January 1, 2010;
Distinct roles for Robo2 in the regulation of axon and dendrite growth by retinal ganglion cells.
Guidance factors act on the tip of a growing axon
to direct it to its target. What role these molecules play, however, in the control of the dendrites that extend from that axon''s cell body is poorly understood. Slits, through their Robo receptors, guide many types of axons, including those of retinal ganglion
cells (RGCs). Here we assess and contrast the role of Slit/Robo signalling in the growth and guidance of the axon
and dendrites extended by RGCs in Xenopus laevis. As Xenopus RGCs extend dendrites, they express robo2
, while slit1
are expressed in RGCs and in the adjacent inner nuclear layer. Interestingly, our functional data with antisense knockdown and dominant negative forms of Robo2
(dnRobo2) and Robo3
(dnRobo3) indicate that Slit/Robo signalling has no role in RGC dendrite
guidance, and instead is necessary to stimulate dendrite
branching, primarily via Robo2
. Our in vitro culture data argue that Slits are the ligands involved. In contrast, both dnRobo2 and dnRobo3 inhibited the extension of axons and caused the misrouting of some axons. Based on these data, we propose that Robo signalling can have distinct functions in the axon
and dendrites of the same cell, and that the specific combinations of Robo receptors could underlie these differences. Slit acts via Robo2
in dendrites as a branching/growth factor but not in guidance, while Robo2
function in concert in axons to mediate axonal interactions and respond to Slits as guidance factors. These data underscore the likelihood that a limited number of extrinsic factors regulate the distinct morphologies of axons and dendrites.
[+] show captions
Fig. 1. Slit1 and slit2 are expressed in the developing Xenopus retina. (Aâ F) Transverse vibratome sections through the retinas of Xenopus embryos processed for wholemount in situ hybridization against slit1 (A, C, E) or slit2 (B, D, F). slit1 mRNA is expressed in the floor plate of the neural tube at stages 30â 35/36 (A, C). Weak mRNA expression is evident in the ganglion cell layer (GCL, as marked by **, A and E) by stage 30, and maintained at stages 33/34 (E) and 35/36 (C). Slit2 mRNA is expressed in the inner nuclear layer (INL) of the developing Xenopus retina (B, D, F).Slit2 is expressed in the floor plate and roof plate at stages 35/36 and 37/38 (B, D). At stage 35/36, expression in the retina is observed in the proliferative ciliary marginal zone (CMZ; arrowheads), and in central retinal cells of the INL (B). CMZ expression is lost by stage 37/38, while INL expression spreads across the whole layer (D). A higher magnification image of a stage 37/38 retina (F) highlights the location of the slit2 expression in the inner portion of the INL (INL is delineated by the black bars) and not in the GCL (**). Scale bars are 50â î¼m in A for Aâ D and in E for Eâ F. br, brain; D, dorsal; fp, floor plate; INL, inner nuclear layer; L, lens; nr, neural retina; rp, roof plate; V, ventral.
Fig. 2. robo genes are differentially expressed in the developing Xenopus retina. (Aâ G, Iâ J) Xenopus embryos processed for wholemount in situ hybridization for robo1, robo2, or robo3, shown as whole embryos (Aâ C), or in transverse vibratome sections (Dâ G, Iâ J). Robo1 expression is evident in the brain, otic vesicle and branchial arches (A). In stage 33/34 and 37/38 sections, robo1 label is present in the brain, but not in the retina (Dâ E). Robo2 is strongly expressed in migrating neural crest cells (B). Robo2 expression in the brain and eye is already strong at stage 33/34 (F), and becomes even more robust by stage 37/38 (G). Note the strong robo2 label in the GCL (**), with some weaker expression in the other retinal layers (Fâ G). (H) A Robo2 antibody was used to label a transverse cryostat section of a stage 37/38 retina. Robo2 protein is most readily detected in the RGCL (**) and the optic nerve head (arrowhead). The inset in H is an enlargement of the boxed region and shows a few RGCs and their dendrites (arrows) that are expressing Robo2. In wholemounts, robo3 mRNA is detectable in the brain, lens, branchial arches, and migrating cranial neural crest cells (C). A transverse section through a stage 33/34 embryo shows that robo3 mRNA is strongly expressed in the lens, but also weakly in the GCL (**, I). Expression in both locations is lost at stage 37/38 (J). Scale bar in D is 50â î¼m for Dâ G and Iâ J and 25â î¼m for H. A, anterior; br, brain; ba, branchial arches; D, dorsal; e, eye; L, lens; nc, neural crest cells; nr, neural retina; ot, otic vesicle; P, posterior; V, ventral.
Fig. 4. Expression of dnRobo2-MT in developing RGCs reduces dendrite outgrowth and branching. Examples of transgene-expressing RGCs in sections of stage 40 Xenopus retinas following electroporation at stage 27. Cell morphology was analyzed for RGCs expressing GFP alone (A), GFP plus dnRobo2-MT (B), or GFP plus dnRobo3-MT (C). (D–F) anti-myc immunoreactivity (E) for dnRobo2 accurately reflects the dendritic arbour as revealed by GFP fluorescence (D). Merge is shown in F. (G–L) dnRobo2-expressing cells in the RGC layer (G, J) express differentiated protein markers of RGCs, including the transcription factors islet-1 (H, I) and Brn3b (K, L). (M–O) A dnRobo3-expressing RGC (M) expresses Pax6 protein (N, O). In all panels the cells are oriented so that the IPL would be at the top and the lens the bottom. Arrows point to primary dendrites extended by the RGCs. Note that dnRobo2-MT and dnRobo3-MT were expressed at lower transfection efficiencies than GFP. Scale bar is 5 μm.
Fig. 6. Robo2 and Robo3 have roles in RGC axon extension and guidance. (Aã¢â â B) Brains dissected out of stage 35/36 embryos following wholemount in situ hybridization with antisense riboprobes to slit1 (A) or slit2 (B). (C) For comparison, a stage 35/36 horseradish peroxidase-labelled optic tract (brown) is shown enroute to the tectum. RGC axons cross the midline at the optic chiasm, travel dorsally through the diencephalon, make a caudal turn (*) in the mid-diencephalon, and enter the tectum in the midbrain (approximate tectal border is shown with dotted line). Slit1 and slit2 are expressed near the optic chiasm and the mid-diencephalic turn, and are present dorsal to where the axons enter the tectum. (Dã¢â â I) Retinas of stage 27 Xenopus embryos were electroporated with GFP alone or together with dnRobo2-MT or dnRobo3-MT. Embryos were fixed at stage 40, and brains were removed and processed for anti-GFP or anti-myc immunochemistry. The black cells are melanophores. (Dã¢â â I) Lateral views of axons travelling in the contralateral optic tract of a control brain with GFP-expressing RGC axons (D), or of brains with axons expressing dnRobo2 (Eã¢â â F) or dnRobo3 (Gã¢â â I). In the control, axons come across the optic chiasm, make a caudal turn in the diencephalon (*) and innervate the tectum. Axons expressing dnRobo2 (Eã¢â â F) or dnRobo3 (G) often end in the ventral or mid-diencephalon (arrowheads show the locations of the growth cones). Axon guidance errors were sometimes seen: In (F) one axon (arrows) makes an aberrant turn just anterior to the rostral border of the optic tectum (high power view of the area boxed on the wholemount brain shown in inset), while in (Hã¢â â I), one of the dnRobo3-expressing axons missed the mid-diencephalic turn and travelled dorsally towards the pineal gland. I is a high power view of the boxed area in H, and numbers identify the three axon tips. Scale bar in D is 50ã â ã â¼m for Dã¢â â E and Gã¢â â H, 25ã â ã â¼m for F and 10ã â ã â¼m for I. A, anterior; D, dorsal; di, diencephalon; hb, hindbrain; oc, optic chiasm; pi, pineal gland; P, posterior; tec, tectum; tel, telencephalon; V, ventral. (J) Graph showing defects in axon extension upon inhibition of Robo signalling. The location of the growth cone for each GFP, dnRobo2 or dnRobo3-expressing axon was scored as being in the ventral diencephalon (di), mid-diencephalon, dorsal diencephalon, or at the tectum. The numbers in brackets represent the total numbers of axons analyzed for each condition. Data were pooled from four independent experiments. dnRobo2 and dnRobo3 each caused a significant change in the extension of RGC axons (pã â <ã â 0.001, Chi-square test).