NRP1 and NRP2 cooperate to regulate gangliogenesis, axon guidance and target innervation in the sympathetic nervous system

Abstract

The sympathetic nervous system (SNS) arises from neural crest (NC) cells during embryonic development and innervates the internal organs of vertebrates to modulate their stress response. NRP1 and NRP2 are receptors for guidance cues of the class 3 semaphorin (SEMA) family and are expressed in partially overlapping patterns in sympathetic NC cells and their progeny. By comparing the phenotypes of mice lacking NRP1 or its ligand SEMA3A with mice lacking NRP1 in the sympathetic versus vascular endothelial cell lineages, we demonstrate that SEMA3A signalling through NRP1 has multiple cell-autonomous roles in SNS development. These roles include neuronal cell body positioning, neuronal aggregation and axon guidance, first during sympathetic chain assembly and then to regulate the innervation of the heart and aorta. Loss of NRP2 or its ligand SEMA3F impaired sympathetic gangliogenesis more mildly than loss of SEMA3A/NRP1 signalling, but caused ectopic neurite extension along the embryonic aorta. The analysis of compound mutants lacking SEMA3A and SEMA3F or NRP1 and NRP2 in the SNS demonstrated that both signalling pathways cooperate to organise the SNS. We further show that abnormal sympathetic development in mice lacking NRP1 in the sympathetic lineage has functional consequences, as it causes sinus bradycardia, similar to mice lacking SEMA3A.

Fig. 1

Co-expression of neuropilins in the NC lineage. (A–F) Transverse cryosections of E9.5 or E11.5 wild-type embryos at forelimb level, immunolabelled for NRP1 or NRP2 and markers for NC cells (p75), sympathetic precursors (MASH1) or sympathetic neurons (TH); half of each section is shown; areas in dashed boxes are shown in higher magnification as separate channels and overlays to the right of each panel. Sympathetic NC cells aggregating at the dorsal aorta were NRP1- and NRP2-positive at both anterior (double arrowheads; A–C) and posterior (double arrowheads; E and F) somite level; migratory intermediate wave NC cells were also NRP1- and NRP2-positive (arrowheads; A,B,C); NRP1-postive vascular endothelium is indicated (white arrows in A and E). (G and H) Sympathetic neurons in the definite sympathetic ganglia were NRP1- and NRP2-positive. Abbreviations: neural tube, nt; DRG anlagen, d; dermomyotome, dm; dorsal funiculus, df; ventral funiculus, vf; dorsal aorta, da; blood vessel, bp. Scale bars: A–F: 100 μm, G and H: 200 μm.

Fig. 2

Loss of SEMA3A and SEMA3F disrupts SNS patterning. (A) Schematic of the SNS in a E12.5 mouse embryo; sympathetic neurons cluster into sympathetic ganglia (sg) that align either side of the aorta; they are connected by sympathetic axons (sa) into the trunk sympathetic chains (sc) and to the cervical ganglia (cg), including the stellate ganglia (stg) and superior cervical ganglia (scg). (B–F) Ventral view of the SNS in E12.5 embryos after wholemount TH immunolabelling and removal of the internal organs and scg; areas in dashed boxes are shown in higher magnification to the right of each panel. Whilst ganglia in wild-types (B) appeared compact, ganglia were more dispersed in Nrp1-null (C) and Sema3a-null (D) mutants (white arrowheads in C″ and D″); these genotypes also contained chains of medially placed neurons (wavy arrows in C,D), ectopic axons extending across the midline (black arrows in C″ and D″) and ectopic distal neurons (black arrowheads in C′ and D′). Sema3f-null mutants (E) contained ectopic neurons on top of the aorta (white arrows in E′ and E″) and extended aberrant axons on top of the aorta (E″). The ganglia of Sema3a/Sema3f-null mutants (F) were poorly compacted (white arrowhead in F″), and there were medially (wavy arrow in F) and distally displaced neurons (arrowhead in F′); ectopic neurons and axons were present on top of the aorta (white arrow in F″). Scale bar: 500 μm.

Fig. 3

Loss of NRP1 from the NC, but not endothelial cell lineage, disrupts sympathetic NC migration and sympathetic chain organisation. (A–C) Wholemount in situ hybridisation of E9.5 embryos for Sox10. In wild-types (A) and mutants lacking NRP1 in endothelial cells (B), NC cells migrated normally through the anterior somite, avoiding the posterior somite (white brackets) and aggregated into the primary sympathetic ganglia (white arrowhead). Mutants lacking NRP1 in the NC lineage (C) showed excessive NC cell migration through the intersomitic furrows (black wavy arrow) and delayed assembly of the primary sympathetic ganglia (asterisk). (D–E) Wholemount immunolabelling of E13.5 embryos for TH to visualise the sympathetic chains. In controls (D) and mutants lacking NRP1 in endothelial cells (E), the sympathetic chains appeared to be organised into tight ganglia. In the absence of NRP1 in the NC lineage (F), sympathetic ganglia appeared less compact (white arrow), with medially displaced ganglia (wavy arrow) and ectopic axon extension towards the midline (black arrows); ectopic sympathetic neurons were also present, particularly at cervical level (black arrowheads). Scale bars: A–C, 250 μm; D–F, 500 μm.

Fig. 4

NRP1 and NRP2 cooperate to pattern the sympathetic chains. (A–D) Wholemount TH immunolabelling of E13.5 sympathetic chains in littermate embryos from matings of Wnt1-Cre Nrp1^+/−Nrp2^+/− with Nrp1^fl/flNrp2^+/− mice; the position of the aorta is indicated with (a); boxed areas are shown at high magnification in (A′–D″). In wild-types (A), areas distal to the chains were free from sympathetic neurons (A′) and chain ganglia were compact (A″). Nrp2-null mutants (B) contained a few ectopic neurons distal to the cervical ganglia (black arrowhead in B), some ganglia were disorganised at both cervical and trunk level (white arrows in B′ and B), and a few axons extended towards or on top of the aorta (black arrows in B,B″). Mutants lacking NRP1 in the NC lineage (C) contained ectopic ganglia in medial positions (wavy arrow in C), distal ectopic neurons (black arrowhead in C′), poorly compacted ganglia (white arrowhead in C″) and axons that projected across the midline (black arrow in C″). The SNS of mutants lacking both NRP2 and NRP1 in the sympathetic lineage (D) contained medial ganglia (wavy arrow in D), distal ectopic neurons (black arrowhead in D′) and dispersed ganglia (white arrowhead in D″), like other types of Nrp1 mutants; they also had ectopic ganglia that extended axons on top of the aorta (white arrow in D″). Scale bar: 500 μm.

Fig. 5

Abnormal target organ innervation in mice lacking NRP1 in the NC lineage. (A–D) TH immunolabelling of sympathetic fibres on the ventricular surface of E17.5 hearts. (E and F) TH immunolabelling of 20 μm cryosections through the P7 ventricular wall reveals the sympathetic fibre distribution between epicardium (epi) and endocardium (endo); double-headed red arrows indicate the width of the subepicardial (1) and subendocardial (3) areas and the intervening myocardial area (2) quantified in (G). (H) Mean sympathetic fibre density in the subendocardium of the indicated genotypes, normalised to fibre density in the subepicardium of the same genotype (set to a value of 1, indicated with a dashed line; n=4). (I and J) 100 μm transverse vibratome sections through P7 aorta, double immunolabelled for TH (green) and alpha smooth muscle actin (SMA, red). (K) Quantification of sympathetic fibres in 20 μm transverse cryosections through P7 aorta; n=4. Error bars: s.e.m.; the asterisks indicate a P-value <0.05. Scale bars: A–D 500 μm; E,F 20 μm; I,J 100 μm.

Fig. 6

Loss of NRP1-mediated SNS patterning leads to sinus bradycardia. (a) An idealised schematic representation of a mouse ECG trace; abbreviations: P, Q, R, S, T indicate the respective ECG waves; QRS, complex of the Q, R and S waves; QT or PR indicate intervals between the respective waves. (b) The RR interval, which is inversely related to heart rate, is significantly prolonged in Nrp1^fl/−Wnt1-Cre mice compared to control Nrp1^fl/+ and Nrp1^fl/− littermates lacking Cre (*** indicates P<0.001). (c) Scatterplot of individual RR-intervals (crosses) in a 2-minute ECG recording from two representative littermate mice of Nrp1^fl/− and Nrp1^fl/−Wnt1-Cre genotypes; the squares indicate tight clustering of the values for each mouse. The form factor (Y axis) provides a measure of the consistency of QRS complex morphology; clustering around a specific RR-interval indicates a normal QRS complex.