XB-ART-56593Dev Biol April 15, 2020; 460 (2): 108-114.
Heparan sulfate proteoglycans regulate BMP signalling during neural crest induction.
Bone morphogenetic protein (BMP) signalling is key to many developmental processes, including early regionalisation of the ectoderm. The neural crest is induced here by a combination of BMP and Wnt signals from nearby tissues with many secreted factors contributing to its initial specification at the neural plate border. Gremlin 1 (Grem1) is a secreted BMP antagonist expressed in the neural crest in Xenopus laevis but its function here is unknown. As well as binding BMPs, Grem1 has been shown to interact with heparan sulfate proteoglycans (HSPGs), a family of cell surface macromolecules that regulate a diverse array of signalling molecules by affecting their availability and mode of action. This study describes the impact of HSPGs on the function of Grem1 in neural crest induction. It shows for the first time that Grem1 is required for neural crest development in a two-step process comprising an early HSPG-independent, followed by a late HSPG-dependent phase.
PubMed ID: 31883440
PMC ID: PMC7196931
Article link: Dev Biol
Genes referenced: emx1 foxd3 grem1 msx1 otx2 pax3 sdc2 sox10 sox2 tbxt twist1 zic1
Morpholinos: grem1 MO1 grem1 MO2
Article Images: [+] show captions
|Fig. 1. Endogenous grem1 expression pattern, and effects of knockdown and overexpression on the neural crest. (A–C) Stage 36 grem1 expression pattern. (A) Whole-mount stained embryo showing strong expression in the kidney primordia (k), posterior roofplate (rp) and tailbud (tb), and neural crest (nc). (B-C) Transverse sections from an embryo like the one shown, with approximate anteroposterior locations marked by dashed lines in A. (B) In the most anterior region, grem1 is expressed in neural crest cells ventral to the developing eye fields (e). (C) In the pharyngeal region, expression is seen in neural crest cells flanking the neural tube (nt) and invading the tissue around the pharyngeal pouches (pp). (D–E) Effect of grem1 overexpression on the neural crest. (D) Cranial neural crest (nc) around stage 18 is marked by the expression of twist1, flanking the anterior neural plate. (E) Overexpression by injection of Xenopus grem1 RNA causes a substantial expansion (arrowhead) of the neural crest (n = 12/19). (F–G) Effect of grem1 knockdown. (F) At stage 21 twist1 marks the neural crest as it begins to surround the eye fields (e) and to migrate in two more posterior streams. It is unaffected by unilateral injection of a standard control morpholino (n = 30/30). (G) Knockdown of grem1 by injection of a translation-blocking morpholino severely impairs neural crest formation (asterisk, n = 80/93). (H) Injection of murine Grem1 RNA is also able to expand the domain of twist1 expression (arrowhead), indicating functional conservation between the mouse and Xenopus orthologues (n = 59/74). (I–K) Rescue of neural crest formation with murine Grem1. (J) Bilateral grem1 knockdown depletes the neural crest from both sides (asterisks). (K) Unilateral rescue using murine Grem1 RNA, mismatched with the morpholino, which only targets the Xenopus orthologue, is able to restore neural crest development (arrowhead, n = 14/20). (L–Q) Effect of gain and loss of grem1 on other neural crest markers. (L) Control neural crest expresses sox10 (n = 25/25). (M) Unilateral grem1 knockdown abolishes expression on the injected side (asterisk, n = 19/24). (N) Unilateral overexpression expands the domain of sox10 (arrowhead, n = 21/30). (O) Control neural crest also expresses foxd3. (P) Knockdown similarly abolishes expression (asterisk, n = 24/24). (Q) Likewise, the domain is expanded by overexpression (arrowhead, n = 12/15). Embryo in A is shown in side view with anterior towards the left, sections in B-C are transverse with dorsal towards the top, and embryos in D-Q are shown in frontal view.|
|Fig. 2. Grem1 affects regionalisation of the ectoderm and is required for neural plate border formation. (A, D, G, J, M, P, S) Control embryos, (B, E, H, K, N, Q, T) unilateral Grem1 knockdown, (C, F, I, L, O, R, U) unilateral murine Grem1 overexpression. (A–I) Stage 17 embryos assayed for neural plate markers, (J–R) stage 14 embryos assayed for neural plate border markers, and (S–U) stage 11 embryos assayed for mesoderm markers. (A–C) The neural plate (np) expresses sox2, which appears to be unaffected by control morpholino (n = 25/25) or loss of Grem1 (n = 26/26) but is expanded (arrowhead) by overexpression (n = 20/33). (D–F) Prospective midbrain and forebrain (fb) regions express otx2, which is unaffected by control morpholino (n = 24/24), while the most anterior part is reduced by knockdown (n = 5/25). It is unaffected by Grem1 overexpression (n = 16/16). (G–I) The future telencephalon (t) expresses emx1, which is unaffected by control morpholino (n = 17/17), requires Grem1 (asterisk, n = 17/25), but is unaffected by overexpression (n = 18/18). (J–L) The neural plate border (npb) is marked by msx1, which is unaffected by control morpholino (n = 19/19), lost (asterisk) after Grem1 knockdown (n = 30/45), and expanded (arrowhead) by overexpression (n = 21/26). (M–O) It is also marked by pax3, which is unaffected by control morpholino (n = 19/19), does not require Grem1 expression (n = 21/23), but can be similarly expanded (arrowhead) by overexpression (n = 20/24). (P–R) Likewise, zic1, which is expressed additionally in the anterior neural fold, is unaffected by control morpholino (n = 18/18) or knockdown (n = 25/25) but expanded (arrowhead) by exogenous Grem1 (n = 21/23). (S–U) The mesoderm (m) expresses tbxt (S), which is unaffected by either Grem1 knockdown (n = 8/8) or overexpression (n = 8/8). Embryos in A-R are shown in frontal view with dorsal-posterior towards the top. Embryos in S–U are shown in vegetal view.|
|Fig. 3. The activity of Grem1 in vivo is dependent on HSPG binding. (A) Sequence alignment of the cysteine knot domains of the wild-type mouse protein and two mutant Grem1 (MGR) constructs with impaired HS binding. Residues of the characteristic eight-membered cysteine knot are highlighted in yellow, with key basic residues mediating HS affinity in green, and non-conservative substitutions of these residues in red. (B–F) Axis induction assay by ventral misexpression of the indicated genes, traced by coinjected gfp (B–E), and assessed morphologically (B′-E’) for formation of a conjoined secondary axis. (B) Injection of gfp alone has no effect. (C) 150 pg Grem1, as expected, is able to induce a secondary axis (arrowhead). (D) This ability is abolished in MGR5. (E) The same goes for MGR6. (F) Graph summarising results of axis induction assay across a range of doses: 200 pg gfp (n = 0/92), 50 pg Grem1 (n = 73/78), 150 pg Grem1 (n = 31/32), 300 pg Grem1 (n = 7/7), 500 pg grem1 (n = 3/3), 50 pg MGR5 (n = 0 /41), 150 pg MGR5 (n = 0/55), 300 pg MGR5 (n = 0/24), 500 pg MGR5 (n = 0/18), 50 pg MGR6 (n = 0/39), 150 pg MGR6 (n = 2/49), 300 pg MGR6 (n = 0/27), 500 pg MGR6 (n = 0/31). All embryos are oriented with anterior towards the left.|
|Fig. 4. HSPG binding is required for neural crest but not neural plate border specification by Grem1. (A–I) Stage 14 embryos assayed for neural plate border markers. (A, D, G) Control embryos, (B, E, H) unilateral MGR5 expression, (C, F, I) unilateral MGR6 expression. (A–C) msx1 expression at the neural plate border is unexpectedly expanded and upregulated by both MGR5 (n = 6/8) and MGR6 (n = 7/9). (D–F) pax3 is similarly affected by MGR5 (n = 7/9) and MGR6 (n = 5/9). (G–I) The same goes for zic1, with MGR5 (n = 7/7) and MGR6 (n 7/7). (J–R) Stage 18 embryos assayed for twist1 expression. (J) The cranial neural crest (nc) is marked by twist1. (K) As expected, unilateral overexpression of grem1 RNA, expands the neural crest on the injected side (arrowhead). (L–M) The same dose of MGR5 (L, n = 26/26) or MGR6 (M, n = 29/29) is unable to elicit an expansion. (N–R) Neural crest induction rescue experiment. (O) Bilateral Grem1 knockdown severely impairs neural crest formation (asterisks). (P) As before, unilateral supplementation with 100 pg wild-type mouse grem1 restores neural crest development (arrowhead). (Q–R) Accordingly, the same dose of MGR5 (Q, n = 17/17) or MGR6 (R, n = 15/16) is unable to rescue the neural crest. (S) Schematic summarising a working model for the function of Grem1 in neural crest induction. (i) BMP (blue) antagonism from the organiser initially induces neural identity in the ectoderm. (ii) Wnt (pink) from the underlying mesoderm and non-neural ectoderm, and Grem1 (yellow), perhaps also from the underlying mesoderm, signals to the future neural plate border, independently of HSPGs. (iii) HSPGs in the neural plate border stabilise and/or modulate the activity of Grem1 protein in this tissue. (iv) This enables the eventual induction of the neural crest. Embryos are shown in frontal view with dorsal-posterior towards the top.|