ARTICLE

Received 9 Dec 2014 | Accepted 19 Jun 2015 | Published 24 Jul 2015 DOI: 10.1038/ncomms8857 OPEN Developmental disruptions underlying brain abnormalities in

Jiami Guo1,*, Holden Higginbotham1,*, Jingjun Li1, Jackie Nichols1, Josua Hirt1, Vladimir Ghukasyan1 & E.S. Anton1

Primary cilia are essential conveyors of signals underlying major cell functions. Cerebral cortical progenitors and neurons have a primary . The significance of cilia function for brain development and function is evident in the plethora of developmental brain disorders associated with human ciliopathies. Nevertheless, the role of primary cilia function in corticogenesis remains largely unknown. Here we delineate the functions of primary cilia in the construction of cerebral cortex and their relevance to ciliopathies, using an shRNA library targeting known to cause brain disorders, but whose roles in brain development are unclear. We used the library to query how ciliopathy genes affect distinct stages of mouse cortical development, in particular neural progenitor development, neuronal migration, neuronal differentiation and early neuronal connectivity. Our results define the developmental functions of ciliopathy genes and delineate disrupted developmental events that are integrally related to the emergence of brain abnormalities in ciliopathies.

1 UNC Neuroscience Center and the Department of and Physiology, University of North Carolina School of Medicine, Chapel Hill, North Carolina 27599, USA. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to E.S.A. (email: [email protected]).

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 1 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857

erebral cortex forms as a result of a coordinated sequence associated with the majority of these genes are yet to be clarified. of events: radial progenitor formation, neurogenesis, We compiled a library of shRNAs targeting 30 of the ciliopathy Cneuronal migration, post-migratory neuronal differentia- genes with defined links to neurodeficits in humans to system- tion and connectivity. An efficient intracellular response to atically evaluate their function in cerebral cortical development. extracellular signalling cues is fundamental for coordinating the Previously well-studied ciliopathy-related genes with known various cellular events that underlie the formation of cerebral developmental functions (for example, ARL13B9,10, INPP5E11,12 cortex. Primary cilia, the -based, slender, antenna-like and IFT88 (ref. 13)) were excluded. This library was assembled projections from cells, are essential integrators and conveyors of using the resources available at the UNC Therapy Center signal transduction. Cortical neuronal progenitors and developing (Supplementary Table 4). Three to six different shRNAs for each neurons have a primary cilium. The significance of cilia gene were combined into two pools to achieve knockdown function for cortical development and function1 is evident in efficiency (KE). Effects were confirmed with duplicated pools. developmental brain disorders such as Joubert, Meckel–Gruber, As control for off-target disruptions, a vector containing orofaciodigital and Bardet–Biedl syndromes (commonly scrambled shRNA was used. Target gene KE of shRNAs referred to as ciliopathies), where disrupted cilia function and was validated using quantitative reverse transcription–PCR the resultant changes in cortical formation underlie cognitive (qRT–PCR) (Supplementary Table 5). Validity of specific deficits and intellectual disabilities2–4. Moreover, neuro- shRNAs was further checked by randomly analysing off-target psychiatric disorders such as spectrum disorders and effects on other ciliopathy genes in the list. No such off-target are also associated with human ciliopathies2–5. knockdown was detected. Further, several shRNAs were found These observations suggest that impaired cilia function can not to induce expected changes in specific . They hinder the development of neural circuitry and activity, leading to were used as additional controls and did not produce any changes significant functional deficits. Therefore, delineating specific in our cortical development assays. functions of ciliopathy-associated genes in distinct stages of cortical development can help decipher the biological basis of brain function abnormalities in ciliopathies. shRNA library screen for ciliopathy gene functions in brain.To Functional genomic analysis of a cilium has led to the understand how ciliopathy-related genes affect early brain identification of a diverse group of cilia and -specific development, different ciliopathy gene-specific shRNA or control proteins1–4,6–8. Among them, in 87 of these genes have shRNA vectors were electroporated into the developing cerebral been associated with human ciliopathies (Supplementary Tables 1 cortex at E14.5. For some of the genes (BBS7, BUBR1 and KIF7), and 2); mutations in 77 of these genes thus far are known to result rescue of shRNA effects was tested using co-electroporation of in neurobehavioural or neurodevelopmental deficits in humans shRNA-resistant human complementary DNAs (cDNAs) (UNC (Supplementary Tables 1–3). A necessary first step in our ability Gene Therapy Center). Electroporated embryos were allowed to to devise effective therapeutic interventions for these cilia-related survive for 2 or 4 days in vivo, and were analysed for progenitor neurodevelopmental diseases is an understanding of their proliferation/organization, neuronal migration, laminar organi- biological basis. Disrupted neural circuitry due to genetic zation and differentiation of post-migratory cortical neurons and mutations in cilia-specific genes in ciliopathy patients could their projections. The effect of shRNAs on distinct aspects of arise as a result of deficits in the generation of appropriate cortical development were assayed as follows. number and types of neurons, disrupted neuronal migration, Cortical progenitors electroporated with shRNAs at E14.5 were placement or connectivity in the developing brain. Pinpointing labelled with radial glial (RG)-specific anti-RC2 or anti-Nestin which of these processes is disrupted as a result of a genetic antibodies or intermediate progenitor-specific anti-Tbr2 anti- can be a complex and time-consuming process. bodies at E16.5. E16.5 embryos were pulse labelled with However, a short hairpin RNA (shRNA) library-based gene 5-bromodeoxyuridine (BrdU) 1 h prior to removal. RG knockdown approach coupled with routinized assays for distinct morphology and the apical and basal endfeet of GFP þ /RC2 þ stages of cortical development allow for rapid and efficient RG were examined for changes in RG polarity. The apical assessment of the brain developmental roles of ciliopathy genes. molecular polarity of RG was evaluated by the characteristic With this goal in mind, we compiled a library of shRNA targeting apical b-catenin enrichment at the ventricular surface. The RG 30 genes known to be linked to human ciliopathies with progenitor division was evaluated using PH3 and BrdU neurological deficits (Supplementary Table 4). We used this immunolabelling. The number and position of GFP þ /BrdU þ library to query how ciliopathy-related genes affect distinct stages or PH3 þ nuclei were used to examine changes in the rate of cell of brain development, in particular neural progenitor division and the position of progenitor nuclei during cell cycle. development, neuronal migration, neuronal differentiation and Actively proliferating intermediate progenitors were co-labelled early neuronal connectivity, following shRNA-mediated gene with anti-Tbr2 and anti-BrdU antibodies to monitor the changes knockdown or control shRNA expression in embryonic brains. in their proliferation patterns (Fig. 1). These studies give us a comprehensive insight into how human The extent of migration of newly generated GFP þ neurons ciliopathy-related genes affect distinct stages of cerebral cortical was assessed by measuring the distance between ventricular formation, and reveal hitherto undefined neurodevelopmental surface and the leading front of GFP þ migratory neurons, and by pathways whose disruption is likely to be integrally related to the binning the distribution of GFP þ neurons across the developing development of neurological deficits in human ciliopathies. cortical wall. Directionality (polarity) of neuronal migration was categorized as ‘oriented’ or ‘misoriented’ based on the angle of orientation of the leading process of migrating neurons relative to Results the pial surface (oriented: 75°–90°, misoriented:o75°). The final shRNA library of ciliopathy genes. Mutations in 87 cilia-related positioning of cortical neurons within the developing cortical genes have thus far been associated with human ciliopathies layers was investigated by co-labelling GFP þ neurons with (Supplementary Tables 1–3). Mutations in 77 of these genes are antibodies to layer-specific markers Tbr1 (layer VI), Ctip2 known to be associated with neurobehavioural or neurodeve- (layer V) and Cux1 (layers II and III). Distribution of co-labelled lopmental deficits in humans (Supplementary Tables 1–3). The neurons were quantified and compared between control and developmental and neurological consequences of mutations shRNA-expressing groups (Fig. 1).

2 NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 ARTICLE

a bc

CP

PH3 Nestin GFP Tbr2 Brdu GFP d Cux1 Ctip2 GFP

IZ

VZ

e f gh

Figure 1 | In utero electroporation-based assays for distinct stages of cortical development. (a,b) Assays for cortical progenitor proliferation. shRNA- electroporated E14 embryos were allowed to survive for 2 or 4 days and shRNA-expressing (GFP þ ) radial progenitors were co-labelled with anti-nestin (red) and mitotic marker anti-PH3 (blue) antibodies (a). Similarly, actively proliferating intermediate progenitors were co-labelled with anti-Tbr2 (blue) and BrdU (red) in b. Changes in proliferation patterns of shRNA-expressing cells (GFP þ ) can be monitored in such assays. (c) Assay for neuronal migration. At 2 days after electroporation, the extent of migration of GFP þ shRNA-expressing neurons away from the ventricular zone (VZ) towards the cortical plate (CP) can be observed and measured in this assay. (d) The laminar identity of shRNA-expressing neurons in the cortical plate can be evaluated by co- labelling with different layer-specific markers. Cortical plate neurons in d were co-labelled with anti-Cux-1 (layer 2/3) and anti-Ctip2 (layer 5). (e) At E18.5, projections of the post-migratory cortical neurons to the contralateral cortex can also be visualized (arrow, e). In higher-magnification images of the cortical plate (f), the growing axon (arrowhead) and dendrites (arrow) of post-migratory neurons in the emerging cortical plate can be clearly visualized. (g,h)The developing dendritic morphology of two different post-migratory neurons. GFP þ cells in these images express control shRNA vector. The use of a ciliopathy gene shRNA library in such assays allow rapid evaluation of the role of ciliopathy-related genes in different aspects of cortical development. CP, cortical plate; IZ, intermediate zone; SVZ, subventricular zone; VZ, ventricular zone. Scale bar, 160 mm(a–c); 210 mm(d); 330 mm(e); 65 mm(f,g); 45 mm(h).

Post-migratory differentiation of GFP þ , shRNA-expressing The effect of ciliopathy-specific genes on cortical progenitors. neurons in the cortical plate (CP) was evaluated by determining The generation and maintenance of the apicobasally polarized RG whether these neurons display an axon, and if the axon projects scaffold is an essential first step in cerebral cortical formation. Of first apically and then laterally towards contralateral cortical or the 30 ciliopathy genes tested, knockdown of four genes (BBS1, subcortical regions. The elaboration of apical and basal neurites BBS7, BBS10 and TMEM216) resulted in disruption of the apical– was evaluated by measuring the length and number of the basal polarity of RG cells (Supplementary Table 6). Control RG primary neurite branches, and the density of filopodia on primary progenitors display characteristic polarized morphology, with cell apical neurites (Fig. 1). soma in the ventricular zone, an apical endfoot, an elongated

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 3 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 polarized basal process oriented towards the pial surface and intermediate precursor (IP) cells. IPs then undergo symmetric branched basal endfeet attached to the pial basement membrane neurogenic divisions and predominantly give rise to the (Fig. 2a–d). Knockdown of BBS1 (Bardet–Biedl syndrome 1) neurons of upper-layer identities. Knockdown of BUBR1 resulted in wavy RG processes (Fig. 2e) and excessively branched (Bub1-related kinase; BUB1B), IFT80 (intraflagellar transport 80 basal endfeet (Fig. 2f, asterisk). In contrast, TMEM216 (trans- homologue), KIF7 ( family member 7) and TMEM216 led membrane 216) or BBS10 (Bardet–Biedl syndrome 10) to decreased cell divisions of both apical radial progenitors deficiency led to short or retracted basal RG processes with (Fig. 3a–e, arrows; Supplementary Fig. 1) and basal IPs as aberrant branching (Fig. 2g,i,j) and clubby endfeet structures indicated by the reduced GFP þ /PH3 þ cells (Fig. 3a–e, (Fig. 2h,k; asterisk). Further, BBS10 knockdown also led to the arrowheads) and basally localized GFP þ /Tbr2 þ , or GFP þ / loss of apical enrichment of b-catenin, and significantly increased Tbr2 þ /BrdU þ cells (Fig. 3f–j; Supplementary Fig. 1; percentage of RG cells without apical endfeet (Fig. 2d,l,m). In Supplementary Table 6). These findings suggest that distinct sharp contrast to the effects of the above three genes, BBS7 ciliopathy genes may selectively modulate progenitor cell division knockdown caused the ventricular epithelium to invaginate and thus, cortical neurogenesis. towards the pial surface, resulting in rosette-like structures with actively dividing progenitors localized in the centre of the rosette lumen and migrating neurons radiating outwards (Fig. 2n,o; Ciliopathy genes and neuronal migration. Newborn neurons Supplementary Fig. 1). Taken together, these results indicate that migrate radially through the intermediate zone (IZ) to reach their different ciliopathy-specific genes (BBS1, BBS7, BBS10 and specific cortical layers in the developing CP. During the initial TMEM216) modulate distinct aspects of apical–basal polarity of phase of radial migration, newborn neurons within the lower IZ the RG scaffold and the integrity of the proliferative niche transiently assume a characteristic ‘multipolar’ morphology to organization. explore their microenvironment for directional cues, and then As cerebral cortex forms, precise control of radial progenitor transform into a bipolar morphology as they migrate radially cell cycle dynamics is required for appropriate precursor pool (Fig. 4a). The bipolar neurons subsequently migrate towards CP expansion and neurogenesis. Symmetrical divisions of RG by adhering to the basal RG process, with their leading process produce daughter radial progenitors and help expand the oriented towards the pial surface (Fig. 4a). Knockdown of BBS1, progenitor pool. As neurogenesis begins, radial progenitors BBS7, BBS10, AHI1 and ALMS1 resulted in a significantly divide asymmetrically and produce daughter neurons and higher percentage of newborn GFP þ neurons remaining as

Ctrl shRNA BBS1shRNA TMEM216 shRNA a b c * e fgh * *

d

RC2 β-Catenin

BBS10 shRNA BBS7 shRNA

i jk* m n o

70 * 60 50 l 40 30 endfeet 20 10 0

RC2 β-Catenin % Of RG cells without apical PH3 RC2 β-Catenin

Figure 2 | Ciliopathy-related gene effects on polarized radial glial progenitor morphology. (a–d) Control radial glial (RG) progenitor scaffold. Polarized RG are labelled with anti-GFP (a) and anti-RC2 (b). The basal processes of control radial glia (a, arrowhead) spanning the entire width of the developing embryonic cerebral cortices were labelled with electroporated GFP. (c) Higher-magnification view of basal endfeet (adjacent to asterisk) labelled with GFP. (d) Higher-magnification view of GFP þ radial glial cell bodies in VZ and apically enriched b-catenin (red) lining the ventricular surface. Radial glial cell soma is localized either adjacent to the ventricular surface (arrowhead) or just above with elongated apical endfeet (arrow). (e,f) BBS1 shRNA knockdown results in RG with wavy basal processes (square; e) and disrupted basal endfeet (see endfeet adjacent to asterisk; f). (g,h) TMEM216-deficient radial glia do not fasciculate as densely as in controls (square; g) and are often shorter and misoriented (arrowhead; g). (h) Disrupted basal endfeet in TMEM216 shRNA-expressing RG cells (see endfeet adjacent to asterisk). (i–m) Knockdown of BBS10 lead to aberrantly branched, short or retracted (arrowhead; f,i) RG scaffold. (k) Enlarged view of GFP þ -expressing BBS10 shRNA basal endfeet illustrates the unbranched morphology of basal endfeet (see endfeet adjacent to asterisk). (l) BBS10 shRNA expression results in loss of apical endfeet, and apical enrichment of b-catenin (red). (m) Quantification of percentage BBS10 shRNA-expressing RG without apical endfeet. Number of cells: 134 (control); 189 (BBS10). (n,o) BBS7 shRNA expression results in rosette-like organization of the ventricular zone and aberrant b-catenin expression. Data shown are mean±s.e.m. *Po0.05 (Student’s t-test). Number of brains per group ¼ 4. Scale bar, 20 mm(a,b,e,g,i,j,n,o); 5 mm(c,d,f,h,k,l).

4 NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 ARTICLE

Ctrl shRNA BUBR1 shRNA TMEM216 shRNA IFT80 shRNA KIF7 shRNA a bc d e PH3 GFP

fghi j BrdU Tbr2 GFP

klCtrl BURU1 TMEM216 IFT80 KIF7 * * * * * * 18 40 * * * * 16 * 35 * 14 30 * cells + * 12 25 * 10 * /PH3

% 20 + 8 15 6 4 10

% Of GFP 2 5 0 0 Ctrl BUBR1 TMEM216 IFT80 KIF7 GFP+/BrdU+ GFP+/Tbr2+ GFP+/BrdU+/Tbr2+

Figure 3 | Effect of ciliopathy genes on progenitor proliferation. (a–e) Co-labelling with anti-GFP and PH3 antibodies show proliferating, GFP þ/PH3 þ RG progenitors in VZ (arrow) and IPs in SVZ (arrowhead), respectively, in control cortices (a). (b–e) In contrast, knockdown of BUBR1 [BUB1B] (b), TMEM216 (c), IFT80 (d) and KIF7 (e) resulted in significantly reduced proliferation of RG progenitors and IPs. (f–j) Co-labelling with anti-GFP, BrdU and Tbr2 antibodies indicates that knockdown of BUBR1 [BUB1B] (g), TMEM216 (h), IFT80 (i) and KIF7 (j) resulted in reduced percentage of Tbr2 þ cells (Tbr2 þ /GFP þ (open arrowhead)) and mitotic progenitors (BrdU þ /GFP þ (arrow), Tbr2 þ /BrdU þ /GFP þ (filled arrowhead)). (k) Quantification of percentage of PH3 þ /GFP þ cells in control and BUBR1 [BUB1B], TMEM216, IFT80 and KIF7 groups. Data shown are mean±s.e.m. *Po0.05 (Student’s t-test). (l) Quantification of percentage of BrdU þ /GFP þ, Tbr2 þ /GFP þ and Tbr2 þ /BrdU þ /GFP þ cells in control and BUBR1 [BUB1B], TMEM216, IFT80 and KIF7 shRNA groups. Data shown are mean±s.e.m. *Po0.05 (Student’s t-test). Number of brains per group ¼ 4. Scale bar, 10 mm(a–j).

Tuj-1 þ (and Tbr2 À ) multipolar neurons in the lower IZ at E16.5 (Fig. 4i,k,l,v (arrows); Supplementary Fig. 2d,e,h,i (arrows)), and (Fig. 4b–g), suggesting a defect in the appropriate multipolar-to- misorientation in BBS1-, NPHP8 [RPGRIP1L]- and TCTN2- bipolar transition. Notably, knockdown of these genes only deficient neurons (Fig. 4k,t,x; Supplementary Fig. 2l (arrowhead); delayed but not permanently stalled this transition, as indicated Supplementary Table 7). Moreover, knockdown of BBS4, BBS12, by the presence of bipolar migratory neurons at E18.5 in these BUBR1 [BUB1B], IFT80, KIF7, NPHP1, NPHP8 [RPGRIP1L] cortices (Fig. 4; Supplementary Fig. 1). Compared with control, and TUB resulted in shortened leading processes (Fig. 4; knockdown of BBS7, ALMS1 (Alstrom syndrome 1) and KIF7 Supplementary Fig. 2; Supplementary Table 7), whereas knock- resulted in thinner and shorter processes of the multipolar neu- down of BBS10 and NPHP8 [RPGRIP1L] led to aberrantly rons, whereas knockdown of BBS10 resulted in longer processes elongated trailing processes (Fig. 4u,x (arrowheads)). Taken (Fig. 4m–r; Supplementary Fig. 2; Supplementary Table 7). BBS7 together, these results demonstrate that ciliopathy genes can knockdown also led to an increased number of processes, whereas regulate distinct aspects of cortical neuron migration, including the knockdown of ALMS1 and KIF7 led to decreased number of transient multipolar stage, multipolar-to-bipolar transition and processes in multipolar neurons (Fig. 4m–p,r; Supplementary glial-guided radial migration. Although control shRNA expression Fig. 2; Supplementary Table 7). consistently did not reveal an altered array of migration patterns, a In total, knockdown of 17 ciliopathy genes including AHI1, recent study on doublecortin and neuronal migration highlights ALMS1, BBS1, BBS4, BBS7, BBS9, BBS10, BBS11 [TRIM32], BBS12, the importance of further validation of neuronal migration effects BUBR1 [BUB1B], IFT80, KIF7, NPHP1 (nephronophthisis 1), observed with shRNAs using additional in vivo mouse genetic NPHP8 [RPGRIP1L], TCTN2, TMEM216 and TUB (Tubby tools14. protein homologue) resulted in delayed migration (Fig. 4a–l; Supplementary Figs 1 and 2; Supplementary Table 7). The migration delay is associated with increased leading process Ciliopathy genes and the identity and organization of neurons. branching in BBS1-, BBS4-, BBS10-, BUBR1 [BUB1B]-, Upon reaching the CP, newly generated neurons migrate past IFT80-, NPHP8 [RPGRIP1L]- and TUB-deficient neurons older neurons to occupy more superficial regions of the CP,

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 5 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 where they terminate their migration and acquire distinct laminar resulted in a significantly increased percentage of GFP þ /Cux1 þ and neuronal subtype identities. Compared with control, neurons localized at deeper CP positions where Ctip2 þ or knockdown of AHI1, ALMS1, BBS1, BBS4, BBS7, BBS9, BBS10, Tbr1 þ neurons are normally localized (Supplementary Table 8; BBS11 [TRIM32], BBS12, BUBR1 [BUB1B], IFT80, KIF7, Supplementary Fig. 3). Therefore, consistent with their role in NPHP1, NPHP8 [RPGRIP1L], TCTN2, TMEM216 and TUB neuronal migration, knockdown of these genes impair the

E16.5 g Ctrl shRNA BBS1 shRNA BBS7 shRNA BBS10 shRNA AHI1 shRNA ALMS1 shRNA Ctrl BBS1 BBS7 AHI1 ALMS1 0 abcdef 10 CP 20 30

cells 40 + 50 60 IZ 70 % Of GFP 80 90 VZ /SVZ 100 VZ/SVZ IZ CP

E18.5 Ctrl shRNA BBS10 shRNA BBS12 shRNA NPHP8 shRNA TUB shRNA h ijkl

CP

IZ

VZ /SVZ

Ctrl shRNA BBS10 shRNA BBS12 shRNA NPHP8 shRNA TUB shRNA 0% 10% 20% 30% 40% 50% 0% 10% 20% 30% 40% 50% 0% 10% 20% 30% 40% 50% 0% 10% 20% 30% 40% 50% 0% 10% 20% 30% 40% 50% Pia 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5 6 6 6 6 6 7 7 7 7 7 8 8 8 8 8 9 9 9 9 9

Position in cerebral wall 10 10 10 10 10 vs

Ctrl shRNA BBS7 shRNA KIF7 shRNA ALMS1 shRNA BBS10 shRNA mn o pq

r Ctrl BBS7 KIF7 ALMS1 BBS10 Ctrl shRNA BBS1 shRNA BBS10 shRNA TUB shRNA KIF7 shRNA NPHP8 shRNA 3 * * stuvwx * 2.5 * 2 * * 1.5 * 1 0.5 processes (fold basal) 0 Changes in multipolar neuron Length Number

6 NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 ARTICLE laminar placement of neurons with upper-layer identities. function remains largely unknown. In this study, we carried out a Quantification of the percentage of GFP þ /layer marker þ (Cux1, systematic in vivo characterization of the role of 30 ciliopathy Ctip2 or Tbr1) neurons indicate no significant differences genes in major steps of cerebral cortical formation, ranging from between control and shRNA groups, suggesting that neuronal progenitor development, neuronal migration, neuronal differ- layer identity was not altered by shRNA expression entiation and early neuronal connectivity. Importantly, this (Supplementary Table 8). Together these data demonstrate a electroporation-based approach enables the exploration of cilia- critical role for groups of ciliopathy genes including AHI1, related gene functions in cerebral cortical development while ALMS1, BBS1, BBS4, BBS7, BBS9, BBS10, BBS11 [TRIM32], avoiding hydrocephalus, noted often in cilia mutants. Our BBS12, BUBR1 [BUB1B], IFT80, KIF7, NPHP1, NPHP8 observations provide a comprehensive survey of ciliopathy gene [RPGRIP1L], TCTN2, TMEM216 and TUB in the coordination of functions in the context of cerebral cortical development. The migration and placement of neurons in the developing cerebral developmental expression patterns of ciliopathy genes often cortex. correlate with their distinct functions during cerebral cortical development16–18 (Supplementary Fig. 4). Our results link deficiency in 17 of the ciliopathy genes to a variety of Ciliopathy genes in post-migratory differentiation of neurons. neurodevelopmental defects that may help us understand the After arriving at their laminar destination within the developing diverse clinical features of ciliopathies, including cortical CP, projection neurons extend axons and dendrites as they hypoplasia, ectopias and axonal fibre tract defects, and the assemble into functional neural circuits. During this process, neu- resultant functional outcomes such as intellectual disabilities1,4,19. rons initially extend an axon, through the corpus callosum to the Currently, the known ciliopathy-related gene mutations are contralateral cortex or through the internal capsule to sub-cerebral thought to lead to loss of function of respective . The targets. In addition, neurons also extend apical dendrites towards loss-of-function approach used in our screen is consistent with the pial surface and extensive basal dendrites from the cell soma. this, but is limited in its ability to model gain-of-function The dendrites of cortical neurons also develop small protrusions mutations or human alleles that are hypomorphic. However, called spines, which are the major sites of synaptic contacts. We further examination of these gene functions with shRNA-resistant found that knockdown of BBS5, BBS7 resulted in significantly gene constructs (Supplementary Fig. 1), gene-specific small reduced axon outgrowth (Supplementary Table 9; Fig. 5a,f,i,j). The hairpin microRNA (shmiRNA), mouse genetic tools and related axons of BBS5-, BBS7-, BBS9-, BBS11 [TRIM32]-, BBS12- and human gene mutations will be necessary to fully understand the TMEM216-deficient neurons also display aberrant trajectory and role of these genes in brain developmental abnormalities in fasciculation (Fig. 5a–e,k,l). Knockdown of BBS1, BBS5, BBS7, ciliopathies14. Use of human mutant alleles will be particularly BBS11 [TRIM32], BBS12 and KIF7 caused a substantial reduction helpful in uncovering potential gain-of-function mutations. Gene in axons that project across the midline to the contralateral cortex manipulations at developmental stages other than what were (Fig. 5f–l; Supplementary Fig. 1; Supplementary Table 9). This described in this study may help confirm or reveal additional defect in midline crossing in BBS5, BBS7-deficient axons is caused stage-specific effects of ciliopathy genes, especially for the ones not only by reduced axon outgrowth but also by disrupted axon without any detectable phenotypes in our assays. Cell biological guidance. These axons reached midline, but instead of crossing, exploration of these gene functions in cilia compartments and the project aberrantly towards subcortical targets (Fig. 5i,j). Although associated centrosome is also necessary. Nevertheless, this study the dendrites and dendritic spines are not fully formed at E18, provides a template for further exploration of the biological basis deficiency in BBS9, BBS10, BBS11 [TRIM32], BBS12, NPHP1 or of the brain structural and functional deficits associated with KIF7 resulted in reduced apical neurite branching and total length ciliopathies. These new insights into the functions of ciliopathy (Fig. 5m–u; Supplementary Fig. 1; Supplementary Table 9). In genes help link malformations seen in ciliopathies addition, BBS9, BBS11, BBS12 and NPHP1 knockdown also to disruptions in specific cortical developmental events (Fig. 6). resulted in a lower density of filopodia (Supplementary Table 9), 15 The four ciliopathy genes (BUBR1 [BUB1B], IFT80, KIF7 and many of which eventually form dendritic spines . In summary, TMEM216) we identified as modifiers of progenitor proliferation these results suggest that ciliopathy-related genes BBS1, BBS5, encode proteins localized to distinct ciliary compartments BBS7, BBS9, BBS10, BBS11 [TRIM32], BBS12, KIF7, NPHP1 and (Supplementary Table 1). BUBR1 [BUB1B], localized in centro- TMEM216, differentially modulate appropriate patterns of post- some/, is a key spindle assembly checkpoint migratory neuronal differentiation in cerebral cortex, including protein, whose insufficiency leads to cell cycle misregulation axonal growth, axonal guidance, neurite extension and and defects20. Importantly, mutations in BUBR1 arborization. [BUB1B] cause a rare human disorder mosaic-variegated aneuploidy or premature chromatid separation syndrome, a Discussion novel ciliopathy syndrome characterized by microcephaly Structural and functional neurological deficits are a common and mental retardation20,21. TMEM216 is a component of feature of ciliopathies in humans. However, the role of primary transition zone and is required for centrosome/basal body cilia-related genes linked to ciliopathies in brain development and docking to initiate ciliogenesis22–24. TMEM216 is a causative

Figure 4 | Effect of ciliopathy genes on neuronal migration. (a–l) GFP þ neurons express shRNAs against BBS1 (b), BBS7 (c), BBS10 (d,ji)i, AHI1 (e), ALMS1 (f), RTTN (i), BBS12 (j) and NPHP8 [RPGRIP1L] (k)andTUB(l) shRNAs do not migrate normally. shRNAs were electroporated at E14.5. a–f and h–l are from cortical analyses at E16.5 and E18.5, respectively. Arrows in i, k and l indicate branched leading processes. Arrowhead in k indicates a misoriented leading process. (g) Quantification of GFP þ cell position in the developing cortical wall (E16.5). Quantification of GFP þ cell position in the developing cerebral wall (E18.5) are shown under each panel in h–l.(m–q) Higher-magnification images of GFP þ multipolar neurons in control, BBS7, KIF7, ALMS1 and BBS10 shRNA groups. Red arrowheads indicate processes. (r) Quantification of average number and length of process in GFP þ multipolar neurons in control and BBS7, KIF7, ALMS1 and BBS10 shRNA groups. Data shown are mean±s.e.m. *Po0.05 (Student’s t-test). (s–x) Representative images of GFP þ migrating neurons from control and BBS1, BBS10, TUB, KIF7 and NPHP8 [RPGRIP1L] shRNA groups. Arrowhead, trailing process of migrating neurons. Number of brains per group ¼ 4. CP, cortical plate; IZ, intermediate zone; SVZ, subventricular zone; VZ, ventricular zone. Scale bar, 60 mm(a–f); 65 mm(h–l); 15 mm(m–x).

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 7 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857

Ctrl shRNA BBS5 shRNA BBS7 shRNA BBS9 shRNA TMEM216 shRNA abcde

Ctrl shRNA KIF7 shRNA BBS1 shRNA fgh

BBS5 shRNA BBS7 shRNA BBS11 shRNA i j k

BBS12 shRNA Ctrl shRNA BBS9 shRNA BBS10 shRNA BBS11 shRNA BBS12 shRNA NPHP1 shRNA KIF7 shRNA l mn o p q r s

tu* v * * * * * * * * * * 6 * 160 * * * 140 5 m) of 1.8 * 120 μ 1.6 4 m) 1.4

μ 100 3 80 1.2 1 2 60 0.8 Length ( 40 filopodia 0.6 1 0.4 20 0.2

# of apical neurite braches 0 0 0

Ctrl Ctrl Ctrl BBS9 KIF7 BBS9 KIF7 density (per 10 Average BBS9 BBS10 BBS11 BBS12 NPHP1 BBS10 BBS11 BBS12 NPHP1 BBS11 BBS12 NPHP1 Figure 5 | Effect of ciliopathy genes on post-migratory neuronal differentiation. Embryos were electroporated with shRNAs at E14.5 and analysed at E18.5. (a–e,k,l)GFPþ neurons expressing BBS5 (b), BBS7 (c), BBS9 (d), TMEM216 (e), BBS11 (k) and BBS12 (l) shRNAs display axonal fasciculation defects (compare areas indicated by green arrowheads, insets (a–e)). (f–l) GFP þ neurons expressing shRNA against KIF7 (g), BBS1 (h), BBS5 (i), BBS7 (j), BBS11 (k) and BBS12 (l) display midline crossing defects in the developing cerebral wall. GFP þ neurons expressing shRNA against BBS5 (i), and BBS7 (j) also display reduced axonal outgrowth (compare areas indicated by yellow arrowheads in f, i and j. Right panels in f–l show higher-magnification images of the outlined (box) area in f–l.(m–s) Neurons expressing BBS9 (n), BBS10 (o), BBS11 (p), BBS12 (q), NPHP1 (r) and KIF7 (s) shRNAs display significantly reduced total non-axonal neurite length and branching. (t,u) Quantification of average apical neurite number (t) and total length (u) in control and BBS9-, BBS10-, BBS11-, BBS12-, NPHP1- and KIF7-deficient neurons. (v) Quantification of filopodial density in control and BBS9-, BBS11-, BBS12- and NPHP1-deficient neurons. Data shown are mean±s.e.m. *Po0.05 (Student’s t-test). Number of brains per group ¼ 4. Scale bar, 80 mm(a–e); 148 mm(f–l); 18 mm(m–s).

8 NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 ARTICLE

L1

7 L2 BBBS9,BS9, BBS10,BBS10, BBBS11,BS11, BBS12,BBS12, KKIF7,IF7, NNPHP1.PHP1. L3 PMN 4 L4 AHI1,AHI1, ALMS1,ALMS1, BBS1,BBS1, BBBS4,BS4, BBS7,BBS7, BBBS9,BS9, BBS10,BBS10, BBBS11,BS11, BBS12,BBS12, BBUBR1,UBR1, L5 IFT80,IFT80, KIF7,KIF7, NPHP1,NPHP1, RMN NPHP8,NPHP8, TCTN2,TCTN2, TMEM216,TMEM216, TUB.TUB. CP L6

5 BBBS5,BS5, BBBS7.BS7. 6 BBBS1,BS1, BBBS5,BS5, BBBS7,BS7, 3 BBBS9,BS9, BBBS11,BS11, BBBS12,BS12, BBS7,BBS7, BBS10,BBS10, KKIF7,IF7, TTMEM216.MEM216. IZ MPN KIF7,KIF7, AALMS1.LMS1.

MPN 2 SVZ IPC ORGC

VZ RGC 1 2 BBS1,BBS1, BBBS7,BS7, BUBR1,BUBR1, IFT80,IFT80, BBS10,BBS10, TMEM216.TMEM216. KIF7,KIF7, TMEM216.TMEM216.

Figure 6 | Ciliopathy genes and cerebral cortical development. Schematic of major steps (1–7) of cerebral cortical development. Polarized radial progenitor formation and organization (1), radial glial progenitor and intermediate progenitor proliferation (2), multipolar-to-bipolar transition of newborn neurons (3), radial glial-guided neuronal migration (4), axon outgrowth (5), axon guidance and fasciculation (6) and apical dendritic outgrowth (7) are illustrated. Ciliopathy-related genes identified to play a role in each of these steps are indicated. CP, cortical plate; IPC, intermediate progenitor cell; IZ, intermediate zone; L1–L6, layers 1–6; MPN, multipolar neuron; ORGC, outer radial glial cell; PMN, post-migratory neuron; RGC, radial glial cell; RMN, radially migrating neuron; SVZ, subventricular zone; VZ, ventricular zone.

gene for ciliopathy syndromes and related heterotopias are an abnormal brain patterning defect commonly disorders (JSRD) and Meckel–Gruber syndrome (MKS), with seen in Bardet–Biedl syndrome (BBS), Joubert syndrome (JBTS), associated mental retardation22–24. Our observation of disrupted Meckel–Gruber syndrome (MKS)9,13,28, caused by defects in the neural progenitor proliferation in BUBR1 [BUB1B]- and formation and maintenance of polarized RG progenitors and RG- TMEM216-deficient brains is consistent with the clinical guided neuronal migration. Importantly, mutations in BBS1, symptoms, such as microcephaly, found in human patients. On BBS7 and BBS10 cause BBS, whereas mutations in TMEM216 the other hand, IFT80 and KIF7 are members of IFT complexes cause JBTS and MKS19,24,29. The presence of cortical heterotopia and deficiency of IFT80 or KIF7 results in impaired ciliary in these ciliopathies is associated with epilepsy and mental signalling transduction without loss of cilia25–27. Mutations of retardation2,4,7,30,31. Taken together, these data demonstrate the IFT80 cause Jeune asphyxiating thoracic dystrophy and short-rib converging role of multiple ciliopathy genes in the development polydactyly type III (SRP type III)25. Mutations in KIF7 cause and function of RG progenitors. Disruption of these processes hydrolethalus and acrocallosal syndromes26. The four ciliopathy may underlie cortical malformations such as microcephaly and genes we identified as regulators of progenitor proliferation could heterotopias seen in ciliopathies related to mutations in these represent two distinct molecular pathways: centrosome genes. dependent and ciliary signalling dependent, underlying some of Defects in neuronal migration and differentiation of cortical the cortical deficits (for example, microcephaly) associated with neurons can impair the organization and activity of the cortical these gene mutations. circuitry, and are thought to underlie a broad spectrum of Knockdown of BBS1, BBS10, BBS7 and TMEM216 lead to neurodevelopmental and psychiatric disorders32,33. Neuronal disruption of distinct aspects of the polarized RG scaffold. BBS1, migration defects such as cortical periventricular heterotopia BBS10 and TMEM216 are required for the maintenance of the and polymicrogyria are common features of various ciliopathies RG scaffold; BBS10 is required for the polarized organization of (Supplementary Table 1). Surprisingly, our results show that 17 the RG scaffold, whereas knockdown of BBS7 resulted in the different ciliopathy genes modulate distinct phases of cortical formation of rosette-like heterotopias caused by invagination of neuron migration, highlighting the critical nature of primary cilia the progenitor niche. These observations suggest their converging in neuronal migration during cortical development. These results yet differential influence of cilia-related genes on the RG polarity. also provide molecular bases and genotype–phenotype Further understanding of the functions of these proteins localized correlations for the brain abnormalities caused by aberrant to different ciliary compartments in the regulation of distinct neuronal migration in ciliopathies. For example, ciliopathy aspects of RG polarity could help dissect the molecular patients carrying mutations in AHI1 or BBS genes show cortical mechanisms underlying RG organization and function. Cortical heterotopia and polymicrogyria29,34. Consistently, our results

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 9 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 suggest that neurons expressing AHI1 or BBS (BBS1, BBS4, BBS5, example, Shh functions to guide commissural axons by acting BBS7, BBS9, BBS10, BBS11 [TRIM32] and BBS12) shRNA show directly as a chemoattractant for the axon growth cones. migration defects. We also observed that knockdown of ciliopathy Consistently, mutations in KIF7, known to disrupt the Shh genes ALMS1, BUBR1 [BUB1B], IFT80, NPHP1, NPHP8 pathway, cause acallosal syndromes in humans and lead to [RPGRIP1L], TCTN2, TMEM216 and TUB retarded neuronal dysgenesis or agenesis of the corpus callosum in mice26,27,54. Our migration. Neurons deficient in these genes display excessive observations indicated that in addition to KIF7, knockdown of branching or misorientated leading processes. Since knockdown five of the BBS genes, BBS1, BBS5, BBS7, BBS11 [TRIM32] and of BBS1, BBS10 and TMEM216 also led to disrupted RG BBS12, also led to axonal guidance and midline crossing defects. morphology, the migration defects seen in these three ciliopathy How these cilia-related proteins, localized primarily in cilium gene knockdowns may in part result from a compromised RG near the cell soma, modulate dynamics of the axonal growth scaffold, whereas the migration defects with the other genes could cone remains unclear. The answer might lie in the non-ciliary be due to their specific neuronal functions. Knockdown of five expression of ciliopathy genes at other sub-cellular compart- genes (BBS7, BBS10, ALMS1, KIF7 and AHI1) led to defects in the ments, such as the growth cone. To date, several ciliopathy genes mutipolar–bipolar transition stage, suggesting a selective have been shown to have non-ciliary expression (for example, requirement for primary cilia in the early stages of neuronal KIF7 in the growth cone, KIF3A, BBS4 and BBS8 in synapse) or migration. Though the molecular mechanisms involved in the cilia-independent functions (for example, KIF7 in axonal transition from multipolar to radial migration are yet to be fully guidance and BBS4 in synaptic transmission), suggesting understood, multiple cues such as adhesion molecules (for additional key roles for some of the ciliopathy genes in protein example, TAG1, N-cadherin and Connexin 43), cytoskeletal trafficking and membrane transport54–59. Future effort aimed at regulators (for example, CDK5 and ARX), intracellular signalling dissecting the role of ciliary versus non-ciliary function of molecules (for example, , RhoA, Mark2 and LKB1), ciliopathy genes during corticogenesis will be essential to transcription factors (for example, FoxG1, Scratch 1, 2 and delineate how they contribute to the aetiology of constellation KLF4) and extracellular signals such as Neurog2, Semaphorin 3A, of symptoms associated with ciliopathies. Netrin and Reelin are involved in this process35–43. Our results Many of the causative genes of ciliopathy syndromes display suggest that primary cilium and its associated signalling proteins genetic and/or direct physical interactions, leading possibly to could help sense, integrate and convey signalling cues necessary perturbations in common cellular processes6,60. For example, it for this transition. has been shown that KIF7 genetically interacts with BBS loci26. Consistent with the migration defects, we found an aberrant Deficiency of KIF7 and several BBS genes led to similar defects in laminar placement of neurons following disruption of 17 progenitor proliferation and neuronal migration, indicating ciliopathy genes (AHI1, ALMS1, BBS1, BBS4, BBS7, BBS9, convergent biological influences of genetic interaction between BBS10, BBS11 [TRIM32], BBS12, BUBR1 [BUB1B], IFT80, KIF7 and BBS genes. On the contrary, some ciliopathy-related KIF7, NPHP1, NPHP8 [RPGRIP1L], TCTN2, TMEM216 and genes can exert effects on multiple aspects of human brain TUB). Recent studies also showed that genetic deficiency of BBS1 development (Supplementary Table 10). For example, mutations or BBS4 in mice results in perturbed migration and lamination in AH1 can disrupt neurogenesis, neuronal migration and axon defects32,44. The identification of multiple ciliopathy genes with growth, leading to hypoplasia, heterotopia and failed axonal previously unknown roles in neuronal migration and placement decussation in humans (Supplementary Table 10). If and how suggest that this stage of cortical development is highly vulnerable AH1 convergently interacts with other ciliopathy genes to in ciliopathies, and the resultant changes in the formation of modulate these diverse cellular processes remains unclear. cortical circuits may underlie cognitive deficits associated with On the basis of localization in cilia and functional interactions, ciliopathies. ciliopathy genes have also been classified into functional Post-migratory cortical neurons undergo extensive neurite networks6,61. BBSome is the best studied of the ciliopathy outgrowth to generate the appropriate axon–dendritic architec- protein complexes. BBSome contains seven highly conserved BBS ture of the neuronal circuits. A steady and polarized delivery of proteins (BBS1, BBS2, BBS4, BBS5, BBS7, BBS8 [TTC8] and membrane proteins, such as guidance cue receptors, and BBS9) that localize to the basal body and functions as a unit that cytoskeletal components to the growing neurites is required to mediates vesicle trafficking to the ciliary membrane62. Three enable the neurites to extend in appropriate directional patterns -like BBS proteins (BBS6 [MKKS], BBS10 and as they contact and form synapses with appropriate partners. BBS12) form a higher-order complex, mediate BBSome subunit Consistently, primary cilia-mediated signalling cascades (for association and are required for BBSome assembly62,63. example, GPCR and Wnt signalling) and cilia-associated centro- Deficiency of BBS genes cause Bardet–Biedl syndrome, one of some mechanisms have both been shown to regulate neuronal the well-characterized ciliopathies29. Our results show that dendritic and axonal outgrowth and refinement45–48. We found deficiency in eight BBS protein family members (BBS1, BBS4, that two BBS genes (BBS5 and BBS7) are required for axonal BBS5, BBS7, BBS9, BBS10, BBS11 [TRIM32] and BBS12) led to outgrowth, whereas one transition zone gene (NPHP1), four BBS significant cortical defects including a disrupted RG scaffold, genes (BBS9, BBS10, BBS11 [TRIM32] and BBS12) and an IFT perturbed neuronal migration and impaired neuronal regulator, KIF7, modulate neurite outgrowth and filopodial differentiation, thus highlighting the essential roles of BBS formation, indicating the differential and specific influence of genes in brain development. Importantly, the results of our ciliopathy genes over post-migratory neuronal differentiation. functional assays further clarify some of the known effects of Defects in axonal tract development, including callosal BBSome ciliopathy in humans: it has been suggested that agenesis, disrupted pyramidal decussation and superior cerebellar deficiency in any member of BBSome or chaperonin-like BBS peduncle decussation, are often seen in ciliopathy patients, and complex disrupts the functionality of the BBSome and causes these axonal tract defects are thought to underlie some of the similar phenotypes in human patients and BBS animal motor and cognitive deficits in ciliopathies2,4,18,26,30,49.Itis models60,63–67. Consistent with this, our results show that known that midline glial structures (for example, glial wedge) and deficiency of BBSome members and chaperonin-like BBS secreted guidance molecules (for example, WNT, Slit2, Shh and proteins often led to defects in common neurodevelopmental Netrin1) released from these structures are crucial for guiding events. We found that BBS1, BBS7 and BBS10 are involved in RG axonal midline crossing and corpus callosum formation49–53. For progenitor development; knockdown of BBS1, BBS4, BBS7, BBS9,

10 NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 ARTICLE

BBS10, BBS11 [TRIM32] or BBS12 resulted in neuronal cells). Transfected IMCD3 cells were then lysed for RNA purification using a migration defects, whereas loss of BBS1, BBS5, BBS7, BBS9, Qiagen RNeasy mini kit according to the manufacturer’s instructions. BBS10, BBS11 or BBS12 led to post-migratory, neuronal differentiation defects. However, our analysis also revealed RT–PCR and qPCR. RT–PCR was performed using the Invitrogen SuperScript III other previously unknown distinctive requirements for First-Strand Synthesis System. Real-time PCR was performed using Applied Bio- system Power SYBR Green PCR master mix and the appropriate forward and individual BBSome members and chaperonin-like BBS proteins reverse primers (Supplementary Table 5). The real-time PCR reactions were per- during corticogenesis. For example, knockdown of BBS1 and formed in triplicates. Peptidylprolyl isomerase A (Ppia) was used as an endogenous BBS10 resulted in excessive branching, and reduced branching of reference. The comparative (DDCT) method of relative quantification was used to RG basal endfeet, respectively (Fig. 2). Also, knockdown of BBS7 determine the level of gene knockdown. and BBS10 led to shorter, and longer multipolar neuronal processes, respectively (Fig. 4). Thus, it is likely that even though In utero electroporation. In utero electroporations were performed as described 68,69 À 1 BBSome and chaperonin-like BBS proteins function to synergize previously . Briefly, 1–2 ml of a pool of plasmid DNA (1.5 mg ml ) were injected into the lateral ventricles of E14.5 brains and electroporated using five and converge on common developmental pathways, they still pulses at 30 V for 50 ms at 950-ms intervals through the uterine wall using a BTX exert different regulatory roles during corticogenesis. The same ElectroSquarePorator (ECM 830). In shRNA þ cDNA rescue experiments, human also appears to be the case for other ciliopathy networks such as cDNAs were used at 1 mg ml À 1 concentration. Each pool of shRNAs or control the NPHP–JBTS–MKS complex. vectors was electroporated into four different embryos. Embryos were then allowed Ciliopathies display a broad range of clinical features, high to develop for 2 or 4 days prior to analyses. degree of pathological variability, genetic heterogeneity and phenotypic overlaps, thus complicating their diagnosis and Immunohistochemistry. The primary antibodies used were anti-RC2 (1:1, Iowa 4 Hybridoma), anti-b-catenin (rabbit polyclonal, 1:1,000;), anti-GFP (chicken, treatment . Elucidation of multifaceted functions of cilia and 1:1,000; Abcam), anti-BrdU (mouse monoclonal, 1:50; BD Biosciences), anti- associated protein complexes is fundamental to our effort to phospho-histone H3 (PH3) (rabbit polyclonal, 1:200; Millipore), anti-Tbr2 (1:500, understand the biological bases of ciliopathies and develop AB23345, Abcam), anti-Ctip2 (1:500, AB18645, Abcam), anti-Cux-1 (1:100, efficient therapeutic interventions. This comprehensive in vivo SC13024, Santa Cruz Biotechnology), anti-Tbr1 (1:500, AB31940, Abcam) and anti-Brn1 (1:1,000, gift from A. Ryan, McGill University)10. Secondary antibodies functional investigation of the causative genes of ciliopathies in were AlexaFluor 488 or Cy3-conjugated (Invitrogen and Jackson the context of cortical development not only delineates varied ImmunoResearch). Nuclei were counterstained with 4’,6-diamidino-2- functions of cilia in the construction of the cerebral cortex, phenylindole (Sigma). Electroporated mouse brains were removed and fixed with but also reveals potential developmental bases of brain 4% paraformaldehyde overnight at 4 °C. Brains were then embedded in 3% agarose and sectioned at 50 mm with a Vibratome (VT1000S; Leica Microsystems). Sections abnormalities in ciliopathies. In the future, the use of causative were blocked with PBS/10% goat serum/0.2% Triton X-100 for 1 h, and incubated human mutant alleles with defined links to cortical abnormalities in primary antibodies (see above) overnight at 4 °C. After three washes in 1 Â PBS, in this type of screen or analysis of genotyped neural cells sections were incubated with secondary antibodies (1:1,000) at room temperature for 2 h, washed and mounted with Citifluor anti-fading solution (Agar). For BrdU generated from ciliopathy patient-derived induced pluripotent À 1 stem cells can help further define the biological basis of ciliopathy labelling, BrdU (50 mg kg ) was injected intraperitoneally into pregnant mice 1 h prior to brain tissue processing. mutations and enhance our understanding of cilia-related gene functions in brain development. Establishing a causal link Quantitative analysis of progenitor proliferation, polarity and niche between genotype and phenotype of ciliopathies with such organization. GFP þ RG progenitors were co-immunolabelled with RC2 or approaches will be vital to develop rational therapeutic anti-nestin antibodies. The characteristic apical endfeet attached to the pial surface interventions for ciliopathies. and branched basal endfeet were imaged and quantified per section in control and shRNA brains. Lack of apical endfeet and increased or decreased branching of basal endfeet were measured. The apical molecular polarity of RG was detected by Methods apical b-catenin enrichment at the ventricular surface of GFP þ radial progenitors. Mice. Mice were cared for according to guidelines approved by the University Changes in RG progenitor proliferation were detected using PH3 and BrdU of North Carolina. Light/dark cycle in the vivarium is 7/7 h. Animals were immunolabelling. The number and position of GFP þ /BrdU þ or PH3 þ nuclei in housed in groups of three adults per cage. C57/BL6 pregnant female mice the ventricular zone was measured (25 Â 103 mm2 area) to examine changes in the were used for in utero electroporation. The day of vaginal plug detection was rate of cell division and the position of progenitor nuclei during cell cycle. Actively considered E0.5. proliferating intermediate progenitors in control and shRNA brains were co- labelled with anti-Tbr2 and anti-BrdU antibodies and quantified (25 Â 103 mm2 area) to monitor the changes in their proliferation patterns. shRNA library of ciliopathy genes. A library of shRNAs specific to the 30 human ciliopathy genes was assembled using the resources available at the UNC Gene Therapy Center. Lists of shRNA constructs and sources are provided in the Quantitative analysis of neuronal migration and placement. Cerebral wall was divided into 10 equal bins and the percentage of GFP þ neurons in each bin was Supplementary Information (Supplementary Table 4). Plasmids used for in utero 68 electroporation were prepared using the EndoFree Plasmid kit (Qiagen). For each counted . Number of leading processes per neuron was counted in migrating gene, two pools of shRNAs containing two to three individual shRNAs were neurons in the IZ. The percentage of migrating neurons with leading processes oriented at an angle o75° to the pial surface was also measured. Percentage of randomly combined and used for electroporation. Non-silencing scrambled þ þ shRNA (sc-108080, Santa Cruz), GIPZ or pLKO.1 shRNA vectors (GE Dharmacon GFP /layer marker (Tbr1, Ctip2 or Cux1) neurons located below CP, away from and UNC Gene Therapy Center) were used as control. KE of shRNAs (fold change their normal laminar locations, was quantified to detect changes in the laminar compared with control) was measured using quantitative real-time PCR and are placement of shRNA-expressing neurons. Images were acquired using Zeiss 710 indicated for each shRNA group. KE of 70% was used as a threshold for shRNA and Zeiss 780 confocal systems. Imaging analysis was done using Zeiss LSM Image pools. Each pool consisted of an equivalent amount of different shRNAs (2 or 3 per Browser and ImageJ software (NIH). pool) listed. For BBS8, only the effective pool of ShRNA containing three different shRNAs was used. shRNAs from Santa Cruz are prepooled (four shRNAs per gene) Quantitative analysis of post-migratory differentiation of neurons. Number of by the vendor. GFP þ , shRNA-expressing neurons in the CP displaying an axon and the projec- tion of the axon apically and then laterally towards contralateral cortical or sub- cortical regions was quantified per section. The length of apical and basal neurites Tissue culture, transfection, RNA isolation and purification . Mouse and the number of the primary neurite branches were counted per GFP þ neuron. inner medullary collecting duct (IMCD3) epithelial cells were grown in DMEM/ The density of immature filopodia on primary apical neurites was quantified per F12 media supplemented with 10% fetal bovine serum and 1% penicillin– 10-mm length. streptomycin (P/S) antibiotics at 37 °C with 5% CO2. The cells were allowed to reach 80% confluency and ciliogenesis was induced by serum starvation of the cells for 24 h. IMCD3 cells were then transfected using Invitrogen Lipofectamine 2000 Statistical analysis. GraphPad or Excel was used for data analysis. Two-tailed reagent according to the manufacturer’s instructions (Sigma). For each well of cells Student’s t-test and two-way analysis of variance with Tukey–Kramer multiple in a six-well plate, 1.5 mg of shRNAs and pCIG2 was used. The cells were imaged 3 comparison test were performed using GraphPad. Data were collected and pro- days post transfection to measure transfection efficiency (percentage of GFP þ cessed blindly. All data are expressed as means±s.e.m. Statistical comparisons

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 11 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857

between different knockdown groups were not performed, as there are differences 32. Kamiya, A. et al. Recruitment of PCM1 to the centrosome by the cooperative in the efficacy of the knockdown of different genes. action of DISC1 and BBS4: a candidate for psychiatric illnesses. Arch. Gen. Psychiatry 65, 996–1006 (2008). 33. Me´tin, C., Vallee, R. B., Rakic, P. & Bhide, P. G. Modes and mishaps of neuronal References migration in the mammalian brain. J. Neurosci. 28, 11746–11752 (2008). 1. Guemez-Gamboa, A., Coufal, N. G. & Gleeson, J. G. Primary cilia in the 34. Dixon-Salazar, T. et al. Mutations in the AHI1 gene, encoding jouberin, developing and mature brain. Neuron 82, 511–521 (2014). cause Joubert syndrome with cortical polymicrogyria. Am. J. Hum. Genet. 75, 2. Lancaster, M. A. & Gleeson, J. G. The primary cilium as a cellular signaling 979–987 (2004). center: lessons from disease. Curr. Opin. Genet. Dev. 19, 220–229 (2009). 35.Jossin,Y.&Cooper,J.A.Reelin,Rap1andN-cadherinorientthemigrationof 3. Louvi, A. & Grove, E. A. Cilia in the CNS: the quiet organelle claims center multipolar neurons in the developing neocortex. Nat. Neurosci. 14, 697–703 (2011). stage. Neuron 69, 1046–1060 (2011). 36. Pacary, E. et al. Proneural transcription factors regulate different steps of 4. Hildebrandt, F., Benzing, T. & Katsanis, N. Ciliopathies. N. Engl. J. Med. 364, cortical neuron migration through Rnd-mediated inhibition of RhoA signaling. 1533–1543 (2011). Neuron 69, 1069–1084 (2011). 5. Marley, A. & Von Zastrow, M. A simple cell-based assay reveals that diverse 37. Tabata, H. & Nakajima, K. Multipolar migration: the third mode of radial neuropsychiatric risk genes converge on primary cilia. PLoS ONE 7, e46647 (2012). neuronal migration in the developing cerebral cortex. J. Neurosci. 23, 6. Sang, L. et al. Mapping the NPHP-JBTS-MKS protein network reveals 9996–10001 (2003). ciliopathy disease genes and pathways. Cell 145, 513–528 (2011). 38. Miyoshi, G. & Fishell, G. Dynamic FoxG1 expression coordinates the 7. Novarino, G., Akizu, N. & Gleeson, J. G. Modeling human disease in humans: integration of multipolar pyramidal neuron precursors into the cortical plate. the ciliopathies. Cell 147, 70–79 (2011). Neuron 74, 1045–1058 (2012). 8. Eggenschwiler, J. T. & Anderson, K. V. Cilia and developmental signaling. 39. Ohshima, T. et al. Cdk5 is required for multipolar-to-bipolar transition during Annu. Rev. Cell Dev. Biol. 23, 345–373 (2007). radial neuronal migration and proper dendrite development of pyramidal 9. Higginbotham, H. et al. Arl13b-regulated cilia activities are essential for neurons in the cerebral cortex. Development 134, 2273–2282 (2007). polarized radial glial scaffold formation. Nat. Neurosci. 16, 1000–1007 (2013). 40. Friocourt, G. et al. Cell-autonomous roles of ARX in cell proliferation and 10. Higginbotham, H. et al. Arl13b in primary cilia regulates the migration neuronal migration during corticogenesis. J. Neurosci. 28, 5794–5805 (2008). and placement of interneurons in the developing cerebral cortex. Dev. Cell 23, 41. Torii, M., Hashimoto-Torii, K., Levitt, P. & Rakic, P. Integration of neuronal 925–938 (2012). clones in the radial cortical columns by EphA and ephrin-A signalling. Nature 11. Jacoby, M. et al. INPP5E mutations cause primary cilium signaling defects, 461, 524–528 (2009). ciliary instability and ciliopathies in human and mouse. Nat. Genet. 41, 42. Evsyukova, I., Plestant, C. & Anton, E. S. Integrative mechanisms of oriented 1027–1031 (2009). neuronal migration in the developing brain. Annu. Rev. Cell Dev. Biol. 29, 12. Conduit, S. E., Dyson, J. M. & Mitchell, C. A. Inositol polyphosphate 5- 299–353 (2013). phosphatases; new players in the regulation of cilia and ciliopathies. FEBS Lett. 43. Namba, T. et al. Pioneering axons regulate neuronal polarization in the 586, 2846–2857 (2012). developing cerebral cortex. Neuron 81, 814–829 (2014). 13. Willaredt, M. A. et al. A crucial role for primary cilia in cortical morphogenesis. 44. Ishizuka, K. et al. DISC1-dependent switch from progenitor proliferation to J. Neurosci. 28, 12887–12900 (2008). migration in the developing cortex. Nature 473, 92–96 (2011). 14. Baek, S. T. et al. Off-target effect of doublecortin family shRNA on neuronal 45. Guadiana, S. M. et al. Arborization of dendrites by developing neocortical migration associated with endogenous microRNA dysregulation. Neuron 82, neurons is dependent on primary cilia and type 3 adenylyl cyclase. J. Neurosci. 1255–1262 (2014). 33, 2626–2638 (2013). 15. Tada, T. & Sheng, M. Molecular mechanisms of dendritic spine morphogenesis. 46. Kumamoto, N. et al. A role for primary cilia in glutamatergic synaptic Curr Opin Neurobiol 16, 95–101 (2006). integration of adult-born neurons. Nat. Neurosci. 15, 399–405 (2012). 16. Kang, H. J. et al. Spatio-temporal transcriptome of the human brain. Nature 47. Sakakibara, A. et al. Dynamics of centrosome translocation and microtubule 478, 483–489 (2011). organization in neocortical neurons during distinct modes of polarization. 17. Thompson, C. L. et al. A high-resolution spatiotemporal atlas of gene Cereb. Cortex 24, 1301–1310 (2013). expression of the developing mouse brain. Neuron 83, 309–323 (2014). 48. Higginbotham, H. R. & Gleeson, J. G. The centrosome in neuronal 18. Miller, J. A. et al. Transcriptional landscape of the prenatal human brain. development. Trends Neurosci. 30, 276–283 (2007). Nature 508, 199–206 (2014). 49. Richards, L. J., Plachez, C. & Ren, T. Mechanisms regulating the development 19. Oh, E. C. & Katsanis, N. Cilia in vertebrate development and disease. of the corpus callosum and its agenesis in mouse and human. Clin. Genet. 66, Development 139, 443–448 (2012). 276–289 (2004). 20. Miyamoto, T. et al. Insufficiency of BUBR1, a mitotic spindle checkpoint 50. Norris, C. R. & Kalil, K. Guidance of callosal axons by radial glia in the regulator, causes impaired ciliogenesis in vertebrates. Hum. Mol. Genet. 20, developing cerebral cortex. J. Neurosci. 11, 3481–3492 (1991). 2058–2070 (2011). 51. Shu, T. & Richards, L. J. Cortical axon guidance by the glial wedge during the 21. Suijkerbuijk, S. J. E. et al. Molecular causes for BUBR1 dysfunction in the development of the corpus callosum. J. Neurosci. 21, 2749–2758 (2001). human cancer predisposition syndrome mosaic variegated aneuploidy. Cancer 52. Hand, R. & Polleux, F. Neurogenin2 regulates the initial axon guidance of Res. 70, 4891–4900 (2010). cortical pyramidal neurons projecting medially to the corpus callosum. Neural. 22. Szymanska, K. & Johnson, C. A. The transition zone: an essential functional Dev. 6, 30 (2011). compartment of cilia. Cilia 1, 10 (2012). 53. Paul, L. K. et al. Agenesis of the corpus callosum: genetic, developmental and 23. Garcia-Gonzalo, F. R. et al. A transition zone complex regulates mammalian functional aspects of connectivity. Nat. Rev. Neurosci. 8, 287–299 (2007). ciliogenesis and ciliary membrane composition. Nat. Genet. 43, 776–784 (2011). 54. Maurya, A. K. et al. Positive and negative regulation of Gli activity by Kif7 in 24. Valente, E. M. et al. Mutations in TMEM216 perturb ciliogenesis and cause the zebrafish embryo. PLoS Genet. 9, e1003955 (2013). Joubert, Meckel and related syndromes. Nat. Genet. 42, 619–625 (2010). 55. Muresan, V., Lyass, A. & Schnapp, B. J. The kinesin motor KIF3A is a 25. Rix, S., Calmont, A., Scambler, P. J. & Beales, P. L. An Ift80 mouse model of component of the presynaptic ribbon in vertebrate photoreceptors. J. Neurosci. short rib polydactyly syndromes shows defects in hedgehog signalling without 19, 1027–1037 (1999). loss or malformation of cilia. Hum. Mol. Genet. 20, 1306–1314 (2011). 56. Corbit, K. C. et al. Kif3a constrains beta-catenin-dependent Wnt signalling 26. Putoux, A. et al. KIF7 mutations cause fetal hydrolethalus and acrocallosal through dual ciliary and non-ciliary mechanisms. Nat. Cell Biol. 10, 70–76 syndromes. Nat. Genet. 43, 601–606 (2011). (2008). 27. Liem, K. F., He, M., Ocbina, P. J. R. & Anderson, K. V. Mouse Kif7/Costal2 is a 57. Pazour, G. J. et al. The intraflagellar transport protein, IFT88, is essential for cilia-associated protein that regulates Sonic hedgehog signaling. Proc. Natl vertebrate photoreceptor assembly and maintenance. J. Cell Biol. 157, Acad. Sci. USA 106, 13377–13382 (2009). 103–113 (2002). 28. Abdelhamed, Z. A. et al. Variable expressivity of ciliopathy neurological 58. Sedmak, T. & Wolfrum, U. Intraflagellar transport molecules in ciliary and phenotypes that encompass Meckel-Gruber syndrome and Joubert syndrome is nonciliary cells of the retina. J. Cell Biol. 189, 171–186 (2010). caused by complex de-regulated ciliogenesis, Shh and Wnt signalling defects. 59. Abd-El-Barr, M. M. et al. Impaired photoreceptor protein transport and Hum. Mol. Genet. 22, 1358–1372 (2013). synaptic transmission in a mouse model of Bardet-Biedl syndrome. Vision Res. 29. Rooryck, C. et al. Bardet-biedl syndrome and brain abnormalities. 47, 3394–3407 (2007). Neuropediatrics 38, 5–9 (2007). 60. Zhang, Q., Seo, S., Bugge, K., Stone, E. M. & Sheffield, V. C. BBS proteins 30. Lee, J. H. & Gleeson, J. G. The role of primary cilia in neuronal function. interact genetically with the IFT pathway to influence SHH-related phenotypes. Neurobiol. Dis. 38, 167–172 (2010). Hum. Mol. Genet. 21, 1945–1953 (2012). 31. Lee, J. E. & Gleeson, J. G. Cilia in the nervous system: linking cilia function and 61. Lee, J. E. & Gleeson, J. G. A systems-biology approach to understanding the neurodevelopmental disorders. Curr. Opin. Neurol. 24, 98–105 (2011). ciliopathy disorders. Genome Med. 3, 59 (2011).

12 NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8857 ARTICLE

62. Nachury, M. V. et al. A core complex of BBS proteins cooperates with the GTPase K. Stabingas for technical assistance and Dr Tamara Caspary (Emory University) for Rab8 to promote ciliary membrane biogenesis. Cell 129, 1201–1213 (2007). helpful comments. 63. Seo, S. et al. BBS6, BBS10, and BBS12 form a complex with CCT/TRiC family and mediate BBSome assembly. Proc. Natl Acad. Sci. USA 107, Author contributions 1488–1493 (2010). J.G., H.H., J.L. and E.S.A. designed the experiments and supervised the project. H.H., J.G., 64. Badano, J. L. et al. Identification of a novel Bardet-Biedl syndrome protein, J.L., J.N., J.H. and V.G. conducted the experiments and analysed the data. J.G., H.H., J.L. BBS7, that shares structural features with BBS1 and BBS2. Am. J. Hum. Genet. and E.S.A. wrote the manuscript. 72, 650–658 (2003). 65. Billingsley, G. et al. Mutations in chaperonin-like BBS genes are a major contributor to disease development in a multiethnic Bardet-Biedl syndrome Additional information patient population. J. Med. Genet. 47, 453–463 (2010). Supplementary Information accompanies this paper at http://www.nature.com/ 66. Hernandez-Hernandez, V. et al. Bardet-Biedl syndrome proteins control the naturecommunications cilia length through regulation of polymerization. Hum. Mol. Genet. 22, 3858–3868 (2013). Competing financial interests: The authors declare no competing financial interests. 67. Zhang, Q., Yu, D., Seo, S., Stone, E. M. & Sheffield, V. C. Intrinsic protein- protein interaction-mediated and chaperonin-assisted sequential assembly of Reprints and permission information is available online at http://npg.nature.com/ stable bardet-biedl syndrome protein complex, the BBSome. J. Biol. Chem. 287, reprintsandpermissions/ 20625–20635 (2012). 68. Yokota, Y., Ring, C., Cheung, R., Pevny, L. & Anton, E. S. Nap1-regulated How to cite this article: Guo, J. et al. Developmental disruptions underlying neuronal cytoskeletal dynamics is essential for the final differentiation of brain abnormalities in ciliopathies. Nat. Commun. 6:7857 doi: 10.1038/ncomms8857 neurons in cerebral cortex. Neuron 54, 429–445 (2007). (2015). 69. Yokota, Y. et al. The adenomatous polyposis coli protein is an essential regulator of radial glial polarity and construction of the cerebral cortex. Neuron 61, 42–56 (2009). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise Acknowledgements in the credit line; if the material is not included under the Creative Commons license, This research was supported by NIH grants MH060929 to E.S.A. and by the confocal users will need to obtain permission from the license holder to reproduce the material. imaging core of an NINDS institutional center core grant (5P30NS045892-12). We thank To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

NATURE COMMUNICATIONS | 6:7857 | DOI: 10.1038/ncomms8857 | www.nature.com/naturecommunications 13 & 2015 Macmillan Publishers Limited. All rights reserved.