Neogenin Recruitment of the WAVE Regulatory Complex ... - Cell Press

Article

Neogenin Recruitment of the WAVE Regulatory Complex to Ependymal and Radial Progenitor Adherens Junctions Prevents Hydrocephalus Graphical Abstract

Authors Conor J. O’Leary, Cathrin C. Nourse, Natalie K. Lee, ..., Kai Sempert, Stacey J. Cole, Helen M. Cooper

Correspondence [email protected]

In Brief Loss of ependymal adhesion leads to hydrocephalus. O’Leary et al. show that adherens junction assembly relies on Neogenin recruitment of the WRC and Arp2/3 to promote actin polymerization and that its loss results in hydrocephalus. Neogenin therefore stabilizes ependymal junctions to buffer the forces generated by increasing CSF in the neonate.

Highlights d

Loss of Neogenin leads to hydrocephalus due to ependymal adherens junction failure

d

Depletion of Neogenin in the embryonic cortex disrupts radial glial junctions

d

Neogenin anchors the WRC and Arp2/3 to radial glial and ependymal junctions

d

Blocking Neogenin-WRC binding leads to actin depolymerization and junctional loss

O’Leary et al., 2017, Cell Reports 20, 370–383 July 11, 2017 ª 2017 The Author(s). http://dx.doi.org/10.1016/j.celrep.2017.06.051

Cell Reports

Article Neogenin Recruitment of the WAVE Regulatory Complex to Ependymal and Radial Progenitor Adherens Junctions Prevents Hydrocephalus Conor J. O’Leary,1,2 Cathrin C. Nourse,1 Natalie K. Lee,1,3 Amanda White,1 Michael Langford,1 Kai Sempert,1 Stacey J. Cole,1,4 and Helen M. Cooper1,5,* 1Queensland

Brain Institute, The University of Queensland, Brisbane, QLD 4072, Australia address: Queensland Emory Drug Discovery Initiative, The University of Queensland, Brisbane QLD 4072, Australia 3Present address: Department of Structural Biology, St. Jude Children’s Research Hospital, Memphis, TN 38105-3678, USA 4Present address: The University of Queensland Diamantina Institute, The University of Queensland, Brisbane, QLD 4102, Australia 5Lead Contact *Correspondence: [email protected] http://dx.doi.org/10.1016/j.celrep.2017.06.051 2Present

SUMMARY

Denudation of the ependyma due to loss of cell adhesion mediated by cadherin-based adherens junctions is a common feature of perinatal hydrocephalus. Junctional stability depends on the interaction between cadherins and the actin cytoskeleton. However, the molecular mechanism responsible for recruiting the actin nucleation machinery to the ependymal junction is unknown. Here, we reveal that loss of the netrin/ RGM receptor, Neogenin, leads to severe hydrocephalus. We show that Neogenin plays a critical role in actin nucleation in the ependyma by anchoring the WAVE regulatory complex (WRC) and Arp2/3 to the cadherin complex. Blocking Neogenin binding to the Cyfip1/Abi WRC subunit results in actin depolymerization, junctional collapse, and denudation of the postnatal ventricular zone. In the embryonic cortex, this leads to loss of radial progenitor adhesion, aberrant neuronal migration, and neuronal heterotopias. Therefore, Neogenin-WRC interactions play a fundamental role in ensuring the fidelity of the embryonic ventricular zone and maturing ependyma. INTRODUCTION Perinatal hydrocephalus, a prevalent neurodevelopmental condition, is characterized by ventromegaly due to an increase in cerebrospinal fluid (CSF) and is associated with severe intellectual impairment and motor dysfunction (Guerra et al., 2015; Kousi and Katsanis, 2016). Non-communicating hydrocephalus is caused by obstruction of the aqueducts, leading to increased CSF pressure, whereas communicating hydrocephalus results from excessive CSF secretion or failure to resorb the CSF without obstruction. Ependymal cells form a cohesive multi-ciliated epithelium that acts as a bidirectional barrier between the ventricle and the brain parenchyma and transports CSF components and interstitial fluids between these domains (Del Bigio, 2010; Rodrı´guez et al., 2012). Tight cell-to-cell adhesion be-

tween ependymal cells is essential for ventricular integrity and is mediated by cadherin-based adherens junctions situated at the subapical membrane (Guerra et al., 2015; Kousi and Katsanis, 2016). Ependymal cell denudation, as a consequence of junctional failure, is emerging as a major cause of hydrocephalus (Guerra et al., 2015; Rodrı´guez et al., 2012). Ependymal cells are derived from radial glial cells (RGCs) within the embryonic ventricular zone (VZ) (Del Bigio, 2010; Spassky et al., 2005). RGCs generate pyramidal neurons, which then migrate along RGC processes (radial migration) to establish the cortical plate (Wilsch-Braüninger et al., 2016). RGCs are apicobasally polarized, with adherens junctions situated at the subapical membrane within the RGC endfeet at the ventricular surface. Adherens junctions are essential for the establishment of RGC polarity and adhesion and are therefore major determinants of RGC function. Increased junctional stability leads to progenitor expansion and reduced neurogenesis, whereas loss of junctions induces premature neurogenesis (Bizzotto and Francis, 2015; Wilsch-Braüninger et al., 2016). Junctional instability leads to loss of adhesion, the collapse of RGC structure, disruption of the ventricular surface and radial migration. It is now clear that junctional failure underpins the etiology of neuronal migration defects in humans (Bizzotto and Francis, 2015; Ferland et al., 2009; Guerra et al., 2015; Lian and Sheen, 2015). Human neuronal accumulations (heterotopias) also accompany denudation of the ependyma (Guerra et al., 2015; Rodrı´guez et al., 2012). Therefore, cortical malformations and some forms of hydrocephalus have a common origin—the failure of adherens junctions. Despite the importance of adhesion in the maintenance of ventricular integrity, the molecular mechanisms governing RGC and ependymal cell adherens junction assembly and stabilization are poorly understood. In the embryonic and postnatal brain, adherens junctions must resist the mechanical stresses exerted as a consequence of dynamic cellular behaviors (e.g., cell division and interkinetic nuclear migration) and increased CSF volume. The maintenance of VZ integrity in the embryo and postnatal animal therefore relies on the adhesive strength generated by adherens junctions. In all epithelia, junctional stability is dependent on the interaction between cadherins and the closely apposed actin ring running parallel to the junction (Priya and Yap, 2015; Verma et al., 2012).

370 Cell Reports 20, 370–383, July 11, 2017 ª 2017 The Author(s). This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Failure to activate actin nucleation via the Arp2/3 enzyme complex at the cadherin adhesion complex results in junctional collapse and loss of epithelial cohesion (Bernadskaya et al., 2011; Kovacs et al., 2011). Arp2/3-mediated actin nucleation precisely at the junction is critically dependent on its association with the WAVE regulatory complex (WRC), comprising five subunits organized into the Cyfip/Nap and WAVE/Abi/HSPC300 subcomplexes (Chen et al., 2010, 2014). The maintenance of RGC junctional assembly in the embryonic cortex is also known to be reliant on Arp2/3 and the WRC (Wang et al., 2016; Yoon et al., 2014). Despite the importance of the actin ring for junctional stability, the mechanism by which the WRC and Arp2/3 are recruited to the cadherin complex in RGCs and ependymal cells is unknown. Neogenin is an axon guidance receptor for both Netrin-1 and the repulsive guidance molecules (RGMs) and also regulates neurogenesis, neuronal migration, and neural regeneration (Kam et al., 2016; O’Leary et al., 2013, 2015; Rajagopalan et al., 2004; Siebold et al., 2017; van Erp et al., 2015). In the early neural tube, Neogenin is required for the establishment of neuroepithelial adhesion and apicobasal polarity (Kee et al., 2008, 2013; Mawdsley et al., 2004). We have recently demonstrated in an in vitro epithelial cell model that Neogenin is responsible for the tight spatial control of WRC-Arp2/3 recruitment to adherens junctions and that junctional tension is governed by Neogenin-dependent assembly of the actin ring (Lee et al., 2016). Here, we identify Neogenin as a pivotal component of adherens junctions in the RGCs of the embryonic cortex and differentiating ependymal cells in the postnatal VZ, where it promotes actin nucleation by recruiting the WRC and Arp2/3 to the cadherin adhesion complex. The critical role of Neogenin in the maintenance of ventricular integrity in the embryo and neonate is demonstrated by the presence of cortical heterotopias and hydrocephalus when Neogenin-WRC interactions are prevented. RESULTS Perinatal Onset of Hydrocephalus in Neogenin Hypomorphic Mice In Neogenin loss-of-function (Neogt/gt) mice created using a gene trap approach, Neogenin protein is reduced to 19% of wild-type levels and the majority of animals (80%) die prior to birth from unknown causes (Leighton et al., 2001; O’Leary et al., 2015). Within the cohort of surviving Neogt/gt juveniles, we noted that 20% exhibited a dome-shaped cranium indicative of hydrocephalus (Figure S1A). Coronal sections through matched regions of Neogt/gt and wild-type (Neo+/+) brains revealed extensive dilation of the Neogt/gt telencephalon at postnatal day 28 (P28) (Figure 1Ai) accompanied by a severe reduction in the width of the neo- and piriform cortices. Close inspection of the ependyma revealed extensive denudation of ependymal cells in the lateral ventricle (Figure 1Ai, arrow). Abnormal expansion of the third ventricle was also observed at this age (Figure 1Aii, arrow). All surviving Neogt/gt animals exhibited some degree of hydrocephalus, ranging in severity from subtle compression of the cortex and adjacent regions to substantial dilation throughout the

ventricular system as seen in Figure 1 (20% of survivors). Hydrocephalus was never observed in heterozygous animals or their wild-type littermates. Expansion of the lateral ventricles and ependymal denudation was already evident at P12 (Figure 1B), but examination of the cerebral and sylvian aqueducts failed to reveal any obvious obstruction (Figure 1C; data not shown). Diffusion of Fast Green dye into the Neogt/gt fourth ventricle after injection into the lateral ventricle (Figure S1B) confirmed that the aqueducts were not blocked at P21. Although specified embryonically in the mouse, ependymal cells in the lateral ventricles do not transition into mature multi-ciliated cells until after birth (Spassky et al., 2005; Tramontin et al., 2003). Substantial dilation was observed throughout the Neogt/gt ventricular system as early as P3, when denudation at the ventricular surface was already evident in the lateral and fourth ventricles (Figure 1D, arrows), indicating that denudation was initiated concomitant with ependymal cell differentiation. Loss of Adherens Junctions in RGCs and Differentiating Ependymal Cells in Neogt/gt Mice We next examined the integrity of the P28 ventricular surface using an antibody to the ependymal cell-specific marker S100b. In wild-type Neo+/+ animals, a continuous layer of S100b-positive (+) ependymal cells was observed surrounding the lateral ventricle (Figure 2A), and N-cadherin immunolabeling revealed that wild-type ependymal cells were polarized and formed an uninterrupted epithelium. In contrast, there was a marked reduction of S100b and N-cadherin in the Neogt/gt ependymal layer (Figure 2A), indicating severe depletion of ependymal cells. RGCs give rise to both astroglia and ependymal cells (Spassky et al., 2005), suggesting that the depletion of ependymal cells in the Neogt/gt VZ may have resulted from the increased production of astrocytes at the expense of ependymal cells. To test this, dams were injected with bromodeoxyuridine (BrdU) at embryonic day 14.5 (E14.5) and the number of BrdU+/GFAP+ cells lining the lateral wall was quantified at P3 (Figures S1C–S1E). We found no significant difference between genotypes in the number of GFAP+ cells specified at E14.5 (BrdU+/GFAP+) or the total number of GFAP+ cells. At P0–P3, the VZ comprises differentiating GLAST+ RGCs, which already express the ependymal cell marker FoxJ1 (differentiating ependymal cells; Jacquet et al., 2011; Tramontin et al., 2003). Depletion of FoxJ1+ cells was observed in the Neogt/gt VZ compared to that of wild-type littermates (Figure 2B, arrow), and quantitative analysis confirmed a significant decrease in the number of FoxJ1+ cells within the Neogt/gt VZ (Figure 2C). Furthermore, examination of the apical polarity marker, Par3, and the junctional marker, b-catenin, demonstrated that the remaining FoxJ1+ cells were not apicobasally polarized and that adherens junctions were absent (Figures 2B and 2D). Taken together, the above data suggest that Neogenin is required for adherens junction assembly in FoxJ1+ cells and that its loss triggers their denudation from the neonatal VZ. This conclusion was supported by examination of Neogenin expression, which was absent from the ependymal layer within the P28 lateral ventricles (Figure 2E) but localized with GLAST and the adherens junction component N-cadherin at P3 (Figure 2F), indicating that it was

Cell Reports 20, 370–383, July 11, 2017 371

Figure 1. Depletion of Neogenin Results in Ependymal Denudation and Severe Hydrocephalus (A) Cresyl violet staining of P28 Neogt/gt mice revealed expansion of the lateral ventricles, thinning of the cortex (Ai), and extensive denudation of the ependymal layer (Ai, lower panel, arrow). The third ventricle was also expanded (Aii, arrow). (B) The P12 Neogt/gt forebrain exhibited expanded lateral ventricles with ependymal denudation (right panel, arrow). (C) The cerebral aqueduct was not obstructed at P12. (D) At P3, expansion of the lateral and fourth ventricles was evident in Neogt/gt neonates, and denudation of the VZ was observed (lower panels, arrows). (A and D) coronal view; (B and C) sagittal view. Aq, Sylvian aqueduct; cAq, cerebral aqueduct; Cb, cerebellum; Ctx, cortex; Hc, hippocampus; Lv, lateral ventricle; 3v, third ventricle; 4v, fourth ventricle. Composite images are shown in (Ai, top), (Aii and B, left), (C, left), and (D, top). The scale bars represent 1 mm (enlarged panels 50 mm). See also Figure S1.

restricted to the junctions of differentiating RGCs. Furthermore, in P3 Neo+/+ animals, Cyfip1, WAVE2, and Arp3 were concentrated at the ventricular surface (Figures 2G and S2, arrows) but were markedly depleted from this region in Neogt/gt animals, suggesting that Neogenin is necessary for the localization of the WRC and Arp2/3 to the junctions of FoxJ1+ cells.

372 Cell Reports 20, 370–383, July 11, 2017

Neogenin-WRC Interactions Are Required for RGC Adhesion and Radial Migration Ependymal cells are specified in a subpopulation of RGCs in the VZ between E14 and E16 (Spassky et al., 2005). In the E14.5 VZ, Neogenin colocalized with the RGC marker, nestin, and the junctional markers, N-cadherin and b-catenin, within the RGC endfeet (Figures 3A, 3B, and S3A), demonstrating that it was tightly associated with RGC adherens junctions. Neogenin also colocalized with Arp3 and the WRC components, WAVE2 and Cyfip1, at RGC junctions (Figures 3C and 3D). To assess whether Neogenin directly interacted with the WRC, we employed the antibody-based Duolink proximity assay, which detects protein-protein interactions occurring within 40 nm (Lee et al., 2016). Abundant proximity signals were detected for the NeogeninWAVE2 antibody pair at RGC junctions, whereas few proximity signals were generated by the control antibody pair (Figure 3E), indicating that Neogenin and the WRC were closely associated at RGC junctions. Although the WRC is known to be important for RGC junctional assembly (Yoon et al., 2014), the mechanism by which it is recruited to the cadherin complex has yet to be identified. The WRC-interacting receptor sequence (WIRS) found in WRC binding partners, including Neogenin, binds to a highly conserved binding pocket within the WRC composed of the Cyfip1 and Abi subunits (Chen et al., 2014; Lee et al., 2016). To determine whether Neogenin recruits the WRC to the RGC cadherin

(legend on next page)


adhesion complex via its WIRS motif, we generated cDNAs encoding the wild-type Neogenin intracellular domain (NeoICD) or a mutated form in which the key residues of the WIRS motif were changed to alanine (NeoICDDWIRS; Figure S3B). A myristoylation sequence was added at the N terminus to ensure insertion into the plasma membrane. Transfection into HEK293T cells demonstrated that NeoICD and NeoICDDWIRS were expressed at equivalent levels (Figures S3D and S3E). Wild-type or mutant Neogenin ICD cDNAs (containing GFP) or GFP alone were electroporated into the E12.5 mouse cortex. At E14.5, expression of NeoICDDWIRS or GFP alone in the Pax6+ RGC population did not perturb the architecture of the VZ (Figures 3F and S3C). RGCs remained tightly packed and attached to the ventricular surface, and the Pax6+ nuclei were elongated, indicating that apicobasal polarity was maintained. However, wildtype NeoICD severely disrupted RGCs within the VZ where Pax6+ cells had adopted a rounded morphology and were loosely associated (Figures 3F and S3C). In addition, a large population of Pax6+/GFP+ cells was ectopically distributed in the subventricular and intermediate zones. In contrast, RGCs were never detected above the VZ after electroporation of NeoICDDWIRS or GFP alone. Quantification of Pax6+/GFP+ cells confirmed a significant reduction in the proportion of RGCs in the NeoICD VZ and a concomitant increase in the more superficial layers (Figure 3G). This phenotype was observed in 100% of embryos electroporated with the NeoICD construct. In summary, expression of NeoICD, but not NeoICDDWIRS, perturbed RGC adhesion and apicobasal polarity, leading to the ectopic localization of RGCs. Examination of neuronal distribution in the E14.5 cortex after electroporation of NeoICD revealed that a significant population of bIII-tubulin+ neurons had failed to migrate out of the VZ and instead accumulated at the ventricular surface (Figure 4A, arrow). However, in cortices expressing GFP alone or NeoICDD WIRS, neurons migrated normally through the VZ and few neurons were seen abutting the ventricular surface. Quantification of bIII-tubulin+/GFP+ cells across the cortex confirmed a significant increase in the number of neurons within the VZ in the presence of NeoICD (Figure 4C). In some instances, NeoICD also induced a periventricular heterotopia phenotype in which aggregates of neurons and progenitors protruded into the ventricular space (Figure 4B, arrow). Periventricular heterotopias were never observed after electroporation of NeoICDDWIRS or GFP alone. Thus, expression of NeoICD disrupted RGC adhesion and morphology, leading to a failure in neuronal migration. RGC Junction Stability Requires Neogenin Recruitment of the WRC The detachment of RGCs from the ventricular surface in the presence of NeoICD suggested that NeoICD induced the disas-

sembly of adherens junctions. In GFP- and NeoICDDWIRS-expressing RGCs, intact adherens junctions were identified as concentrated patches of N-cadherin, b-catenin, and ZO-1 within the apical endfeet (Figures 5A and S6A). Conversely, these junctional proteins were no longer localized to the endfeet in NeoICDexpressing cells, indicating junctional failure. The localization of both NeoICD and NeoICDDWIRS to the RGC endfeet was confirmed by Myc immunolabeling (Figure S3C). To further test the hypothesis that Neogenin plays a central role in maintaining the integrity of RGC junctions, we electroporated full-length wild-type Neogenin (FLNeo) or mutated Neogenin (FLNeoDWIRS) into the E12.5 cortex (Figures 5B and S6C). As seen for NeoICDDWIRS, the expression of FLNeoDWIRS did not perturb N-cadherin+ junctions or the VZ architecture and Pax6+/GFP+ RGC morphology was unaffected (Figure 5B). In contrast, FLNeo induced depolarization of Pax6+/GFP+ RGCs and detachment from the ventricular surface due to junctional disruption in RGC endfeet, as indicated by loss of N-cadherin (Figure 5B, arrow). To further confirm that Neogenin maintains the integrity of RGC junctions, we introduced Neogenin-specific short hairpin RNAs (shRNAs) into the E12.5 mouse cortex by electroporation and analyzed RGC junctions at E14.5 (Figure S4). In embryos expressing Neo shRNA, Pax6+/GFP+ RGCs were observed in the subventricular and intermediate zones as well as the cortical plate (Figures S4A and S4B). Pax6+/GFP+ RGCs within the VZ were no longer polarized, and adherens junctions had failed to form (Figures S4A and S4C). Subcortical and periventricular heterotopias were also observed after depletion of Neogenin (Figure S5). Depletion of Neogenin, therefore, exactly replicated the phenotype generated by the expression of NeoICD or FLNeo. Control shRNAs had no effect on Pax6+/GFP+ RGC distribution, junctional stability, or neuronal migration. Disassembly of RGC junctions in the presence of NeoICD or FLNeo indicated that they may prevent the recruitment of the WRC and Arp2/3 to the cadherin complex within the RGC endfeet by sequestering the WRC away from the cadherin adhesion complex, as suggested by the increased levels of NeoICD and FLNeo throughout the basolateral membrane (Figures S3C and S6C). Conversely, the WIRS mutants were unable to associate with Cyfip1/Abi and therefore were unable to sequester the WRC. This was supported by the observation that, in the presence of NeoICDDWIRS or GFP alone, the Cyfip1 and WAVE2 WRC subunits and Arp3 remained tightly associated with the junctional complex within the RGC endfeet (Figures 5C and 5D), whereas NeoICD promoted the dissociation of these proteins from the subapical membrane. To provide further evidence for this interpretation, we performed a Duolink proximity assay using antibodies to WAVE2 and the Myc epitope on the NeoICD C terminus and quantified the number of proximity signals in the

Figure 2. Adherens Junctions Are Lost in the Postnatal Neogt/gt VZ (A) At P28, S100b+ ependymal cells were depleted in the Neogt/gt lateral wall. N-cadherin was absent from the ependymal layer. (B) At P3, FoxJ1+ cells were reduced in the Neogt/gt VZ and Par3 was lost from the subapical membrane (arrow). (C) Quantification of FoxJ1+ cells in the lateral ventricle (Lv) (unpaired Mann-Whitney U test; p = 0.028; mean ± SEM; n = 4). (D) At P3, b-catenin was no longer localized to the subapical membrane of Neogt/gt FoxJ1+ cells (arrows), indicating loss of junctions. (E) At P28, Neogenin was seen in the subventricular zone (SVZ) and was absent from the S100b+ ependymal layer (EL) (arrows). (F) At P3, Neogenin was expressed by GLAST+ cells in the VZ and localized to N-cadherin+ junctions (arrows). (G) Arp3 and Cyfip1 were depleted from the subapical membrane in P3 Neogt/gt mice. Composite images: (A, left). The scale bars represent (A) 1 mm (enlarged panels 5 mm), (B and F) 5 mm, and (D, E, and G) 10 mm. See also Figure S2.




RGC endfeet (Figure 5E). This analysis clearly demonstrated that there was a strong interaction between NeoICD and WAVE2, whereas minimal association was observed between NeoICDD WIRS and WAVE2, thereby confirming that NeoICD, but not the mutant, was able to interact with the WRC. This led us to conclude that the interaction between the Neogenin WIRS motif and Cyfip1/Abi is required to anchor the WRC to the RGC junction. Neogenin Is Required for Arp2/3-Mediated Actin Nucleation at the RGC Junction Arp2/3 exists in an inactive complex and is activated upon interaction with the VCA subdomain of WAVE2 (Chen et al., 2010; Verma et al., 2012). As seen for WAVE2 and Cyfip1 (Figures 5C and 5D), expression of NeoICD, but not NeoICDDWIRS or GFP alone, in the E14.5 VZ resulted in the loss of Arp3 from the RGC subapical membrane (Figure 5D). Furthermore, depletion of NeoICD resulted in the complete loss of F-actin in the endfeet and a concomitant increase in G-actin (Figure 6A). In contrast, the actin ring remained intact in the presence of NeoICDDWIRS or GFP alone, and the level of unpolymerized G-actin observed in the endfeet was minimal (Figure 6A). Quantification of the F-actin fluorescence intensity relative to that of G-actin (F-actin:G-actin ratio; Schmid et al., 2014) confirmed that incorporation of G-actin into the actin ring was significantly reduced in the presence of NeoICD (Figure 6B). Therefore, Arp2/3-mediated actin nucleation and hence the reinforcement of the actin ring is governed in RGCs by the ability of Neogenin to spatially restrict the WRC and Arp2/3 to the adherens junction, an essential prerequisite for the establishment of junctional tension. Neogenin Maintains Junctional Stability in Differentiating Ependymal Cells in the Postnatal VZ To determine whether Neogenin is required for the recruitment of the WRC and Arp2/3 to the cadherin complex in differentiating ependymal cells in the postnatal VZ, we performed ex vivo electroporation to introduce NeoICD, NeoICDDWIRS, or GFP into the lateral wall of the P0 cortex. Slice cultures were generated from electroporated brains and analyzed 2 days later (equivalent to P2). In the presence of NeoICDDWIRS or GFP alone, high levels of GFP expression were observed in a large proportion of RGCs, identified by their highly polarized morphology and long basal processes extending away from the cell body at the ventricular surface (Figure 7A, arrowheads). In striking contrast, in the presence of NeoICD, very few GFP+ RGCs were observed in the lateral ventricle and the remaining GFP+ cell bodies failed to

polarize (Figure 7A). GFP+ cells were also observed within the ventricular space (Figure 7A, arrows), demonstrating that RGC adhesion was disrupted, leading to detachment from the VZ and subsequent extrusion into the ventricular space. Examination of N-cadherin localization in S100b+ ependymal cells revealed that NeoICD severely affected junctional assembly, whereas the junctions remained unperturbed in the presence of NeoICDDWIRS (Figure 7B, arrow). Moreover, NeoICD, but not NeoICDDWIRS, induced a severe depletion of F-actin in FoxJ1+ cells, as indicated by a reduction in phalloidin staining at the ventricular surface (Figure 7C, arrow), leading to a loss of adhesion due to destabilization of the actin ring. These observations provide a mechanistic explanation for the hydrocephalic phenotype seen in Neogt/gt mice, whereby depletion of Neogenin in the postnatal VZ prevents the recruitment of the WRC and Arp2/3 to the junction of differentiating ependymal cells. As a result, the failure to initiate actin nucleation destabilizes the actin ring, leading to the disassembly of adherens junctions and the inability of the maturing ependyma to resist extant stresses. DISCUSSION Denudation of the human ependyma, a common feature of hydrocephalus, has been attributed to the failure to establish a cohesive epithelium due to the disruption of adherens junctions in ependymal cells and RGCs (Guerra et al., 2015; Rodrı´guez et al., 2012). In the embryonic VZ, the inability to assemble adherens junctions leads to neuronal heterotopias, which are often coincident with denudation of the ependyma (Guerra et al., 2015; Rodrı´guez et al., 2012). Thus, loss of adherens junctions is a common feature of hydrocephalus and some cortical malformations. In this study, we demonstrate that the netrin/RGM receptor, Neogenin, plays a pivotal role in the maintenance of adherens junction stability in the VZ and that its loss results in severe hydrocephalus and cortical heterotopias. Adherens junction stability is dependent on the interaction between cadherins and the actin cytoskeleton (Priya and Yap, 2015; Verma et al., 2012). However, the critical molecular mechanism responsible for the recruitment of the actin nucleation machinery to the cadherin adhesion complex in RGCs and ependymal cells has not been investigated. Here, we identify an essential role for Neogenin in the embryonic and postnatal VZ, where it promotes actin nucleation at the junctions of RGCs and differentiating ependymal cells by recruiting the WRC and Arp2/3 to the cadherin adhesion complex. We further show that blocking the interaction between Neogenin and the WRC subunits Cyfip1 and Abi

Figure 3. NeoICD Induces Loss of RGC Polarity and Ectopic Positioning in the E14.5 Cortex (A–D) Neogenin colocalized with the RGC marker Nestin (A) and the adherens junction marker N-cadherin (B) in E14.5 RGC endfeet. The WRC subunits Cyfip1 (C) and WAVE2 (D) as well as the Arp3 subunit (D) colocalized with Neogenin in RGC endfeet. (E) Duolink analysis of Neogenin and WAVE2 (antibody pairs: anti-Neogenin and anti-WAVE2; control anti-Myc and anti-WAVE2) revealed abundant Neo-WAVE2 proximity signals (red puncta) in RGC endfeet at E14.5, indicating a close association between Neogenin and the WRC. (F) In GFP- and NeoICDDWIRS-expressing E14.5 cortices, Pax6+ RGCs were apicobasally polarized and remained in the VZ. NeoICD induced loss of polarity, and RGCs were ectopically positioned above the VZ. (G) Quantification of ectopic Pax6+/GFP+ RGCs above the VZ (two-way ANOVA, ****p < 0.0001, mean ± SEM; Mann-Whitney U test, *p = 0.0286, mean ± SEM; n = 4). ICD, NeoICD; ICDDWIRS, NeoICDDWIRS; IZ, intermediate zone. Composite images are shown in (A, top) and (F). The scale bars represent (A) 50 mm (enlarged panels 10 mm), (B–E) 10 mm, and (F) 50 mm. See also Figures S3 and S4.


Figure 4. At E14.5, Expression of NeoICD Prevents Radial Migration, Leading to Neuronal Heterotopias (A) NeoICD, but not NeoICDDWIRS or GFP alone, resulted in the accumulation of bIII-tubulin+ neurons within the E14.5 VZ (arrow). (B) Aggregates of Nestin+ progenitors protruded into the ventricle in the NeoICD cortices (arrow). (C) Quantification of the distribution of bIII-tubulin+/GFP+ cells across the E14.5 cortex (two-way ANOVA, ****p < 0.0001, mean ± SEM; Mann-Whitney U test, *p = 0.0286, mean ± SEM; n = 4). CP, cortical plate; V, ventricle. Composite images are shown in (A, top). The scale bars represent (A and B) 50 mm. See also Figure S5.

prevents Arp2/3-mediated actin nucleation, leading to the collapse of the actin ring and loss of cell adhesion, which then promotes denudation of the postnatal VZ. In the embryonic cortex, the prevention of Neogenin-WRC interactions destabilizes RGC junctions and disrupts the ventricular surface, leading to aberrant neuronal migration. Depletion of Neogenin in the Neogt/gt VZ provoked denudation prior to P3, which progressed to a substantial loss of mature ependymal cells and severe hydrocephalus throughout the ventricular system. Introduction of NeoICD into the P0 VZ induced loss of adherens junctions, which triggered substantial

de-adhesion of RGCs and differentiating ependymal cells, confirming that inhibition of Neogenin function was directly responsible for the denudation seen in the Neogt/gt postnatal VZ. Moreover, the absence of a cohesive multi-ciliated ependymal epithelium would have a direct impact on CSF laminar flow through the Neogt/gt ventricular system, thereby further exacerbating ventricle dilation. Thus, loss of Neogenin in the postnatal VZ has a detrimental effect on ependymal cell differentiation and maturation, which is essential for the maintenance of ependymal integrity and function. Therefore, the Neogt/gt mice provide a useful model with which to investigate the aberrant




molecular and cellular processes contributing to communicating hydrocephalus. The WRC-interacting receptor sequence (WIRS), found in the intracellular domain of Neogenin, binds to a highly conserved binding pocket within the WRC composed of the Cyfip1 and Abi subunits (Chen et al., 2014; Lee et al., 2016). In the embryonic VZ, blocking Neogenin-WRC interactions resulted in the loss of RGC apicobasal polarity due to NeoICD-induced destabilization of adherens junctions, which subsequently prevented neuronal migration into the cortical plate, leading to neuronal heterotopias. In addition, RGC de-adhesion disrupted the ventricular surface, thereby generating a periventricular heterotopia phenotype in which aggregates of neurons and progenitors protruded into the ventricular space. Despite the severe denudation in the Neogt/gt postnatal VZ, we did not observe any phenotype in the Neogt/gt embryonic VZ. The low level of Neogenin protein (20% of wild-type) remaining in the Neogt/gt brain appears to be sufficient to maintain junctional integrity and thus RGC adhesion within the embryonic ventricular system. However, it is likely that adhesive strength is diminished and, as such, may not be sufficient to resist the forces arising from increased CSF volume and other mechanical stresses associated with expansion of the ventricles in the Neogt/gt neonate. In the embryonic and postnatal brain, adherens junctions must resist the local forces exerted as a consequence of dynamic cellular behaviors and increased CSF volume. The maintenance of VZ integrity relies on adhesive strength, which is dictated by reciprocal interactions between the cadherins and the actin ring, and is buffered by junctional tension generated through actomyosin contractility within the actin ring (Priya and Yap, 2015; Verma et al., 2012). To maintain adhesive strength, the actin ring must undergo continual turnover (Kovacs et al., 2011). The essential role of actin in maintaining the fidelity of the VZ is reinforced by the emergence of hydrocephalus and neuronal heterotopias in mice lacking actin-associated proteins, including myosin IIB, which drives actin contractility (Jain et al., 2014; Ma et al., 2007; Schmid et al., 2014). Recent studies have shown that, as in other epithelial cells, actin nucleation at RGC junctions is initiated by the WRC and the Arp2/3 enzymic complex (Wang et al., 2016; Yoon et al., 2014). However, these studies failed to address the fundamental question of how WRC and Arp2/3 activity is confined to the adherens junction. Here, we identify Neogenin as the missing link that anchors the nucleation machinery to the junctions of RGCs and differentiating ependymal cells. In the absence of junctional Neogenin, actin nucleation fails to initiate due to the dissociation of the WRC and Arp2/3 from the RGC cadherin complex, leading to the loss of F-actin and collapse of the actin ring. Therefore, within RGCs and maturing ependymal cells, Neogenin is a critical junc-

tional component that ensures Arp2/3-mediated actin nucleation occurs at the junction, thereby promoting actin ring stability and adhesion. The emergence of denudation of the ventricular lining only in the postnatal Neogt/gt animal indicates that Neogenindependent assembly of the actin ring directly impacts the flexibility and resilience of the maturing ependyma in response to increasing environmental stresses. Based on our previous in vitro study using the Caco2 epithelial cell line (Lee et al., 2016), we predicted that the preservation of the actin ring in RGCs and ependymal cells would be governed by the ability of Neogenin to spatially restrict the WRC to the cadherin adhesion complex via a direct interaction between its WIRS domain and the Cyfip1/Abi-binding pocket within the WRC. In agreement with this hypothesis, the Duolink proximity assay indicated that Neogenin and the WAVE2 subunit of the WRC were closely associated at RGC junctions (Figure 3E). Confirmation that a direct binding interaction occurred between Neogenin and the WRC was provided by the introduction of NeoICD. Expression of NeoICD in RGCs promoted the dissociation of WAVE2 and Cyfip1 from the RGC subapical membrane, whereas in the presence of the WIRS mutant, the WRC subunits remained tightly associated with the cadherin complex (Figures 5C–5E). Moreover, only NeoICD induced the disintegration of the actin ring, junctional failure, and loss of RGC polarity and adhesion. Together, these observations indicate that NeoICD, but not the WIRS mutant, prevented spatially restricted Arp2/3-mediated actin nucleation at the junction by sequestering the WRC away from the cadherin complex. Proximity assays confirmed that only the NeoICD was able to interact strongly with the WRC (Figure 5E). We therefore conclude that the Neogenin WIRS motif is responsible for the recruitment of the WRC, and subsequently Arp2/3, to the RGC junction via its interaction with the Cyfip1/Abi subunits and that this interaction is essential for maintaining adherens junction stability and VZ integrity in the embryo and neonate. Depletion of Neogenin in the early neural tube also results in the loss of neuroepithelial adhesion and apicobasal polarity (Kee et al., 2008, 2013; Mawdsley et al., 2004), thereby implicating the abrogation of Neogenin-WRC interactions in the etiology of neural tube closure defects. Neogenin was first identified as an axon guidance receptor for both Netrin-1 and the repulsive guidance molecules and also regulates neural migration in the adult and embryonic brain (O’Leary et al., 2015; Rajagopalan et al., 2004; Siebold et al., 2017; van Erp et al., 2015). In addition, it plays a central role in the mobilization of inflammatory cells in response to injury and infection (Mirakaj and Rosenberger, 2017). During migration, attractive or repulsive responses to guidance cues are dictated by remodeling of the actin cytoskeleton, which promotes the extension and retraction

Figure 5. NeoICD Induces RGC Junctional Collapse by Preventing WRC and Arp2/3 Junctional Recruitment (A) At E14.5, expression of NeoICD, but not NeoICDDWIRS or GFP alone, resulted in loss of N-cadherin (arrow) from GFP+ RGC endfeet, indicating junctional failure. (B) After electroporation at E12.5, FLNeo, but not FLNeoDWIRS, induced loss of RGC polarity in the E14.5 cortex and Pax6+/GFP+ RGCs were detached from the ventricular surface. Loss of N-cadherin from GFP+ RGC endfeet in the presence of FLNeo indicated junctional failure (arrow). (C and D) Neo ICD disrupted Cyfip1 (C), WAVE2, and Arp3 (D) recruitment to GFP+ RGC junctions (arrows). (E) Duolink proximity assay: NeoICD, but not NeoICDDWIRS, interacts with WAVE2 (Mann-Whitney U test; *p = 0.05; mean ± SEM; n = 4). Composite images: (B, left). The scale bars represent (A–D) 10 mm. See also Figure S6.


Figure 6. NeoICD Prevents Arp2/3-Mediated Actin Nucleation at the RGC Junction (A) Electroporation of NeoICD, but not NeoICDDWIRS or GFP alone, resulted in the complete loss of F-actin (phalloidin, red, arrow) in RGC endfeet and a concomitant increase in G-actin (DNase I, white, arrow), indicating the absence of actin nucleation and collapse of the actin ring. (B) Quantification of the F-actin:G-actin ratio in RGC endfeet (Kruskal-Wallis test, ***p = 0.0005, mean ± SEM; Mann-Whitney U test, *p = 0.0286; n = 4). Composite images: (A, top of each panel). The scale bars represent (A) 20 mm.

receptors are able to convey spatial information to the actin cytoskeleton. Conclusions This study has revealed an essential role for Neogenin-mediated recruitment of the WRC and Arp2/3 in promoting actin ring stability, adherens junction assembly, and VZ and ependymal integrity in the embryo and neonate, respectively. We also reveal that the disruption of Neogenin function is associated with cortical malformations. We propose that the primary function of Neogenin is to stabilize adherens junctions in order to buffer the forces generated by dynamic cellular behaviors and increasing CSF volume in the expanding embryonic and neonatal VZ. The emergence of hydrocephalus and cortical heterotopias when Neogenin-WRC interactions are prevented emphasizes the fundamental role of this interaction in ensuring the fidelity of the embryonic and postnatal VZ. This study identifies Neogenin and components of the WRC as potentially important candidate genes that may contribute to the etiology of hydrocephalus. EXPERIMENTAL PROCEDURES

of lamellipodia upon receptor activation. Actin extension within lamellipodia is initiated by WRC-mediated Arp2/3 activation (Gomez and Letourneau, 2014; Swaney and Li, 2016). Our finding that Neogenin focuses actin nucleation to restricted regions of the plasma membrane by directly interacting with the WRC provides a plausible molecular mechanism through which guidance


Animals All experiments involving animals were approved by the Anatomical Biosciences Animal Ethics Committee of The University of Queensland and performed in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Neogenin gene trap 129/Ola x C57BL/6 mice were kindly provided by Prof. Marc Tessier-Lavigne (Stanford University; Leighton et al., 2001). Mice (P3–P28) homozygous for the gene trap allele (Neogt/gt) or wild-type littermates (Neo+/+) were transcardially perfused with 0.9% NaCl followed by 4% paraformaldehyde (PFA) and post-fixed for 3 hr at room temperature. Both males and females were used as the phenotype was equivalent in both sexes. C57BL/6 pregnant dams for in utero electroporations were obtained from The University of Queensland breeding facility. For timed matings, males and females were placed together overnight, and the following morning was designated as E0.5.



For dye injections, P21 brains were removed and 2 mL of Fast Green (1% v/v) was injected into the rostral lateral ventricle using a glass-pulled micropipette and picospritzer (Parker Hannifin). Dye diffusion was assessed after 2 min. In Utero and Ex Vivo Electroporations In utero electroporation into the developing somatosensory area of the embryonic cortex was performed as previously described (Sua´rez et al., 2014). Briefly, time-mated pregnant females were deeply anesthetized, the uterine horns exposed, and 1 mL of plasmid DNA (2 mg/mL) injected into a single lateral ventricle at E12.5. Embryonic brains were subsequently electroporated with five 50 ms pulses of 30 V across the uterus using an ECM 830 electroporator (BTX Harvard Apparatus). The embryos were harvested at E14.5 and drop fixed in 4% PFA. For ex vivo electroporations, the brain was dissected from P0 C57BL/6 pups in ice-cold HEM (minimal essential media, HEPES [pH 7.2]; penicillin, 10 U/mL; and streptomycin, 10 mg/mL) and subsequently placed in ice-cold ACSF (artificial cerebrospinal fluid; 2.5 mM KCl, 1 mM NaH2PO4, 10 mM D-glucose, 26.2 mM NaHCO3, and 119 mM NaCl). The lateral ventricle was injected with 1 mL of plasmid DNA (2 mg/mL) and electroporated with five 100 ms pulses of 25 V. Electroporated brains were then embedded in 4% low melting agarose (SigmaAldrich) and sectioned at 400 mm on a vibratome (Leica VT1000S) containing ice-cold ACSF. Individual slices containing the somatosensory area were cultured on 0.4 mm membrane inserts (Millipore) in slice culture medium (25% HBBS, 67% Eagle’s basal medium, 5% fetal bovine serum, 1.5% D-(+)-glucose, penicillin, and streptomycin) for 48 hr in 5% CO2. Slices were then fixed in 4% PFA for 3 hr at room temperature before processing for immunohistochemistry. In general, a minimum of four animals/condition was analyzed in each experiment. Also see the Supplemental Experimental Procedures. Quantification of Cell Populations FoxJ1+ or GFAP+/BrdU+ cells in the ventricular lining of the lateral ventricles were quantified in coronal sections at comparable rostro-caudal levels. Three sections per genotype (four animals/genotype) were analyzed following confocal image acquisition using a 603 oil objective. Quantitative analysis of electroporated cortices was carried out by analyzing GFP+ cells in tissue sections at comparable rostro-caudal levels. Four sections per animal (four animals/condition) were analyzed by capturing 3 3 3 confocal mosaics using a 603 oil objective. The number and location of GFP+/Pax6+ and GFP+/bIIItubulin+ cells were manually recorded using Imaris Bitplane Scientific Software. Also see the Supplemental Experimental Procedures for Duolink proximity assays and calculation of the F-actin:G-actin ratio. Statistical Analysis Statistical analysis was carried out using GraphPad, Prism (version 6; Graphpad Software). A minimum of four animals/condition was analyzed in each experiment. The unpaired Student’s t test was used for comparisons between Neogt/gt and Neo+/+ littermates. Other data were first assessed for normality using the D’Agostino and Pearson normality test. One-way or two-way ANOVA analysis was then performed followed by a two-tailed non-parametric Mann-Whitney U test. Statistical significance was considered to be p < 0.05. All values are presented as mean ± SEM. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and six figures and can be found with this article online at http://dx.doi.org/ 10.1016/j.celrep.2017.06.051.

AUTHOR CONTRIBUTIONS C.J.O. and H.M.C. conceived the project, were responsible for all data analysis, and wrote the manuscript; C.J.O. performed the majority of the experiments and data analysis. C.C.N. performed and analyzed the Neogenin shRNA experiments. A.W. generated all cDNA constructs and performed western blots. S.J.C. and K.S. were responsible for the generation and analysis of RNAi constructs. N.K.L. and M.L. contributed to experimental design, data analysis, and interpretation. ACKNOWLEDGMENTS This work was supported by the National Health and Medical Research Council of Australia (grants 1063080 and 1024201). N.K.L. and K.S. were supported by University of Queensland International Scholarships. M.L. and S.J.C. were supported by Australian Postgraduate Awards. Imaging work was performed in the Queensland Brain Institute’s Advanced Microscopy Facility and generously supported by an ARC LIEF grant (LE130100078). We thank Assoc. Prof. Julian Heng (Harry Perkins Institute of Medical Research) for providing the CA-b-EGFPm5-silencer 3 vector and Prof. Marc Tessier-Lavigne (Stanford University) for providing the Neogenin gene trap mice. The anti-BrdU antibody was deposited in the Developmental Studies Hybridoma Bank by Dr. S.J. Kaufman. We are also grateful to the QBI microscopy facility for expert advice, Ms. Rowan Tweedale for critical reading of the manuscript, and Mr. Nick Valmus for producing the graphical abstract. Received: April 3, 2017 Revised: May 11, 2017 Accepted: June 20, 2017 Published: July 11, 2017 REFERENCES Bernadskaya, Y.Y., Patel, F.B., Hsu, H.-T., and Soto, M.C. (2011). Arp2/3 promotes junction formation and maintenance in the Caenorhabditis elegans intestine by regulating membrane association of apical proteins. Mol. Biol. Cell 22, 2886–2899. Bizzotto, S., and Francis, F. (2015). Morphological and functional aspects of progenitors perturbed in cortical malformations. Front. Cell. Neurosci. 9, 30. Chen, Z., Borek, D., Padrick, S.B., Gomez, T.S., Metlagel, Z., Ismail, A.M., Umetani, J., Billadeau, D.D., Otwinowski, Z., and Rosen, M.K. (2010). Structure and control of the actin regulatory WAVE complex. Nature 468, 533–538. Chen, B., Brinkmann, K., Chen, Z., Pak, C.W., Liao, Y., Shi, S., Henry, L., Grishin, N.V., Bogdan, S., and Rosen, M.K. (2014). The WAVE regulatory complex links diverse receptors to the actin cytoskeleton. Cell 156, 195–207. Del Bigio, M.R. (2010). Ependymal cells: biology and pathology. Acta Neuropathol. 119, 55–73. Ferland, R.J., Batiz, L.F., Neal, J., Lian, G., Bundock, E., Lu, J., Hsiao, Y.-C., Diamond, R., Mei, D., Banham, A.H., et al. (2009). Disruption of neural progenitors along the ventricular and subventricular zones in periventricular heterotopia. Hum. Mol. Genet. 18, 497–516. Gomez, T.M., and Letourneau, P.C. (2014). Actin dynamics in growth cone motility and navigation. J. Neurochem. 129, 221–234. Guerra, M.M., Henzi, R., Ortloff, A., Lichtin, N., Vıó, K., Jimeńez, A.J., Dominguez-Pinos, M.D., Gonza´lez, C., Jara, M.C., Hinostroza, F., et al. (2015).

Figure 7. NeoICD Induces Loss of Junctions in Differentiating Ependymal Cells and Denudation in the P3 VZ (A) After ex vivo electroporation of P0 brains (n = 4), the forebrain was sectioned and slices cultured for 48 hr in vitro (equivalent to P2). After electroporation of GFP alone or NeoICDDWIRS, highly polarized Sox3+/GFP+ RGCs (Sox3; RGC marker) with extended basal processes (arrowheads) were seen in the VZ, whereas, in the presence of NeoICD, GFP+ cells were severely depolarized and had retracted their basal processes. Some Sox3+/GFP+ cells had delaminated from the VZ (arrows). (B and C) Expression of NeoICD, but not NeoICDDWIRS, resulted in loss of N-cadherin (B) and F-actin (phalloidin; C) from differentiating ependymal cells, indicating that NeoICD prevented the formation of adherens junctions and actin nucleation. Composite images: (A and B, left) and (C, left). The scale bars represent (A) 1 mm (enlarged panels 50 mm) and (B and C) 50 mm (enlarged panels 25 mm).


Cell junction pathology of neural stem cells is associated with ventricular zone disruption, hydrocephalus, and abnormal neurogenesis. J. Neuropathol. Exp. Neurol. 74, 653–671.

Priya, R., and Yap, A.S. (2015). Active tension: the role of cadherin adhesion and signaling in generating junctional contractility. Curr. Top. Dev. Biol. 112, 65–102.

Jacquet, B.V., Muthusamy, N., Sommerville, L.J., Xiao, G., Liang, H., Zhang, Y., Holtzman, M.J., and Ghashghaei, H.T. (2011). Specification of a Foxj1dependent lineage in the forebrain is required for embryonic-to-postnatal transition of neurogenesis in the olfactory bulb. J. Neurosci. 31, 9368–9382.

Rajagopalan, S., Deitinghoff, L., Davis, D., Conrad, S., Skutella, T., Chedotal, A., Mueller, B.K., and Strittmatter, S.M. (2004). Neogenin mediates the action of repulsive guidance molecule. Nat. Cell Biol. 6, 756–762.

Jain, N., Lim, L.W., Tan, W.T., George, B., Makeyev, E., and Thanabalu, T. (2014). Conditional N-WASP knockout in mouse brain implicates actin cytoskeleton regulation in hydrocephalus pathology. Exp. Neurol. 254, 29–40.

Rodrı´guez, E.M., Guerra, M.M., Vıó, K., Gonza´lez, C., Ortloff, A., Ba´tiz, L.F., Rodrı´guez, S., Jara, M.C., Munõz, R.I., Ortega, E., et al. (2012). A cell junction pathology of neural stem cells leads to abnormal neurogenesis and hydrocephalus. Biol. Res. 45, 231–242.

Kam, J.W.K., Dumontier, E., Baim, C., Brignall, A.C., Mendes da Silva, D., Cowan, M., Kennedy, T.E., and Cloutier, J.-F. (2016). RGMB and neogenin control cell differentiation in the developing olfactory epithelium. Development 143, 1534–1546. Kee, N., Wilson, N., De Vries, M., Bradford, D., Key, B., and Cooper, H.M. (2008). Neogenin and RGMa control neural tube closure and neuroepithelial morphology by regulating cell polarity. J. Neurosci. 28, 12643–12653. Kee, N., Wilson, N., Key, B., and Cooper, H.M. (2013). Netrin-1 is required for efficient neural tube closure. Dev. Neurobiol. 73, 176–187. Kousi, M., and Katsanis, N. (2016). The genetic basis of hydrocephalus. Annu. Rev. Neurosci. 39, 409–435. Kovacs, E.M., Verma, S., Ali, R.G., Ratheesh, A., Hamilton, N.A., Akhmanova, A., and Yap, A.S. (2011). N-WASP regulates the epithelial junctional actin cytoskeleton through a non-canonical post-nucleation pathway. Nat. Cell Biol. 13, 934–943. Lee, N.K., Fok, K.W., White, A., Wilson, N.H., O’Leary, C.J., Cox, H.L., Michael, M., Yap, A.S., and Cooper, H.M. (2016). Neogenin recruitment of the WAVE regulatory complex maintains adherens junction stability and tension. Nat. Commun. 7, 11082. Leighton, P.A., Mitchell, K.J., Goodrich, L.V., Lu, X., Pinson, K., Scherz, P., Skarnes, W.C., and Tessier-Lavigne, M. (2001). Defining brain wiring patterns and mechanisms through gene trapping in mice. Nature 410, 174–179. Lian, G., and Sheen, V.L. (2015). Cytoskeletal proteins in cortical development and disease: actin associated proteins in periventricular heterotopia. Front. Cell. Neurosci. 9, 99. Ma, X., Bao, J., and Adelstein, R.S. (2007). Loss of cell adhesion causes hydrocephalus in nonmuscle myosin II-B-ablated and mutated mice. Mol. Biol. Cell 18, 2305–2312. Mawdsley, D.J., Cooper, H.M., Hogan, B.M., Cody, S.H., Lieschke, G.J., and Heath, J.K. (2004). The Netrin receptor Neogenin is required for neural tube formation and somitogenesis in zebrafish. Dev. Biol. 269, 302–315. Mirakaj, V., and Rosenberger, P. (2017). Immunomodulatory functions of neuronal guidance proteins. Trends Immunol. 38, 444–456. O’Leary, C., Cole, S.J., Langford, M., Hewage, J., White, A., and Cooper, H.M. (2013). RGMa regulates cortical interneuron migration and differentiation. PLoS ONE 8, e81711. O’Leary, C.J., Bradford, D., Chen, M., White, A., Blackmore, D.G., and Cooper, H.M. (2015). The Netrin/RGM receptor, Neogenin, controls adult neurogenesis by promoting neuroblast migration and cell cycle exit. Stem Cells 33, 503–514.

Schmid, M.-T., Weinandy, F., Wilsch-Braüninger, M., Huttner, W.B., Cappello, S., and Go¨tz, M. (2014). The role of a-E-catenin in cerebral cortex development: radial glia specific effect on neuronal migration. Front. Cell. Neurosci. 8, 215. Siebold, C., Yamashita, T., Monnier, P.P., Mueller, B.K., and Pasterkamp, R.J. (2017). RGMs: structural insights, molecular regulation, and downstream signaling. Trends Cell Biol. 27, 365–378. Spassky, N., Merkle, F.T., Flames, N., Tramontin, A.D., Garcıá-Verdugo, J.M., and Alvarez-Buylla, A. (2005). Adult ependymal cells are postmitotic and are derived from radial glial cells during embryogenesis. J. Neurosci. 25, 10–18. Sua´rez, R., Fenlon, L.R., Marek, R., Avitan, L., Sah, P., Goodhill, G.J., and Richards, L.J. (2014). Balanced interhemispheric cortical activity is required for correct targeting of the corpus callosum. Neuron 82, 1289–1298. Swaney, K.F., and Li, R. (2016). Function and regulation of the Arp2/3 complex during cell migration in diverse environments. Curr. Opin. Cell Biol. 42, 63–72. Tramontin, A.D., Garcıá-Verdugo, J.M., Lim, D.A., and Alvarez-Buylla, A. (2003). Postnatal development of radial glia and the ventricular zone (VZ): a continuum of the neural stem cell compartment. Cereb. Cortex 13, 580–587. van Erp, S., van den Heuvel, D.M.A., Fujita, Y., Robinson, R.A., Hellemons, A.J.C.G.M., Adolfs, Y., Van Battum, E.Y., Blokhuis, A.M., Kuijpers, M., Demmers, J.A.A., et al. (2015). Lrig2 negatively regulates ectodomain shedding of axon guidance receptors by ADAM proteases. Dev. Cell 35, 537–552. Verma, S., Han, S.P., Michael, M., Gomez, G.A., Yang, Z., Teasdale, R.D., Ratheesh, A., Kovacs, E.M., Ali, R.G., and Yap, A.S. (2012). A WAVE2-Arp2/3 actin nucleator apparatus supports junctional tension at the epithelial zonula adherens. Mol. Biol. Cell 23, 4601–4610. Wang, P.-S., Chou, F.-S., Ramachandran, S., Xia, S., Chen, H.-Y., Guo, F., Suraneni, P., Maher, B.J., and Li, R. (2016). Crucial roles of the Arp2/3 complex during mammalian corticogenesis. Development 143, 2741–2752. Wilsch-Braüninger, M., Florio, M., and Huttner, W.B. (2016). Neocortex expansion in development and evolution - from cell biology to single genes. Curr. Opin. Neurobiol. 39, 122–132. Yoon, K.-J., Nguyen, H.N., Ursini, G., Zhang, F., Kim, N.-S., Wen, Z., Makri, G., Nauen, D., Shin, J.H., Park, Y., et al. (2014). Modeling a genetic risk for schizophrenia in iPSCs and mice reveals neural stem cell deficits associated with adherens junctions and polarity. Cell Stem Cell 15, 79–91.
