A strategy to discover new organizers identifies a putative heart organizer

Abstract

Organizers are regions of the embryo that can both induce new fates and impart pattern on other regions. So far, surprisingly few organizers have been discovered, considering the number of patterned tissue types generated during development. This may be because their discovery has relied on transplantation and ablation experiments. Here we describe a new approach, using chick embryos, to discover organizers based on a common gene expression signature, and use it to uncover the anterior intestinal portal (AIP) endoderm as a putative heart organizer. We show that the AIP can induce cardiac identity from non-cardiac mesoderm and that it can pattern this by specifying ventricular and suppressing atrial regional identity. We also uncover some of the signals responsible. The method holds promise as a tool to discover other novel organizers acting during development.

Figure 1. Transcriptome comparison of known organizers reveals genes co-expressed in organizers and in the Anterior Intestinal Portal endoderm.

(a–c) Experimental design. (a) Hensen’s node (red box) was compared with posterior primitive streak (green box); (b) notochord and ventral neural tube (red box) was compared with dorsal neural tube (green box) and (c) posterior wing bud (red box) was compared with anterior wing bud (green box). (d–i) One gene enriched in organizers is PCSK6 (d–f, red boxes, red arrowhead in f-inset) and SOCS2 exemplifies a depleted gene (g–i, green boxes, green arrowhead in i-inset). Fold-change values are indicated. (j–n) Examples of organizer-enriched transcripts expressed in the anterior intestinal portal endoderm (AIP) (red arrowheads): NRP1 (j), FBLN7 (k), SHH (l), KIRREL3 (m) and VTN (n). Scale bars, 0.5 mm in whole-mounts and 0.1 mm in sections.

Figure 2. AIP induces cardiac and ventricular identity in non-cardiac mesoderm.

(a) AIP or control (Cont.), lateral endoderm from a HH8 quail co-cultured with anterior-medial mesoderm (#3) from a HH5 chick embryo overnight in the anterior area opaca of a host chick. (b–g) Regional myocardium markers VMHC1/MYH15 (b), IRX4 (c), NPPB (d) and GJA5 (e) are induced in #3-mesoderm by quail (brown) AIP (red arrowheads), but not by control endoderm (Cont., black arrows), whereas AMHC1 (f) and SHOX2 (g) are not induced (black arrows). (h) In vitro co-culture of #3-mesoderm with GFP-AIP (j,m,p,s,v,y,b’,e’) or -Cont. (k,n,q,t,w,z,c’,f’) for 24 (j,k,m,n) or 48 h (p,q,s,t,v,w,y,z,b’,c’,e’,f’). No expression is seen in #3-mesoderm cultured alone (i,l,o,r,u,x,a’,d’). AIP (brown) induces MYOCD (j), NKX2.5 (m), VMHC1/MYH15 (p), IRX4 (s), NPPB (v) and GJA5 (y) in #3-mesoderm (red arrowheads), but Cont. does not (brown, k,n,q,t,w,z, black arrows). Neither AMHC1 nor SHOX2 are induced (black arrows) by AIP (b’,e’) or by Cont. (c’,f’). *P≤0.05, **P≤0.005, ***P≤0.0005 using two-tailed Fisher's exact test. Scale bars, 0.5 mm in whole-mounts and 0.1 mm in insets, explants and sections.

Figure 3. AIP specifies ventricular regional identity and represses atrial character.

(a) AIP (green, c,d,h,i,m,n; brown, e–g,j–l,o–q) from a HH8 GFP chick was grafted heterotopically onto cardiogenic mesoderm of a HH5 wild-type chick host (c,h,m, T0) and grown for 12 h (d,i,n, T12). (c–q) AIP expands VMHC1/MYH15 (e–g) and IRX4 (j–l) expression (red arrowheads) and abolishes AMHC1 (o–q, black arrows). (b) In vitro co-culture of #1-mesoderm from a HH5 chick with HH8 GFP-AIP or control (Cont.) endoderm (brown) for 48 h. AMHC1 expression in #1-mesoderm (r, red arrowhead). AIP represses AMHC1 in #1-mesoderm (s, black arrow), Cont. does not (t, red arrowhead). *P≤0.05 using two-tailed Fisher's exact test. Scale bars, 0.5 mm in whole-mounts and 0.1 mm in explants and sections.

Figure 4. Secreted molecules induce ventricular markers.

(a–c) HH5 quail (brown) #3-mesoderm co-cultured in a chick host with cell pellets transfected with NRP1+FBLN7+KIRREL3+VTN (red) either overnight (a,b) or for 6–9 h (c). VMHC1/MYH15 and NPPB are induced in #3-mesoderm (a,b, red arrowheads) but MYOCD is not (c, black arrows). Control pellets (red; pCAB) do not induce VMHC1/MYH15, NPPB or MYOCD (a–c, black arrows). Scale bars, 0.5 mm in whole-mounts and 0.1 mm in insets and sections.

Figure 5. Successive roles of the endoderm in heart induction and patterning.

Each diagram represents a stage in development; in each, tissue interactions are depicted on the left, signals on the right. (a) Initially (HH 4–6), the anterior-lateral endoderm (including some prospective AIP, green) is required for the adjacent cardiac mesoderm (light blue) to express NKX2.5, MEF2C and MYOCD. BMP2 and FGF8 signalling from the anterior-lateral endoderm induce NKX2.5, GATA, MEF2 and HAND transcription factors. The anterior endoderm expresses WNT inhibitors including Crescent, which can induce NKX2.5 (ref. 49). (b) At HH8, progenitor cells are becoming determined as either atrial or ventricular. The AIP endoderm induces ventricular character and represses atrial identity. Signals from the AIP endoderm, including NRP1, FBLN7, KIRREL3 and VTN described here (red text and arrows) and previously reported BMP2, FGF4 and Wnt inhibitors (black text and arrows), induce some ventricular markers, which are integrated with atrial-inducing retinoic acid signalling from the posterior lateral plate mesoderm. The signals from the AIP endoderm that repress atrial identity are unknown. (c) At HH9, when regional markers of anterior–posterior heart tube patterning begin to be expressed, AIP endoderm induces ventricular markers VMHC1/MYH15, IRX4, NPPB and GJA5 (the latter expressed from HH12) and represses the atrial marker AMHC1. Four secreted molecules from the AIP endoderm, NRP1, FBLN7, KIRREL3 and VTN can induce VMHC1/MYH15 and NPPB.