Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 May 17;94(4):790-799.e3.
doi: 10.1016/j.neuron.2017.03.007. Epub 2017 Apr 21.

Netrin1 Produced by Neural Progenitors, Not Floor Plate Cells, Is Required for Axon Guidance in the Spinal Cord

Free PMC article

Netrin1 Produced by Neural Progenitors, Not Floor Plate Cells, Is Required for Axon Guidance in the Spinal Cord

Supraja G Varadarajan et al. Neuron. .
Free PMC article


Netrin1 has been proposed to act from the floor plate (FP) as a long-range diffusible chemoattractant for commissural axons in the embryonic spinal cord. However, netrin1 mRNA and protein are also present in neural progenitors within the ventricular zone (VZ), raising the question of which source of netrin1 promotes ventrally directed axon growth. Here, we use genetic approaches in mice to selectively remove netrin from different regions of the spinal cord. Our analyses show that the FP is not the source of netrin1 directing axons to the ventral midline, while local VZ-supplied netrin1 is required for this step. Furthermore, rather than being present in a gradient, netrin1 protein accumulates on the pial surface adjacent to the path of commissural axon extension. Thus, netrin1 does not act as a long-range secreted chemoattractant for commissural spinal axons but instead promotes ventrally directed axon outgrowth by haptotaxis, i.e., directed growth along an adhesive surface.

Keywords: adhesion; axon guidance; chemotaxis; dcc; floor plate; haptotaxis; netrin1; progenitors; spinal cord; ventricular zone.


Figure 1
Figure 1. FP-derived netrin1 is not required to direct the circumferential trajectory of spinal axons
(A–I, L–AA) Thoracic level transverse sections from E11.5 netrin1+/lacZ; Gli2+/− (control, A, D, G), netrin1+/lacZ; Gli2−/− (Gli2 mutant, B, E, H), netrin1lacZ/lacZ; Gli2−/− (netrin1; Gli2 mutant, C, F,I) Shh::cre; netrin1flox/+ (control, L, N, P–S), Shh::cre; netrin1flox/flox (netrin1ΔFP M, X–AA), netrin1lacZ/lacZ (T–W) mouse spinal cords. (A–C) Netrin1 expression is specifically lost from the FP in Gli2 mutants (B, FP region is shown magnified in A′–C′), and is completely absent from Gli2; netrin1 mutants. (D–I) NF+ axons grow circumferentially in control and Gli2 mutants avoiding the netrin1::β-gal+ VZ (G′, H′). In contrast, NF+ axons extend robustly into the VZ in the Gli2; netrin1 mutants (arrows, I′). (J, K, O) Quantification showed that there are 2–3 fold more NF+ axons extending towards the VZ in the Gli2; netrin1 mutants (48.1±3.1 NF+ axons/section; n=26 sections from 2 embryos) compared to either control (14.2±1.4 NF+ axons/section; n=26 sections from 2 embryos) or Gli2 mutants (16.2±1.0 NF+ axons/section; n=50 sections from 4 embryos). Gli2; netrin1 mutant axons extend aberrantly into the VZ in all zones of the spinal cord (O). (L–M) Netrin1 is specifically lost from the FP in the netrin1ΔFP mice compared to control (FP region magnified in L′ and M′). (N) Cre is only present in FP cells in both control and netrin1ΔFP embryos (FP region magnified in N″). (O) Quantification schematic of four zones along the dorsal-ventral axis of the spinal cord. (P–AA) The NF+, Tag1+ and Robo3+ populations of axons project apparently normally around the VZ in netrin1ΔFP spinal cords (X–AA), very similar to littermate controls (P–S) and distinct from the multiple phenotypes observed in netrin1lacZ/lacZ mutants (T–W) (Laumonnerie et al., 2015; Serafini et al., 1996). (BB, CC) Quantification demonstrated that there is no significant difference (p>0.31) between the number of NF+ axons extending towards the VZ in control (14.2±1.1 NF+ axons/section; n=53 sections from 3 embryos) and netrin1ΔFP (12.8±0.8 NF+ axons/section; n=67 sections from 4 embryos) mice. In contrast, NF+ axons profusely project into the VZ in netrin1lacZ/lacZ mutant embryos (56.4±1.4 NF+ axons/section; n= 68 sections from 6 embryos) at all zones of the spinal cord (O). Data represented as mean ± SEM. Probability of similarity ** p<0.005, *** p< 0.0005, Student’s t-test. Scale bar: A–C: 140 urn; D–I, L–AA: 105pm
Figure 2
Figure 2. Selective depletion of netrin1 from the VZ results in axon guidance defects
(A–U, V–HH) Thoracic level transverse sections of E11.5 Pax3::cre; netrin1flox/+ (control, A–C, F, H–K, R, T), Pax3::cre; netrin1flox/flox (netrin1ΔdVZ, D–E, G, L–O, S, U), Dbx1 ::cre; ROSA26R::gfp (control, V–Z), Dbx1 ::cre; Rbpjflox/flox; ROSA26R::gfp (Notch OFF, AA–EE), Olig2::cre; Rbpjflox/flox; ROSA26R::gfp (Notch OFF, FF–II) mouse spinal cords. (A–G) The Pax3:: cre line drives expression of cre specifically in the dorsal spinal progenitors (C), resulting in the loss of netrin1 from the dorsal spinal cord of netrin1ΔdVZ embryos (D, E) and not in control littermates (A, B). The NF+ axons move laterally in the netrin1ΔdVZ mice to be immediately adjacent to the laminin+ pial surface (B, E). (H–K, R, T) In control littermates, NF+ axons generally avoid the VZ (I) and dorsal-most spinal cord (R), while Robo3+ (J) and Tag1+ (K) commissural axons project in a tightly fasciculated bundle around the VZ and towards the FP (arrows, T). (L–O) In contrast, there are many axon guidance defects in the netrin1ΔdVZ embryos. NF+ axons extend into the dorsal VZ, with some axons reaching the roof plate (magnified panel in S). Robo3+ axons are defasciculated as they extend ventrally (N, arrows, U), and the number of Tag1+ axons reaching the FP appears to be diminished (O). (P, Q) Quantification demonstrated that 2-fold more NF+ axons extend towards the VZ in the netrin1ΔdVZ embryos 29.4±0.9 NF+ axons/section; n=l02 sections from 5 embryos) compared to controls (l6.0±l.0 NF+ axons/section; n=64 sections from 3 embryos). These NF+ axons only grew into the VZ in the dorsal zones (i.e. zones l and 2), where netrin1 was no longer present, while no significant difference was observed in zones 3 and 4 (p>0.ll and p>0.32 respectively). (V–Z) The Dbx1::cre driver line targets GFP reporter gene expression to the p0 domain (box in V shown magnified in panel Z). (AA–EE) The Dbx1::cre driver line is used to deplete Notch signaling from p0 domain, the Sox2+ progenitors in this region (brackets,CC) rapidly differentiate into post-mitotic neurons (Kong et al., 20l5), which do not express netrin1 (bracket, DD). NF+ axons now extend around the ectopic netrin1 boundary (arrows, BB). (FF–HH) Loss of Notch signaling in the Olig2+ pMN domain has no effect on Sox2+ progenitors (bracket, HH), and does not create an ectopic netrin1 boundary or perturb NF+ axon trajectories. (II) There are >2 fold more NF+ axons/μm entering the VZ in the GFP+ p0 region in the Notch OFF spinal cord compared to controls. Of these, 2-fold more project precisely along the p0 GFP boundary. There was no significant difference (p>0.15) in the number of NF+ axons projecting into the VZ outside the GFP+ p0 region in the Notch OFF and control spinal cords. Control: n=47 sections from 4 mice; Notch OFF: n= 66 sections from 6 mice. Data represented as mean ± SEM. Probability of similarity between control and mutant, *** p< 0.0005 Student’s t-test. Scale bar: 105 μm
Figure 3
Figure 3. Spinal progenitors deposit a netrin1 substrate on the pial surface
(A–M) E11.5 thoracic (A–B and L) and lumbar (J) netrin1lacZ/+ and E10.5 lumbar (K) and E11.5 thoracic (C–G and M) netrin1+/+ mouse spinal cords. Note that for netrin1 immunohistochemisty, panel G was processed without antigen retrieval. (A, B) Netrin1 (A) and netrin1::β-gal (B) are both present in FP cells and neural progenitors in the VZ. The domain of netrin1 and netrin1::β-gal expression extends from the ventral midline to a dorsal boundary at the same level as the dorsal root entry zone (DREZ, dotted line). (C) In contrast, high levels of netrin1 protein are observed around the basal pial circumference of the spinal cord starting at the same dorsal boundary observed for netrin1 expression (dotted line), Netrin1 is also present on commissurally projecting axons (chevrons, C, C′). (D–I) Netrin1 protein co-localizes with both the nestin+ progenitor processes (arrows, H) and the laminin+ pial surface (I). Netrin1 is not present at the pial surface in the dorsal-most spinal cord, i.e. above the DREZ, where netrin1::β-gal is not present in the VZ (E, F). See also Movie S1. (J, K) NF+ axon extension is co-incident with the dorsal border of both netrin1:: β-gal expression and netrin1 on the pial surface (dotted line, J′, K′). (L) By Ell.5, NF+ and Tag1+ axons project around a continuous border of netrin1::β-gal+ cells, that spans from the dorsal VZ to the apical FP (dotted lines, L′). Commissural axons are most fasciculated as they project beneath the domain of netrin1:: β-gal at the FP (L′). (L, M) Axon growth also correlates with distribution of netrin1 protein. NF+ axons and Tag1+ Robo3+ commissural axons extend immediately adjacent to the laminin+ netrin1+ pial surface in the dorsal spinal cord (M′). Scale bar: 120um
Figure 4
Figure 4. Dcc mediates the response to VZ-derived netrin1
(A–O) Thoracic level transverse sections of E11.5 netrin1+/+; dcc+/+ (control, A–C, J–L) netrin1lacZ/lacZ (D–F), and Dcc−/− (G–I, M–O) mouse spinal cords. (A–F) Antigen retrieval (see methods) boosts the netrin1 signal in axons (chevrons, B, C) and the pial surface (see also Movie S2). This staining is lost in netrin1lacZ/acZ embryos (E, F). As previously described (Poliak et al., 2015), netrin1 antibodies detect the netrin1::β-gal fusion protein in VZ. (G–I) Netrin1 accumulation in NF+ axons is greatly diminished in Dcc mutant spinal cords, even though pial-netrin1 remains intact (see also Movie S3). (J–L) Control NF+ (K) and Robo3+ (L) axons project precisely around the VZ. (M–O) In contrast, Dcc mutant NF+ and Robo3+ axons exuberantly project dorsally into the VZ at all levels (N, O). Robo3+ axons are also profoundly defasciculated in the motor columns (O). (P) Quantification of the average number of NF+ axons extending into the VZ demonstrates that a comparable number of NF+ axons extend into the VZ in Dcc and netrin1 (Figure 1BB–1CC) mutant embryos. Control: n=44 sections, 3 embryos, and Dcc−/−: n=l05 sections, 6 embryos. (Q) In the canonical model, netrin1 functions as a long-range chemoattractant secreted by cells in the FP. (R, S) In the growth substrate model, netrin1 produced by neural progenitors is transported to the pial surface in their radial processes to form a growth substrate (green line). Axons then extend adjacent to this substrate in a Dcc dependent manner. (T, U) Our conditional analyses support the growth substrate model, by demonstrating the key requirement for VZ-derived netrin1 in guiding spinal axons. Data is represented as mean ± SEM. Probability of similarity, *** p< 0.0005, ** p<0.005, * p<0.05, Student’s t-test. Scale bar: H5μm

Similar articles

See all similar articles

Cited by 49 articles

See all "Cited by" articles

MeSH terms

LinkOut - more resources