Vascularisation is not necessary for gut colonisation by enteric neural crest cells

Abstract

The vasculature and nervous system share striking similarities in their networked, tree-like architecture and in the way they are super-imposed in mature organs. It has previously been suggested that the intestinal microvasculature network directs the migration of enteric neural crest cells (ENCC) along the gut to promote the formation of the enteric nervous system (ENS). To investigate the inter-relationship of migrating ENCC, ENS formation and gut vascular development we combined fate-mapping of ENCC with immunolabelling and intravascular dye injection to visualise nascent blood vessel networks. We found that the enteric and vascular networks initially had very distinct patterns of development. In the foregut, ENCC migrated through areas devoid of established vascular networks. In vessel-rich areas, such as the midgut and hindgut, the distribution of migrating ENCC did not support the idea that these cells followed a pre-established vascular network. Moreover, when gut vascular development was impaired, either genetically in Vegfa(120/120) or Tie2-Cre;Nrp1(fl/-) mice or using an in vitro Wnt1-Cre;Rosa26(Yfp/+) mouse model of ENS development, ENCC still colonised the entire length of the gut, including the terminal hindgut. These results demonstrate that blood vessel networks are not necessary to guide migrating ENCC during ENS development. Conversely, in miRet(51) mice, which lack ENS in the hindgut, the vascular network in this region appeared to be normal suggesting that in early development both networks form independently of each other.

Keywords: Blood vessels; Enteric nervous system; Migration; Neural crest cells; Vascular system.

Fig. 1

Labelling of the chick enteric nervous and vascular systems. In order to label ENCC and blood vessels, the vagal region of the neural tube is surgically removed from E1.5 (HH10) wild type host chicken embryos (A) and replaced with the equivalent region of neural tube from a GFP+ donor chicken (B). All vagal-derived ENCC are therefore GFP+ [arrows] (C). In the same neural tube-transplanted embryo at E4 (HH23), DiI is injected into the blood stream [black arrow] (D), to reveal the developing blood vessel network (E) Br: Brain; H: Heart; LB: limb bud; A: Allantois. Scale bar A and B=200 μm; C=200 μm; D and E=2 mm.

Fig. 2

Wholemount preparation of chick gut and respiratory system at E4.5 (HH24) and E5.5 (HH28), showing distinct patterns of ENS and vascular network development. (A and C) The vascular networks develop from the pulmonary supply to the lungs, and from the coeliac supply to the stomach and the superior and inferior mesenteric supplies to the distal gut. The oesophagus and ventral midgut are devoid of blood vessels at E4.5 (arrowheads). (B and C) ENCC colonise the gut in a proximo-distal direction. ENCC migrate along and within the oesophagus and midgut despite the absence of vasculature (arrowheads). (E–H) High magnification images of the migration front (boxed in A–C) in the vascular rich region of the umbilicus. ENCC are both in the interspace between vessels (arrowhead) and in close association with blood vessels (arrows). (I–K) High magnification images at E5.5 (HH28). (I) The oesophagus is avascular yet ENCC formed an ENS. (J) ENCC have colonised the majority of the midgut including regions devoid of capillaries (J, arrowheads). (K) At the migration front, most mesenteric projections are perpendicular to the ENCC path. ENCC are found both in the interspace between vessels (arrowhead) and in close association with blood vessels (arrows). Z: z stack projection 3D reconstruction. XZ: 90° rotation around the X axis. Scale bar A–D=0.5 mm; E–H=100 μm; I–K=100 μm.

Fig. 3

Formation of the enteric nervous and vascular systems in the embryonic quail digestive tract. (A) Confocal image of a wholemount preparation of oesophagus at E4.5 (quail stage Q25), labelled with QH1 and HNK1, showing extensive colonisation of the oesophagus by ENCC despite the lack of vasculature. (B) Confocal image of a wholemount preparation of midgut at E4.5 (Q25), labelled with QH1 and HNK1, at the level of the ENCC migration front. XZ projection reveals the localisation of both networks within the same layer of the midgut. Despite this co-localisation, the migration paths of the ENCC do not systematically follow the honeycomb pattern of the QH1+ network (arrowheads). (C) Confocal image of a wholemount preparation of hindgut at E6.5 (Q30), showing the ENCC migration front half way along the hindgut (arrow). Upper box: high magnification image in the region of the migration front reveals a lack of co-patterning between the ENCC and the QH1+ cell network. Both networks are sometimes closely co-localised (arrow) and sometimes apart (arrowheads). Lower box: high magnification demonstrating a lack of co-patterning between the projections of the nerve of Remak and the QH1+ cell network (arrowheads). Rem: Remak nerve. Z: z stack projection 3D reconstruction. XZ: 90° rotation around the X axis. Scale bar A and B=100 μm; C=500 μm.

Fig. 4

Vascular and enteric network formation in the chick hindgut. (A–C) Confocal imaging of the hindgut at E7.5 (HH32). ENCC are located midway along the hindgut, which already has a dense vascular network. At this stage, the most advanced cells of the migration front (◀1) are present within the presumptive SMP and do not overlap with the vasculature, which is located in the outer layers of the gut (C, insets). XZ and YZ orthogonal views confirm the lack of overlap at the level of the SMP (C, insets). A second group of cells (◀2) is located within the presumptive MP where pre-existing blood vessels are present (C, inset). Despite overlap between blood vessels (^⁎) and the developing MP, both networks are not systematically juxtaposed (MP inset in C). (D–F) Confocal imaging of the hindgut at E11.5 (HH37-38), the hindgut is fully colonised by both networks. (F, inset I) Transverse sections demonstrate ENCC and blood vessels at both the myenteric and the submucosal levels. (F, inset II) Transverse sections in the most distal part of the hindgut. The presumptive SMP is sparsely populated by ENCC whereas the mucosal vascular supply is very dense. (G and H) High magnification reconstructions reveal a very robust MP intercalated with many blood vessels. The SMP is thinner, and crossed by blood vessels connecting to a dense mucosal blood network medial to the SMP. (I) Image at the level between the MP and the SMP, showing multiple neuronal projections connecting both plexuses. Blood vessels connections are less numerous and are either closely associated with the neuronal projections (arrows) or isolated (arrowheads) with no systematic associations. MP: myenteric plexus; SMP: submucosal plexus. D→V: Dorso-ventral axis. Z: z stack projection 3D reconstruction. XZ: 90° rotation around the X axis. Scale bar A–C and D–F=0.5 mm; G and I=100 μm.

Fig. 5

ENCC migrate to the terminal hindgut in Vegfa^120/120 and Tie2-Cre;Nrp1^fl/− mutant mice with defective gut vascularisation. (A–F) transverse gut sections from E13.5 mice, labelled with the neuronal marker TuJ1 and the vascular marker endomucin. (G–J) Wholemount preparations of Tie2-Cre;Nrp1^{f /−} and Vegfa^120/120 terminal hindgut at E14.5 and E16.5 respectively. (A and B) Wild type controls have an extensive blood vessel network in the midgut (A), and hindgut (B), with the ENS apparent as an almost continuous ring of TuJ1+ cells encircling the gut. Vegfa^120/120 mice have less extensive vascular networks in both the midgut (C) and hindgut (D), but the ENS is similar to controls in both gut regions. (E and F) Tie2-Cre;Nrp1^fl/− mice exhibit more prominent vascular defects, particularly in the hindgut. The ENS is less extensive but ENCC are present within the terminal hindgut. (G–H) At E14.5, controls have an extensive blood vessel network in the hindgut compared to Tie2-Cre;Nrp1^fl/−mice (red arrowheads). Although less extensive, ENCC have migrated to the terminal end of the gut. (I-J). At E16.5, littermate controls have an extensive blood vessel network in the hindgut compared to Vegfa^120/120 mice (red arrowheads). The ENS of Vegfa^120/120 mice has a disorganised architecture. Z: z stack projection 3D reconstruction. XZ: 90° rotation around the X axis. P→D: Proximo-distal axis, identical for G–J. Scale bar=50 μm.

Fig. 6

ENCC colonise the entire length of the mouse embryo hindgut in vitro in the absence of a vascular network. (A–D) Wholemount and (E–H) transverse sections through Wnt1-Cre;R26R^yfp/+ mouse gut labelled with endomucin (vasculature) and YFP (ENCC). (A, E, and I) At E11.5, only occasional ENCC are present in the hindgut (A, arrowhead, close up in I). E11.5 guts were dissected and placed in culture for up to 4 days. (B, F, J, and K) Confocal images showing guts after 2 days and (C, G, and L) after 4 days of culture. The vascular network collapse is visible by reduced endomucin staining at 2 days (B and F) and 4 days (C and G), and by the presence of apoptotic rounded nucleus in endomucin+ tissue (J, arrows, inset from F). Despite the collapse of the vascular network, ENCC fully colonise the entire length of the hindgut, as seen by YFP staining (B, C, F, and G, arrowheads). Chain migrating ENCC occur with and without the presence of a vascular network (B, F, and K). (C) Wholemount (G) section and (L) high magnification confocal imaging show that, after 4 days in culture, ENCC form a circular plexus comparable with E15.5 non-cultured controls (D, H, and M). MG: midgut; Ce: caecum; HG: hindgut; PG: pelvic ganglion. ^⁎Inferior mesenteric supply. Z: z stack projection 3D reconstruction. XZ/YZ: 90° rotation around the X or Y axis. Scale bar A–D and E–H=100 μm; I–M=50 μm.

Fig. 7

The enteric vascular system appears normal in the aganglionic terminal hindgut of miRet⁵¹ mutant mice at E19.5. (A–D) wholemount preparations of the terminal hindgut of miRet⁵¹ and control littermates at E19.5 labelled with the neuronal marker TuJ1, the vascular marker endomucin and the nuclear marker DAPI. (A and B) confocal images at the level of the myenteric plexus reveal an extensive ENS in control littermates (A) whereas in miRet⁵¹ mice (B) only sparse extrinsic nerves are present. (C–D) Confocal images at the level of the mucosal crypts demonstrate similar vascular networks in control littermates and miRet⁵¹ mice. Scale bar=250 μm.

Fig. S1

Formation of the neural crest-derived nervous and vascular systems in the embryonic chicken lungs. (A–C) Wholemount preparation of the lungs at E5.5 (HH28). NCC are located within the medio-dorsal region of the lungs, whereas the vascular system is located on the lateral side. In the most distal portion of the lungs, the NCC migration front is advanced more distally than the vascular network (C, arrowheads). Some isolated NCC are associated with the vascular system (B and C; arrows). (D) High magnification confocal image of the lungs at E11.5 (HH37-38). NCC form a thin ganglionated network intertwined with the vascular system.

Fig. S2

Formation of the enteric nervous and vascular systems in embryonic chicken oesophagus at E7.5 and E11.5. (A–C) Confocal imaging of the oesophagus at E7.5 (HH32) reveals very sparse vasculature limited to isolated vessels from the adjacent mesentery. (D–E) Confocal imaging of the oesophagus at E11.5 (HH37-38). The vascular system is present on the periphery of the oesophagus, separate from the ENS, and centripetal vessels extend across the ENS (E and F; arrowheads). Congruence is observed between blood vessels and extrinsic vagus nerve fibres (D and F; arrows).

Fig. S3

In the embryonic chicken, at the level of the stomach, a rich vascular bed develops underneath a dense peripheral enteric network. (A–C) Images and (D–F) transverse sections of the stomach at E5.5 (HH28), showing a rich vascular bed throughout the tissue, whereas the ENCC network is located more superficially, with the exception of the tendon attachment region where ENCC never migrate (dotted line; Ten). Single ENCC are present deep in the stomach in blood vessel rich regions, although without apparent association with blood vessels (A–F, arrowheads). (G–I) Images of the stomach at E7.5 (HH32) showing a similar pattern as E5.5 (HH28), with denser networks. (J–L) High magnification imaging of the stomach at E11.5 (HH37-38) showing a robust ENS intercalated with large diameter vessels, with the main vascular network located median to the ENS. Z: z stack projection 3D reconstruction. XZ: 90° rotation around the X axis.

Fig. S4

ENCC colonise the entire length of the embryonic mouse hindgut in vitro, in the absence of PECAM and endomucin+ cells. (A–L) Transverse sections through Wnt1-Cre;R26R^yfp/+ labelled with endomucin (vasculature) and YFP (ENCC). At E11.5, the superior mesenteric artery is PECAM+ and endomucin −ve (A, E, and I; arrow). During in vitro culture, the vascular network collapses due to the lack of blood flow, as seen by reduced endomucin and PECAM staining at 2 days (B and F) and 4 days (C and G). Despite the absence of PECAM and endomucin+ cells in the gut, ENCC fully colonise the entire length of the hindgut, as seen by YFP staining (B, C, F, and G, arrowheads). Chain migrating ENCC occur with and without the presence of PECAM and endomucin cells (B, F, and K). MG: midgut; HG: hindgut.