being highly expressed in the human AGM region at CS14 when HSCs are emerging 17 , supporting a possible role during developmental hematopoiesis but the precise molecular process was unclear.In this study, we show that the addition of IGFBP2 recombinant protein in our in vitro model of EHT results in the emergence of an increased number of functional hematopoietic progenitors and we provide evidence to support the hypothesis that RUNX1T1 could regulated its expression.
Because RUNX1T1 lacks a DNA binding domain, its direct involvement in the regulation of IGFBP2 expression must require association with other cofactors that are yet to be identified.
Our data, together with in vivo data 17 , show that IGFBP2 is expressed predominantly by endothelial cells.The supportive role of the endothelial niche in the development of HSCs has been studied in vivo [38][39][40] and exploited in vitro to support HSCs emergence 39,41 .
We describe here a novel role for IGFBP2 in the remodelling of the metabolism of iPSCderived endothelial cells and that this is associated with the induction of RUNX1, a hallmark of hemogenic fate.The addition of IGFBP2 in the culture induces upregulation of the genes of the oxidative phosphorylation pathway in association with an increased relative mitochondrial contribution to cellular ATP production.This is due to reduced glycolysis in association with the downregulation of genes coding for glycolytic enzymes and their checkpoints.The switch between glycolytic to mitochondrial metabolism has been shown to be essential for the EHT in definitive hematopoiesis 29,30 .This switch is induced in vivo by mechanical cues downstream of the establishment of circulation, and it is required for functional HSCs development 29 , while in vitro, the switch is driven by pyruvate mitochondrial catabolism, leading to definitive EHT as opposed to primitive 30 .Further in vivo studies are required to characterize the role of IGFBP2 during AGM hematopoiesis to overcome the limitation of using an in vitro model such as iPSCs differentiation.
In conclusion, we detected transcriptional differences between in vivo and in vitro developmental hematopoiesis and developed a novel inducible gene activation system to identify novel molecular players during the endothelial-to-hematopoietic transition.Our multidisciplinary approach identified RUNX1T1 and IGFBP2 as novel signalling molecules that support human blood progenitor development in vitro, the latter inducing a metabolic switch from cytoplasmic glycolysis to mitochondrial respiration.This study demonstrates that the combination of CRISPR-mediated activation of target genes with single-cell transcriptomic analysis in differentiating hPSCs can be a powerful approach to alter cell fate, providing a tool for gene function studies during human development.The fine epigenetic manipulation of the transcription can be readily applied to any cell lineage simply by adding specific gRNAs and it will be instrumental in exploring other developmental processes that can be, at least partially, be mimicked in vitro with human iPSCs.
Using this approach, we have identified the supportive role of paracrine IGFBP2, predominantly produced by endothelial cells, which induces transcriptional and metabolic remodelling in association with the induction of RUNX1 expression.one-sample T-test; p=0.7759

Gene expression analysis
Total RNA was purified using the RNAeasy Mini Kit (Qiagen) and cDNA synthesized from 500 ng of total RNA using the High Capacity cDNA synthesis Kit (Applied Biosystem). 2 ng of cDNA were amplified per reaction and each reaction was performed in triplicate using the LightCycler 384 (Roche) with SYBR Green Master Mix II (Roche).A melting curve was performed and analyzed for each gene to ensure the specificity of the amplification.β-Actin was used as reference genes to normalize the data 19 .

Pluripotent Stem Cells differentiation to hematopoietic progenitors
hPSCs were differentiated in a xeno-free composition of SFD medium 14 , BSA was substituted with human serum albumin, HSA (Irvine-Scientific).Day 0 differentiation medium, containing 10 ng/ml BMP4 was added to the colonies prior cutting.Cut colonies were transferred to a Cell Repellent 6 wells Plates (Greniner) to form embryoid bodies and cultured for two days.At day 2 media was changed and supplemented with 3 µM CHIR (StemMacs).At day 3, EBs were transferred into fresh media supplemented with 5 ng/ml bFGF and 15 ng/ml VEGF.At day 6 media was changed for final haematopoietic induction in SFD medium supplemented with 5 ng/ml bFGF, 15 ng/ml VEGF, 30 ng/ml IL3, 10 ng/ml IL6, 5 ng/ml IL11, 50 ng/ml SCF, 2 U/ml EPO, 30 ng/ml TPO, 10 ng/ml FLT3L and 25 ng/ml IGF1.From day 6 onward, cytokines were replaced every two days.

CD34 isolation
CD34+ cells were isolated using CD34 Magnetic Microbeads from Miltenyi Biotec, according to their manufacturing protocol.Briefly, Embryoid bodies were dissociated using Accutase (Life Techonologies) at 37°C for 30'.Cells were centrifuged and resuspended in 150 µl of PBS + 0.5% BSA + 2mM EDTA with 50 µl Fcr blocker and 50 µl of magnetic anti-CD34 at 4°C for 30'.Cells were washed using the same buffer and transferred to pre-equilibrated columns, washed three times and eluted.After centrifugation, cells were resuspend in SFD media, counted and plated for OP9 coculture.

OP9 coculture and colony assay
OP9 cells were maintained in a-MEM supplemented with 20% serum (Gibco) and sodium bicarbonate (Gibco) and passaged with Trypsin every 3-4 days.The day before the coculture, 45.000 OP9 cells were plated for each 12 well plates' well in SFD media.The day of the co-culture the 20.000 iSAM cells or 25.000 SFCi55 or H9 were plated in each well and culture in SFD media supplemented with 5 ng/ml bFGF, 15 ng/ml VEGF, 30 ng/ml IL3, 10 ng/ml IL6, 5 ng/ml IL11, 50 ng/ml SCF, 2 U/ml EPO, 30 ng/ml TPO, 10 ng/ml FLT3L and 25 ng/ml IGF1 and 100ng/ml IGFBP2.Cytokines were replaced twice during one week of coculture.At the end of the coculture, cells were collected by Trypsin and half of the well equivalent was plated in 2 ml of methylcellulose medium (Human enriched H4435, Stemcell Technologies).Cells were incubated in the assay for 14 days and then scored.

Flow cytometry staining and cell sorting
Embryoid bodies were dissociated using Accutase (Life Techonologies) at 37°C for 30'.
Cells were centrifuged and resuspended in PBS + 0.5% BSA + 2mM EDTA, counted and stained at 10 5 cells for a single tube.Cells were stained with antibodies for 30' at room temperature gently shaking.Flow cytometry data were collected using DIVA software (BD).
For the sorting experiments, cells were stained at 10 7 cells/ml in presence of the specific antibodies.Sorting was performed using FACSAria Fusion (BD) and cells were collected in PBS + 1% BSA.Data were analysed using FlowJo version 10.4.2.

Flow cytometry antibodies
For flow cytometry 10

F1Figure 2 -
Figure 2 -The inducible iSAM cassette successfully mediates activation of endogenous gene expression upon DOX induction.A -Schematic of the iSAM cassette containing the TET-on system under the control of EF1a and dCAS9-P2A-MS2-p65-HSF1-T2A-mCherry under the rTTA responsive elements, separated by genetic silencer and flanked by AAVS1 specific homology arms.B -RUNX1C gene expression activation after transient transfection of the iSAM plasmid and gRNAs in presence or absence of DOX in human iPSC line (n=3 from independent transfections).C -RUNX1 protein expression upon iSAM activation after transient transfection of the iSAM plasmid and gRNAs with DOX in human iPSC line detected by immunostaining.D Expression of the iSAM cassette reported by mCherry tag during the differentiation protocol, the representative images (bright field -BF, and fluorescence) show embryoid bodies at day 3 of differentiation.E -Schematic of the gRNA 2.1 containing the capture sequence for detection during the scRNAseq pipeline.F -RUNX1C gene activation level obtained using either the gRNA 2.0 or 2.1 backbone (n=3 from independent transfections of the 4 different gRNAs).G -Statistical analysis of the gRNAs activation level showing no significant variation following addition of the capture sequence (n=3 for each of the 4 different gRNAs).

F2Figure 3 -
Figure 3 -Single Cell RNA sequencing in combination with CRISPR activation identifies arterial cell type and functional hematopoietic expansion in association with activation of the 9 target genes.A -Gene expression profile of target genes following target genes' activation, heatmap shows the expression level of the target genes in the iSAM_NT and iSAM_AGM treated with DOX following normalisation on the -DOX control.B -Dimension reduction and clustering analysis of the scRNAseq data following activation, filtered on cells where the gRNA expression was detected.C -Arterial (GJA4, DLL4), venous (NRP2, APLNR) and hemogenic marker (CD44, RUNX1) expression distribution in the clusters indicated by the colour.D -Expression distribution visualised on the UMAP plot showing the location of arterial cells marked by DLL4, and hemogenic endothelium marked by CD44 and RUNX1.E -Contribution of the different libraries to the clusters showing that arterial cell cluster is overrepresented in the iSAM_AGM treated with DOX, compared to the other libraries.F -Expansion of the arterial population assessed by the membrane marker expression of DLL4+ following targets' activation, quantified by flow cytometry at day 8 of differentiation (Data are normalised on the iSAM_NT + DOX sample, n=5 independent differentiations, * p

Figure 4 -
Figure 4 -IGFBP2 addition to the in vitro differentiation leads to a higher number of functional hematopoietic progenitor cells.A -Violin plot of IGFBP2 expression profile in the arterial cells obtained from the different conditions, in the presence or absence of gRNAs and DOX.B -Number of hematopoietic colonies obtained after coculture on OP9 in presence or absence of IGFBP2 (n=3-4 from independent differentiations, ** p=0.0080,Sidak's Two way ANOVA).C -Percentage of DLL4 + arterial cells differentiation within the CD34 + compartment analysed by flow cytometry in day 8 EBs (n=4 from independent differentiations, Two-way Anova, ns =p>0.99).D -Expansion of hematopoietic progenitors analysed using markers' expression on suspension progenitors derived after coculture of CD34+ cells onto OP9 support (data are expressed as fold over the CTR in the absence of IGFBP2 (n=4 from independent differentiations, * p<0.02,Sidak's Two way ANOVA.E -Single-cell transcriptomic analysis of developing AGM collected from human embryos at Carnegie Stages14 and 15 enriched for CD31 + and CD34 + showing the IGFBP2 expression profile in vivo in the AGM.

Figure 5 -
Figure 5 -IGFBP2 alters cell metabolism by inducing a reduction in glycolytic ATP production.A -Clustering analysis of the single cell transcriptomic time course analysis of differentiating cells at day 10 and day 13 in the absence (CTR) or presence of IGFBP2.Arrows indicate the difference in the clustering due to the addition of IGFBP2 compared to control.B -Expression profile of arterial markers, GJA4 and DLL4, and hemogenic marker RUNX1 (top -the dashed line shows the location of the shift in gene expression of cells treated with IGFBP2) and their expression profile in the endothelial cells cluster marked by Growth factor binding in absence (CTR) and in presence of IGFBP2.C -KEGG enrichment analysis of the genes upregulated at day 13 upon IGFBP2 treatment.The arrow shows the ranking of the Oxidative Phosphorylation pathway.D -Dot Plot showing the expression profile of the genes coding for the enzyme of the Oxidative Phosphorylation pathway.E -Oxygen Consumption Rate (OCR) and Extracellular Acidification Rate (ECAR) profile in cells at day 13 of differentiation reporting mitochondrial respiration and glycolysis, respectively.F -ATP production rate divided by that deriving from glycolysis and from mitochondrial respiration, in cells treated with IGFBP2 and controls at day 13.G -Ratio of the ATP production between glycolysis and mitochondrial respiration in cells treated with IGFBP2 and controls at day 13.