Hemogenic gastruloids (haemGx) produced from mES cells promote hemato-endothelial specification with spatio-temporally accurate ontogeny.

(A) Timeline of mES cells assembly and culture into gastruloids over a 216h period with the addition of specified factors for the promotion of hemato-endothelial specification. The 24-hour pre-treatment in 2i+LIF was omitted when mES cultures showed no signs of differentiation. (B) Imaging of haemGx over time from 72h to 216h at 10x magnification, showing the assembly and growth of the 3D structures and the polarization of the Flk-1-GFP marker from 96h; scale bar: 100mm. (C) Flow cytometry timecourse analysis of haemGx for the presence of CD144 (VE-cadherin), C-Kit, CD41, and CD45 markers; representative plots. (D) Quantification of analysis in (C). Violin plots with median and interquartile range of up to 26 replicates; Kruskal-Wallis CD144 (p=0.0755), C-Kit (p<0.0001), CD41 (p<0.0342), CD45 (p<0.0001). Shown are significant Dunn’s multiple comparisons tests; adjusted q-value. (E) Immunostaining of whole individual haemGx at 216h showing the localized expression of Flk1-GFP (green), C-Kit (yellow), and CD45 (red) markers; nuclear staining is in DAPI (blue); scale bar: 100μm.

HaemGx produce hematopoietic output following the emergence of specialised endothelium.

(A) Immunostaining of whole individual haemGx at 144h showing the establishment of a vascular network by co-expression of CD31 and Flk1-GFP; nuclear staining is in DAPI (blue); scale bar: 100μm. (B) Flow cytometry analysis of haemGx cells at 144h and 216h stained for CD41, CD45, CD31, and CD34. (C) Quantification of flow cytometry analysis in (B) of CD41+ and CD45+ fractions co-expressing CD34 and CD31. Mixed effect analysis with a Ŝidàk test for multiple comparisons. (D) Colony-forming cell (CFC) assay of 216h-haemGx in multipotential methylcellulose-based medium. GEM: granulocyte-erythroid-monocyte; GM: granulocyte-monocyte; M: monocyte; E: erythroid. Please note that the medium does not include TPO and does not assess the presence of megakaryocytic progenitors. CFC frequency in 3 haemGx, n=3, mean±SD. (E) Representative photographs of colonies quantified in C; scale bar=2mm. (F) Real-time quantitative (qPCR) analysis of expression of hemato-endothelial genes by timepoint. Relative gene expression fold change calculated by normalization to Ppia. Bars represent mean of 3 replicates and show individual data points; p values by ANOVA across datapoints; post-hoc statistical significance between specific variables by Tukey’s test and shown by brackets.

Time-resolved analysis of haemGx by scRNA-seq captures successive waves of hematopoietic specification.

(A) UMAP clustering of subsets of sequenced cells expressing CD41, C-Kit, and CD45 at 144h, 192h and 216h, annotated by timepoint (left panel), sorting condition (middle panel), and Leiden clustering (right panel). (B) UMAP panels highlighting the expression of specific hemato-endothelial markers in the transcriptional spaces defined in (A). Color scale indicates expression levels in counts. (C) Heatmaps of differentiation of an arterial endothelial programme in haemGx from scRNA-seq data of sorted on expression of C-Kit at 144h and 192h. Color scale indicates expression in counts. (D) Violin plots quantifying the expression of hemato-endothelial markers in C-Kit+ fraction from clusters in (A) comparing 144h and 192h timepoints. Wilcoxon test; significant p value<0.05. (E) Expression of definitive hematopoietic genes in in C-Kit+, CD41+ and CD45+ cells at 144h and 192/216h timepoints; expression levels as normalised counts per million reads. Mann-Whitney or Kruskal-Wallis with Dunn’s multiple comparisons; significant p or q-value<0.05.

Late-stage haemGx contain hematopoietic output with transcriptional alignment to mouse embryonic populations and in vivo engraftment potential.

(A) Projection of clustered haemGx cells (CD41, C-Kit, and CD45 at 144h, 192h and 216h; Fig. 3A) onto mouse single-cell RNA-seq datasets capturing: arterial and haemogenic specification in the para-splanchnopleura (pSP) and AGM region between E8.0 and E11 (Hou et al. 2020); YS, AGM and FL progenitors and the AGM EHT (Zhu et al., 2020); HSC emergence from the dorsal aorta HE (Thambyrajah et al., 2024). Gastruloid cell projection identified as per their cluster of origin (Fig. 3A). (B) Schematic representation of the experimental workflow for execution and analysis of haemGx implantation in the adrenal gland of immunodeficient mice. (C) PCR detection of haemGx genomic (g)DNA in the bone marrow (BM) and spleen (Spl) of immunodeficient mice 4 weeks after unilateral adrenal implantation of 3 gastruloids/gland. Analysis of 5 replicates of 100ng gDNA/recipient tissue; control animal was injected unilaterally with PBS in an adrenal gland in parallel with experimental implantation. Reaction positive control used 100ng of gDNA from Rosa26-BFP::Flkj1-GFP mES cells (mES) used to generate haemGxs. NTC: no template control. (D) Flow cytometry plots of engraftment detection by BFP expression in bone marrow (BM) of recipient mice 4 weeks following implantation of 216h haemGx. (E) Flow cytometry plots of lineage affiliation of BFP+ engraftment in bone marrow (BM) of recipient mice 8 weeks following implantation of 216h haemGx.

HaemGx with MNX1 overexpression have increased cellularity and enhanced hemogenic endothelial potential.

(A) Cell type enrichment analysis in MNX1-r AML transcriptomes, compared to normal paediatric bone marrow (BM) or other paediatric AML from the TARGET database. GSEA used representative cell type gene sets from the 2021 DB database; bubble plot shows NES scores (color gradient) and statistical significance (– log10(FDR), bubble size). (B) Schematic representation of generation of MNX1-overexpressing and empty vector (EV) mES cells by lentiviral transduction, with subsequent assembly into haemGxs. (C) Imaging of haemGxs with MNX1 overexpression and EV controls at 10x magnification, showing appropriate assembly and polarization of the Flk1-GFP marker. scale bar: 300mm. (D) Size of haemGx at each timepoint, determined by the distance between the furthest extreme points in µm. Mean ± SD of 3 replicate experiments; 2-tailed t-test, p< 0.05 (*), 0.001 (**), 0.0001 (***), and 0.00001 (****). (E-G) Flow cytometry timecourse analysis of (E) C-Kit+, (F) CD41+, and (G) CD45+ cell abundance in MNX1 and EV haemGxs (120-216h). Mean ± standard deviation of 3-7 independent experiments; 2-way ANOVA and Sidak’s multiple comparison test significant at p<0.05 for C-Kit+ cells only (construct contribution to variance p=0.0191; 144h comparison *p=0.0190). (H) Volcano plots of differentially expressed genes (DEGs) by bulk RNA-seq between MNX1 and EV haemGxs at 144h and 216h. Intersection of statistically significant DEGs between 144 h and 216 h (shown in Venn diagram on top right) identified unique DEGs for each condition (labelled in red for 144 h and blue for 216 h). (J) Gene Ontology (GO) term enrichment for differentially upregulated genes between MNX1 and EV haemGx at 144h (red) and 216h (blue), computed in EnrichR using the GO Biological Process repository. (I) Transcription factor (TF) binding site enrichment on differentially upregulated genes between MNX1 and EV haemGx at 144h (red) and 216 hr (blue) using the ENCODE and ChEA Consensus TFs from ChIP database on EnrichR.

MNX1 selects C-Kit+ clonogenic cells and selectively transforms end-stage haemGx.

(A) Representative photographs of serial replating of colony-forming cell assays initiated from EV or MNX1 cells obtained at 144h of the haemGx protocol. (B) Quantification of colony-replating efficiency of EV and MNX1 144h-haemGx (GX) cells. Mean+SD of n>3 replicates; Kruskal-Wallis with Dunn’s multiple comparison testing at significant q<0.05. (C) Representative photographs of serial replating of colony-forming cell assays initiated from EV or MNX1 cells obtained at 216h of the haemGx protocol. (D) Quantification of colony-replating efficiency of EV and MNX1 216h-haemGx cells. Mean+SD of n=5-8 replicates; Kruskal-Wallis with Dunn’s multiple comparison testing at significant q<0.05. (E) Flow cytometry analysis of hemogenic progenitor C-Kit+ cells and myeloid-affiliated CD11b+ cells at first round of plating of 144h and 216h-haemGx initiated from EV and MNX1 cells; representative plots.

HaemGx with MNX1 overexpression recapitulate MNX1-Acute Myeloid Leukemia patient signatures.

(A) Quantification of human MNX1 transcripts in FPKM units from RNA-seq of GX-EV, GX-MNX1, and CFC at 144 hr and 216 hr. Bars represent mean of replicates. (B) Heatmap comparing the expression of all differentially expressed genes (DEGs) between all conditions (GX-MNX1 144 hr vs GX-EV 144 hr; GX-MNX1 216 hr vs GX-EV 216 hr; CFC MNX1 144 hr vs GX MNX1 144 hr; CFC MNX1 216 hr vs GX MNX1 216 hr) as Z score. Hierarchical clustering by Ward D method on Euclidean distances identifies 14 clusters (k). (C) GSEA plots for gene set extracted from k14 from (B) against MNX1-r patients RNA-seq counts vs MLL, core-binding factors (CBF), or other pediatric AML. (D) Cell type analysis of the intersect of leading edge genes LEGs (n=476) from Fig. S5D in MNX1-r patients (k14 LEGs) (B) using the Panglao DB 2021 database. (E) Representative Giemsa-Wright stained dissociated CFC replating (4th plating) MNX1 haemGx cells from 144h and 216h. Solid black arrows, mast cell precursors; dashed black arrows, mast cells.