Multispecies RNA tomography reveals regulators of hematopoietic stem cell birth in the embryonic aorta

Abstract

The defined location of a stem cell within a niche regulates its fate, behavior, and molecular identity via a complex extrinsic regulation that is far from being fully elucidated. To explore the molecular characteristics and key components of the aortic microenvironment, where the first hematopoietic stem cells are generated during development, we performed genome-wide RNA tomography sequencing on zebrafish, chicken, mouse, and human embryos. The resulting anterior-posterior and dorsal-ventral transcriptional maps provided a powerful resource for exploring genes and regulatory pathways active in the aortic microenvironment. By performing interspecies comparative RNA sequencing analyses and functional assays, we explored the complexity of the aortic microenvironment landscape and the fine-tuning of various factors interacting to control hematopoietic stem cell generation, both in time and space in vivo, including the ligand-receptor couple ADM-RAMP2 and SVEP1. Understanding the regulatory function of the local environment will pave the way for improved stem cell production in vitro and clinical cell therapy.

Graphical abstract

Figure 1.

Tomo-seq allows the identification of specific gene expression patterns restricted to the (ventral) aortic microenvironment. (A-D) Schemes of the embryo species (top panels) and samples (bottom panels) used for tomo-seq analyses. E3 chicken anterior slices (A), 35-day-old human slices (B), E10.5 and E11.5 mouse slices (C), 28- and 40-hpf whole zebrafish embryos (D, left panel), and 28- and 40-hpf zebrafish trunks (D, right panel). The red areas mark the (ventral) aortic region where cluster and HSC emergence occurs. (E-G) Immunostainings of chicken (E), human (F), and mouse (G) embryo slices showing the location of clusters in the aorta. (E) Thick transverse slice of an E3 chicken embryo (slice adjacent to the one used for tomo-seq) stained with anti-MEP21 (red), anti-CD45 (green), and anti-RUNX1 (blue) antibodies. (F) Aorta of human embryo slice 1 stained with anti-CD34 (red), anti-CD45 (green), and anti-RUNX1 (white) antibodies and 4′,6-diamidino-2-phenylindole (blue). (G) Aorta of an E10.5, thick mouse embryo slice stained with anti-CD31 (red) and anti-KIT (green) antibodies. Arrows, dorsal clusters. (H) Global view of a Tg(kdrl:mCherry;cd41:eGFP) reporter zebrafish embryo at 40 hpf; mCherry fluorescence highlights the whole vasculature, GFP fluorescence highlights HSPCs. (I) Trunk region containing the aortic region and yolk sac elongation of the 40-hpf zebrafish embryo shown in panel H (dashed boxed). Single fluorescent pictures (right panels). (J-L) Thick slices from E3 chicken (J), 35 day-old human (K), and E11.5 mouse (L) embryos used for tomo-seq. (M) Whole 40-hpf zebrafish embryo used for tomo-seq. (N) Dissected trunk region of a 40-hpf zebrafish embryo used for tomo-seq. (O) Tomo-seq traces for MPL (red) and SHH (black) along the dorsal-ventral axis of the E3 chicken embryo slice shown in panel J. Colored areas corresponding to the ventral part of the aorta (R1, red) or to the dorsal part of the aorta (R2, blue). The numbers on the x-axes of z-score plots represent sequential individual section numbers (in O,R-S,V,X). (P-Q) ISH for MPL (P) and SHH (Q) on transverse sections of E3 chicken embryo, showing the restricted expression of MPL and SHH to clusters and notochord, respectively. Dashed boxed areas are enlarged in the right panels. (R) Tomo-seq traces for BMP4 (blue), GATA2 (red), and WWP2 (black) along the dorsal-ventral axis of human embryo slice 1, shown in panel K. Colored areas corresponding to the ventral part of the aorta (R1, red) or to the dorsal part of the aorta (R2, blue). (S) Tomo-seq traces for Gata2 (red), Bmp4 (blue), and Shh (black), along the dorsal-ventral axis of the E11.5 mouse embryo slice shown in panel L. Colored areas corresponding to the aortic region (R1, red) or to the regions flanking the aorta (R2 and R3, blue). (T-U) ISH for Bmp4 (T) and Gata2 (U) on E11.5 mouse embryo transverse sections. Dashed boxed areas are enlarged in the right panels. (V) Tomo-seq traces for pax6a (blue), nkx2.1a (green), hoxc11a (black), and apoa1a (red) along the posterior-anterior axis of the 40-hpf zebrafish embryo shown in panel M. (W) WISH showing the expression patterns of pax6a, nkx2.1a, and hoxc11a in 40-hpf zebrafish embryos. (X) Tomo-seq traces for apoa1a (blue), shha (black), and cxcl12b (red) along the dorsal-ventral axis of the 40-hpf zebrafish embryo trunk shown in N. Red area (R1): slices corresponding to the yolk sac elongation region (including the aorta) (V) or the aortic region (X). Blue areas (R2 and R3): slices corresponding to regions flanking the aorta region. (Y) WISH showing the expression patterns of apoa1a, shha, and cxcl12b in 40-hpf zebrafish embryos (left panels). Dashed boxed areas (trunk regions) are enlarged in the middle panels. Embryos were cryosectioned post-ISH to show the expression pattern of apoa1a, shha, and cxcl12b along the dorsal-ventral axis of the embryo trunks (right panels). The total number of embryos analyzed for WISH is indicated on the pictures. A, anterior; Ao, aorta; D, dorsal; Nc, notochord; NT, neural tube; Me, mesonephros; P, posterior; R, region; V, ventral; YS, yolk sac; YSe, yolk sac elongation. Bars represent 50 μm (E-G,Y [right panels]), 100 μm (J,P-Q), 150 μm (L,T-U), and 200 μm (H,K,M,W,Y [left panels]).

Figure 2.

Comparative analysis of tomo-seq data sets identifies species-restricted and interspecies-conserved gene expression in the aortic HSC supportive microenvironment. (A) Venn diagram showing shared and distinct ventrally expressed genes between chicken, human, mouse, and zebrafish tomo-seq data sets. Genes in bold: upregulated (messenger RNA level) in HSPC supportive stromal cell line UG26.1B6 compared with the non-/less supportive stromal cell line UG26.3B5 isolated from the embryonic AGM (according to http://stemniche.snv.jussieu.fr/); underlined: known HSPC regulators in the embryonic aorta. Italic denotes connected to hematopoiesis. *Connected to leukemia. (B-F) Tomo-seq traces for PODXL (blue), ALDH1A2 (green), and PPARGC1A (orange) in the (ventral) aortic region (R1) of an E3 chicken slice (B), a 35-day-old human slice (C), an E10.5 mouse slice (D), a 28-hpf whole zebrafish (E), and a 40-hpf zebrafish trunk (F). (G-J) ISH showing the specific expression patterns of PODXL, ALDH1A2, and PPARGC1A in the (ventral) aortic region (R1) of an E3 chicken anterior slice (G), a 35-day-old human slice (H), an E10.5 mouse slice (I), a 40-hpf whole zebrafish embryo (J, left panel), and a zebrafish trunk slice (J, right panel). The number of embryos analyzed for WISH is indicated on the pictures. (K-L) Ten of the top biological process GO terms (K) and most relevant KEGG pathways (L) enriched in at least 2 species tomo-seq data sets (ie, 2, 3, and 4 species data sets). Bars represent 150 μm (G,I-J [left panels]), 200 μm (H), and 50 μm (J [right panels]).

Figure 3.

Ligand-receptor interactions, biological processes, and pathways active during the cross talk between the cluster cells and the aortic microenvironment. (A) Distribution of the ligands and their corresponding receptors in the aortic microenvironment and clusters based on the microenvironment data set (common to at least 2 species, supplemental Table 9), the chicken cluster data set (supplemental Table 10), and a mouse cluster data set. (B) Secretomics analysis using the microenvironment data set (common to at least 2 species) and the chicken/mouse cluster data sets. Of note, not all molecules are linked, because we used stringent active interaction sources (experiments and database) for better readability of the k-means clustering network. (B′) Enlarged view of the blue node shown in the box in panel B. (C-D) Ten first biological processes in GO terms (C) and most relevant KEGG pathways (D) enriched in the ligand-receptor data sets.

Figure 4.

ADM and its receptor RAMP2 are novel and conserved HSPC regulators. (A) Top left panel: cryosection of 35-day-old human embryo stained with anti-ADM (green) and anti-CD34 (red) antibodies and 4′,6-diamidino-2-phenylindole (DAPI; blue). (A′) ADM expression in the ventral aortic microenvironment. Top right panels: cryosection of E10.5 mouse embryo stained with anti-ADM (green) and anti-CD31 (red) antibodies and DAPI (blue). (A′′) Enlargement of the boxed area showing immunostaining for ADM in the ventral aortic microenvironment close to budding hematopoietic cells. Bottom panels: WISH showing the expression pattern of adm in 40-hpf zebrafish embryos (left panel). Post-ISH embryos cryosectioned to show the expression pattern of adm along the dorsal-ventral axis of the embryo trunks (right panel). (B) Top left (B1): cryosection of E10.5 mouse embryo stained with anti-RAMP2 (red) and anti-RUNX1 (green) antibodies and DAPI (blue). (B1′,B1′′,B1′′′) Enlargement of boxed areas showing immunostaining for RAMP2 in the ventral endothelium and clusters. (B2) Top right panels: ISH showing the expression pattern of RAMP2 in an E3 chicken embryo cryosection. (B2′,B2′′,B2′′′) Enlargement of boxed areas showing RAMP2 expression in the ventral aortic endothelium and clusters. Bottom panels: WISH showing the expression pattern of ramp2 in 40-hpf zebrafish embryos (left panel). Post-ISH embryos show the expression pattern of ramp2 along the dorsal-ventral axis of the embryo trunks (right panel). The number of embryos analyzed for WISH is indicated on the photos. (C-D) The number of runx1⁺ HSPC cells was counted per aorta after WISH in noninjected embryos, embryos injected with mismatch MOs (controls), and embryos injected with blocking MOs (ATG and/or SB) for ADM (C, right panels) and Ramp2 (D, right panels). Representative cases of WISH for runx1 expression in controls and adm (C) and ramp2 (D) injected MOs, analyzed at 40 hpf (left panels). Graphs represent the average number (±SEM) of runx1⁺ cells in the dorsal aorta in each condition (n = 3 independent experiments, n ≈ 50 zebrafish embryos each). (E-F) Number of cmyb⁺ HSPCs per aorta in embryos not injected or injected with mismatch MOs (MO_MM) or blocking MOs (MO_ATG or MO_SB) for adm (E) and ramp2 (F). Example of WISH for cmyb in controls and adm (E) and ramp2 (F) injected MOs, analyzed at 48-hpf (left panels). Graphs represent the average number (±SEM) of cmyb⁺ cells in the aorta (n = 3 independent experiments; n ≈ 30 embryos for each condition). (G-H) Number of cd41⁺ HSPCs per aorta in embryos not injected or injected with mismatch MOs (MO_MM) or blocking MOs (MO_ATG or MO_SB) for adm (G) and ramp2 (H). Fluorescent micrographs of controls and adm (G) and ramp2 (H) injected MOs in the Tg(kdrl:mCherry;cd41:eGFP) fish background, analyzed at 40 hpf (left panels). Graphs represent the average number (±SEM) of cd41⁺ cells in the aorta (n = 2 independent experiments, 11-18 embryos for each condition). (I-J) WISH for ephrinB2a (arterial marker) in controls and adm (I) and ramp2 (J) injected MOs, analyzed at 40 hpf. (K) Graph represents the percentage of CD45⁺ hematopoietic cells in E3 chicken trunks after intracardiac injection of MilliQ water (control), ADM, or ADM-Antag. n = 2 independent experiments. Dot: 2 to 3 trunks pooled. (L-M) In vitro clonogenic assay with cells isolated from E9.5 (L) or E10.5 (M) wild-type AGMs cultured as explants with MilliQ (control) or ADM-Antag. Top: the number of CFU-Cs per AGM embryo equivalent (ee). Bottom: the number of CFU-Cs per 100 000 AGM explant cells. One representative experiment of 2 independent experiments. Bars represent 100 µm (A-H) and 200 µm (I-J). *P < .05; **P < .01; ***P < .001; ****P < .0001; n.s., not significant, by Student t test. BFU-E, burst forming unit-erythroid; CFU-G, CFU-granulocyte; CFU-GEMM, CFU-granulocyte-erythroid-macrophage-megakaryocyte; CFU-GM, CFU-granulocyte-macrophage; CFU-M, CFU-macrophage.

Figure 5.

Svep1 regulates the cellularity and fate of clusters. (A-B) Number of cluster cells per E10.5 (A) and E12.5 (B) Svep1^+/+ and Svep1^−/− aortas (at E10.5: n = 5 and 9 Svep1^+/+ and Svep1^−/− embryos, respectively; at E12.5: n = 7 and 5 Svep1^+/+ and Svep1^−/− embryos, respectively; n = 3 independent experiments). Bars represent means ± SEM. (C-D) Enlarged examples of clusters after maximal projection of E10.5 (C) and E12.5 (D) Svep1^+/+ (top panels) and Svep1^−/− (bottom panels) aortas stained with anti-CD31 (endothelial and hematopoietic marker, red), anti-KIT (hematopoietic marker, green), and anti-RUNX1 (hemogenic endothelium and hematopoietic marker, blue) antibodies. (E-F) Number of cells per clusters composed of more than 5 cells in E10.5 (E) and E12.5 (F) Svep1^+/+ and Svep1^−/− aortas. (G-H) Number of clusters per E10.5 (G) and E12.5 (H) Svep1^+/+ and Svep1^−/− aortas. (I) Hematopoietic repopulation analyses after injection of AGM cells isolated from Svep1^+/+ and Svep1^−/− E11 embryos (n = 1-3 independent experiments). Numbers above bars indicate the number of mice repopulated/number of mice injected. Dose of injected cells is indicated as embryo equivalent (ee). (J) Percentage of chimerism in peripheral blood for each injected mouse in panel I, 4 months after transplantation. Dot: 1 transplanted recipient mouse. Red dots: mice used to perform secondary transplantations (labeled a-d, here and in subsequent panels). Dashed line: limit of positivity (>5% of chimerism by flow cytometry). Red line: chimerism average. (K) HSC frequency determined based on the transplantation results in panel I. HSC frequency per AGM was estimated by Poisson statistics applied to the percentage of nonrepopulated recipients (y-axis) and the number of injected cells (x-axis). (L) Long-term donor-derived lymphoid (T cells: CD4⁺, CD8⁺; B cells: B220⁺) and myeloid (CD11b⁺, GR1⁺) contribution in the blood of reconstituted recipients in panel J. Each bar represents a single recipient. (M) Percentage of chimerism in peripheral blood for each secondary recipient mouse transplanted with 3 × 10⁶ bone marrow cells isolated from the primary reconstituted recipients (a-d, indicated in panel J) at 4 months after transplantation. Dot: 1 transplanted repicient mouse. Dashed line: limit of positivity (>5% of chimerism by flow cytometry). Red line: chimerism average. (N) Long-term donor-derived lymphoid (T cells: CD4⁺, CD8⁺; B cells: B220⁺) and myeloid (CD11b⁺, GR1⁺) contribution in the blood of secondary reconstituted recipients shown in panel M. Each bar represents a single recipient. (O-P) Number of runx1⁺ (O) and cmyb⁺ (P) HSPCs per aorta at 40 and 48 hpf, respectively, were determined after WISH in svep1^+/+ and svep1^−/− zebrafish embryos. Representative cases of WISH for runx1 and cmyb expressions in svep1^+/+ and svep1^−/− zebrafish embryos (left side of each graph). Graphs represent the average number (±SEM) of runx1⁺ and cmyb⁺ cells in the dorsal aorta (n ≈ 50 zebrafish embryos each). (Q) Fluorescent micrographs of 2 svep1^−/− (fli1a:GFP;flt1^enh:RFP) embryos analyzed at 40 hpf. Global view (top panel, embryo 1) and enlargement of the dorsal aorta area (bottom panel, embryo 2). (R) WISH for ephrinb2a in svep1^+/+ (top panel) and svep1^−/− (bottom panel) embryos, analyzed at 40 hpf. The total number of embryos analyzed for WISH is indicated on the photographs. (S) The rag1⁺ thymic areas at 4 dpf were determined after WISH in svep1^+/+ and svep1^−/− zebrafish embryos. Representative cases of WISH for rag1 expression in svep1^+/+ and svep1^−/− zebrafish embryos (left panels). Graphs represent the average of rag1⁺ thymic area (±SEM) (n ≈ 10 zebrafish embryos). *P < .05; **P < .01; ***P < .001; ****P < .0001; n.s., not significant, by Student t test. Bars represent 50 µm (C-D), 100 µm (O-Q), and 200 µm (R-S).