Available online at www.sciencedirect.com

ScienceDirect Neuromuscular Disorders 25 (2015) 461–473 www.elsevier.com/locate/nmd

A critical and previously unsuspected role for doublecortin at the neuromuscular junction in mouse and human F. Bourgeois a,b, J. Messéant a,b, E. Kordeli a,b, J.M. Petit a,c, P. Delers a,b, N. Bahi-Buisson a,d, V. Bernard e,f,g, S.M. Sigoillot a,b, C. Gitiaux a,d, M. Stouffer e,h,i, F. Francis e,h,i, C. Legay a,b,* a

Sorbonne Paris Cité, Université Paris Descartes, F75270 Paris, France b CNRS UMR 8194, F75270,Paris, France c Service commun de Microscopie, F75270 Paris, France d INSERM U1016, Hôpital Necker Enfants Malades, Paris, France e Sorbonne Université, Université Pierre et Marie Curie, F75005 Paris, France f UM CR18, Neurosciences Paris Seine, INSERM, UMR-S 1130, F75005 Paris, France g CNRS, UMR 8246, Neuroscience Paris Seine, F75005 Paris, France h INSERM UMR-S 839, F75005 Paris, France i Institut du Fer à Moulin, F75005 Paris, France Received 10 November 2014; accepted 28 January 2015

Abstract Mutations in the microtubule-associated protein doublecortin (DCX) cause type I (X-linked or XLIS) lissencephaly in hemizygous males and subcortical band heterotopia (SBH) in females, with defects in neuron migration during development affecting cortical lamination. We found that besides its well-established expression in migrating neurons of the brain, doublecortin (Dcx in mice) is also expressed in motor neurons and skeletal muscle in embryonic neuromuscular junctions (NMJs), raising the possibility of a role in synaptogenesis. Studies with whole-mount preparations of embryonic mouse diaphragm revealed that loss of Dcx leads to abnormal presynaptic arborization and a significantly increased incidence of short axonal extensions beyond innervated acetylcholine receptor (AChR) clusters in the developing NMJ. This phenotype, albeit relatively mild, suggests that Dcx contributes to a stop/stabilizing signal at the synapse, which normally limits further axonal growth following establishment of synaptic contact with the postsynaptic element. Importantly, we also identified abnormal and denervated NMJs in a muscle biopsy from a 16-year-old female patient with SBH, showing both profound presynaptic and postsynaptic morphological defects. Overall, these combined results point to a critical role of doublecortin in the formation of the NMJ. © 2015 Elsevier B.V. All rights reserved. Keywords: Neuromuscular junction; Doublecortin; Lissencephaly type I; Synapse formation

1. Introduction Type I lissencephaly corresponds to a spectrum of rare neurodevelopmental disorders associated with brain malformations. Most patients suffer from intellectual disability, intractable epilepsy and hypotonia [1,2]. At the cellular level, this disease is characterized by defects in neuronal migration, leading to a disorganization of the cortical layers, and abnormal neuronal differentiation. Doublecortin (DCX) was originally identified as one of the causative genes for this disorder [3,4]. The DCX gene is located on the human X chromosome and DCX mutations cause X-linked lissencephaly (XLIS) in hemizygous males and subcortical band heterotopia (SBH or

* Corresponding author. Université Paris Descartes, 45 rue des Saints Pères, Paris 75270, France. Tel.: +33 142862068; fax: +33 149279062. E-mail address: [email protected] (C. Legay). http://dx.doi.org/10.1016/j.nmd.2015.01.012 0960-8966/© 2015 Elsevier B.V. All rights reserved.

Double Cortex) in females, with both familial and sporadic forms [5,6]. Males with XLIS are severely disabled, with minimal neurological development and refractory epilepsy [7]. Females with SBH present more heterogeneous and milder clinical features. Seizures in SBH patients appear within the first decade of life, and often evolve to multifocal and refractory epilepsy. In addition to these symptoms, delayed motor development as well as hypotonia can be observed [5,7,8], suggesting deficits in the neuromuscular system. The Dcx gene encodes an intracellular 40 kDa protein that is expressed in migrating postmitotic neurons [3,4]. The protein is present in cell bodies and is enriched in the leading process as well as in differentiating axons and dendrites [9]. Dcx is a microtubule-associated protein (MAP) that stabilizes microtubules and favorizes their polymerization [9–13]. This protein contains two evolutionarily conserved DC domains that occur in tandem in its N-terminus [14], which is unique among

462

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

MAPs and is responsible for the binding to microtubules. Most of the missense mutations identified in patients cluster in these DC domains whereas nonsense mutations are spread throughout the gene, mainly in female patients [5,14,15]. The same DC domain is found in other members of the Dcx family, such as Dcx-like kinases (Dclk1 and 2), and is followed in these proteins by a kinase domain [16,17]. Compared to humans, mice deficient for Dcx show a milder phenotype with an abnormal hippocampal pyramidal cell layer but normal neocortical lamination [18–20]. However, the Dclk1 and Dcx double knockout phenotype shows more severe cortical disorganization, resembling the human phenotype for DCX mutations, suggesting a functional redundancy between the two genes [21,22]. Since Dcx transcripts are present in the spinal cord, and abnormal neuromotor skills have been observed in patients with DCX mutations, we asked if the Dcx/DCX protein could play a role in neuromuscular junction (NMJ) formation. The NMJ is a synapse between a motor neuron and a skeletal muscle fiber that forms in several consecutive steps in mammals. This developmental process starts with the differentiation of a postsynaptic domain in the middle of the muscle, pre-patterning the muscle to attract the motor axon. Axons grow toward this postsynaptic domain and once their target is reached, neurite extension stops. Precise apposition of pre- and postsynaptic elements proceeds while further differentiation occurs in both elements. In a final step, synapse elimination occurs, refining synapse innervation and functioning. A number of key molecules have been described that play an instructive role in NMJ formation, although the critical biochemical pathways have not been fully elucidated. The major players characterized in this process are the neural isoform of agrin, a heparan sulfate proteoglycan secreted by the motor neuron, its transmembrane muscle receptor LRP4, and MuSK, a tyrosine kinase co-receptor that transduces agrin activity and its downstream cascades that lead to AChR clustering [23]. However, the mechanisms that control the pattern of presynaptic arborization and neurite extension are still largely unknown. Here, we show that Dcx is expressed both by the motor neuron and the muscle during development. Analysis of the NMJ phenotype in Dcx knockout (Dcx-/Y) mutant mice during embryogenesis reveals significant defects in presynaptic arborization and neurites bypassing the NMJ, suggesting a defect in a stop signal for neuritic extension. Moreover, in a muscle biopsy from a female patient with double cortex, we show that NMJs are also severely disorganized, giving clinical relevance to this phenotype. We show therefore that DCX/Dcx is not only required for brain development, but also critical for correct NMJ formation. 2. Subject and methods 2.1. Clinical history of the patient The patient is a 19-year-old girl without familial history, carrying a DCX mutation (c.176 G>A p.R59H). She came to medical attention at 6 months with infantile spasms. Neurological

examination was normal. EEG showed atypical hypsarrhythmia with a high voltage activity suggestive of a cortical malformation. Infantile spasms were controlled with vigabatrin, but subsequent focal seizures starting at the age of 7 months were only partially controlled with sodium valproate and lamotrigine. Brain imaging (CT scan and MRI) showed thick SBH, with normal overlying cortex, and ventriculomegaly. Her motor development seemed normal, but she developed intellectual disability (estimated IQ 60), with frontal syndrome (frontal inhibition) and poor social interaction. During adolescence, she also developed scoliosis that required surgical treatment at the age of 15. 2.2. Human biopsies Patient muscle biopsies were obtained following scoliosis surgery from Necker-Enfants malades Hospital (Paris, France) and from Bretonneau Hospital (Tours, France) according to local ethical Institutional Review Boards. Consents of the parents were obtained via protocols approved by the Hospital ethics board committees. Control patients and double cortex patient biopsies were treated in exactly the same conditions. Immediately after sampling, they were fixed with paraformaldehyde (PAF) 4% in PBS at 4 °C overnight, cryoprotected in 30% sucrose-PBS for a few hours and then stored frozen at −80 °C until use for immunofluorescence experiments. 2.3. Mice and cells Dcx knockout mice were generated and characterized previously [24]. Housing conditions and experiments with animals were performed according to the ethical guidelines of French and European legislations (86/809/EEC and 00984.02). We used mutant Dcx-/Y and wild-type (WT) male mice littermates at E14, E16.5 and P8 stages. Embryonic mutant and WT mice from the same litter showed no major differences in weight, size and gross morphology regardless of the stage used. Animals from three different litters were used for each of the stages. The myogenic cell line MLCL was generated as described in [25]. Myoblasts were 50% confluent, and myotubes were analyzed at three different times of differentiation (T1, T2 and T3), which have been characterized previously in [26]. 2.4. RNA preparation, microarray and RT-PCR Total RNA was extracted from muscle cells at T1, T2 and T3 or from embryonic mouse muscle using the RNeasy Mini Kit (Qiagen). For microarray studies, cDNAs generated from 500 ng of RNA were converted into cRNA that were labeled with Digoxigenin-UTP using the Applied Biosystems Chemiluminescent RT-IVT labeling kit and hybridized to a mouse microarray following manufacturer’s instructions (Applied Biosystems). Microarray experiments were performed on 3 replicates from 3 independent MLCL muscle cell cultures at T1, T2 and T3. Primers used for PCR were the following: Dcx F: 5′-TACGTTTCTACCGCAATGGGG-3′, Dcx R: 5′-CT GCTTTCCATCAAGGGTGTA-3′, Dclk1 F: 5′-TGTCGTTC GGCAGAGATATG-3′, Dclk1 R: 5′-TCGAACCTTCTTGGC

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

TTTCT-3′, Dclk2 F: 5′-GGAGCATTGAGCTGGAACAT-3′, Dclk2 R: 5′-GCGGTAGAAGCTGCAGTGA-3′. 18S was used as an internal positive loading control. 2.5. Western blot Equal amounts of proteins (40 µg for cells) were first separated by 8% NuPAGE Novex Tris-Acetate Gels according to a previous protocol [26]. Membranes were incubated with an antibody against DCX (goat antibody from Santa Cruz; 1:500) in PBS with 5% donkey serum, then washed and incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (donkey anti-goat from Nordic Immunology; 1:5000) in PBS with 5% donkey serum. Detection and quantification of the signals were performed as described previously [26]. 2.6. Histology Immunostaining of myotubes was performed after fixation in cold methanol, then as in Sigoillot et al. [26]. Dcx and α-tubulin were revealed with a goat polyclonal antibody (Santa Cruz; 1:250) and a mouse antibody (Sigma; 1:2000), respectively. Alexa Fluor 488- and 594-conjugated secondary antibodies (Molecular Probes) were used at 1:500 dilution. AChR clusters were detected with Alexa Fluor 488- or 594-conjugated α-bungarotoxin (α-BTX; Molecular Probes) used at 1:1000 dilution. Image capture and analysis were performed as in Sigoillot et al. [26]. For immunohistochemistry on tissues, entire diaphragms were dissected from embryos and fixed with 4% PAF in PBS. Whole-mount mouse diaphragms were double stained with polyclonal rabbit antibodies against neurofilament (NF68; 1:1000; Merck Millipore) and synaptophysin (Syn; 1:50; Invitrogen) and Alexa Fluor 594-conjugated α-BTX. Diaphragms were first incubated for 2.5 days at 4 °C with rabbit antibodies to NF and/or Syn or with goat antibody to Dcx (1:250; Santa Cruz) then overnight at 4 °C with Alexa Fluor 594-conjugated α-BTX (1:1000) and rabbit or goat secondary antibodies coupled to Alexa Fluor 488 or 594 (1:500). Immunostaining on patient biopsies was performed on isolated muscle fibers. Antibodies to neurofilaments (NF68) were diluted 1:1000, Alexa Fluor 488- or 594-conjugated α-BTX, and secondary antibodies were diluted 1:750 and 1:500 respectively. 2.7. Image acquisition and quantification of the phenotypes Dcx localization was observed in the diaphragms after whole-mount immunostaining using a LSM 510 Carl Zeiss confocal laser-scanning microscope equipped with a 40× and 63× objectives. At E16.5, the width of the synaptic band was defined as the distance between two AChR clusters localized at the maximal distance from the main axis of the hemidiaphragm using tile scan tools. Images were taken with a Nikon TE2000E Fluorescence microscope equipped using a 10× objective (N.A. = 0,3). The length of eighty 40 µm-spaced segments along the entire synaptic band was measured using Nikon NIS Element AR 3.0 software. Nerve branching was analyzed on the same mosaic pictures using Image J software.

463

For quantitative analysis of AChR clusters (total number, innervation, volume and density), confocal microscopy digitized image stacks were collected from diaphragms at E16.5 using a 40× objective. Images were captured at the point of nerve entry in the left hemidiaphragm. Information was obtained from three z-stacks side by side and data were collected with a 0.5 µm z-step. Hemidiaphragms of four WT (n = 18 digitized image stacks) and eight mutant embryos (n = 27 digitized image stacks) from three different litters were quantified. The volume and density of individual AChR clusters were quantified using the Image J plugin “3D object counter” [27]. Only the objects larger than 25 µm3 and smaller than 400 µm3 in volume were selected and analyzed. Axon terminals innervating and bypassing AChR clusters were checked on rendered 3D images obtained using the LSM 510 software. 2.8. Stastistical analysis Quantitative data are means ± SEM. Statistical analyses were performed by pairwise comparisons between two conditions using a Student’s t-test or an ANOVA test (p < 0.05). 3. Results 3.1. Dcx is expressed in mouse muscle cells in vitro and in vivo and is developmentally regulated We assessed Dcx expression in the previously characterized polyclonal myogenic cell line MLCL, which gives rise to differentiated myotubes (MT) from myoblasts (MB) [26]. Following nearly complete fusion of MB into MT (stage T1), cells express aggregates of classic postsynaptic markers including first AChR (stage T2) then AChR and AChE (stage T3). The 3 time-points of MT differentiation (T1, T2 and T3) were analyzed in microarray experiments. Dcx mRNA was expressed at these 3 time-points with higher levels at T2 compared to T1 and T3 (Fig. 1A). Interestingly, this T2 stage corresponds to early postsynaptic differentiation. Using RT-PCR, we did not detect Dcx mRNAs in MLCL MB in culture, although they were detected in MT at T2. Comparatively, mRNAs coding for other members of the Dcx family, Dclk1 and Dclk2, were found in MB as well as in MT at T2 (Fig. 1B). Dcx, Dclk1 and Dclk2 mRNAs were expressed in vivo in embryonic brain (Br) at E18 and embryonic limb muscles at E14 and E18, but unlike Dclk1 and Dclk2, Dcx mRNAs were absent in adult muscle (Fig. 1B). This temporal expression pattern is therefore similar to the one observed in the brain in which Dcx is present mainly in embryonic but not in mature neurons in adult brains [9], whereas Dclk1 and 2 show wider temporal expression [28]. Using Western blots, antibodies against Dcx detected a doublet at approximately 40 kDa in MT at T2, but not in MB corresponding with the pattern of mRNA expression (Fig. 1C, top panel). In controls, the same antibodies recognized bands of similar molecular weight in extracts of embryonic WT but not Dcx-/Y brains (Fig. 1C, bottom panel), in agreement with previous data [9]. Since Dcx is a MAP, we visualized this protein together with α-tubulin in myotubes at T2 by immunohistochemistry (Fig. 1D). Although fainter in expression, the staining of the

464

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

A

B

In vitro

C

In vivo

-R T M B T2

Muscle

Dcx mRNA relative expression

500

***

**

*

Dcx

Dcx

400

MB T2

Br E14 E18 Ad

GAPDH

300

Dclk1 WT Dcx-/Y

200

Dclk2

100

Dcx

0

T1

D

40 kDa

T2

30 kDa

18S

T3

Dcx

40 kDa

α-tubulin

Fig. 1. The expression of Dcx is developmentally regulated in muscle. (A and B) Transient expression of Dcx during in vitro and in vivo differentiation of myotubes in mice. (A) Quantitative microarray analysis at different stages of differentiation (T1, T2 and T3) shows Dcx mRNA expression in myotubes in culture. (B) RT-PCR analysis of mRNAs from muscle cells in culture and embryonic mouse tissues reveals the presence of Dcx mRNAs in myotubes (T2), brain E18 (Br) and skeletal muscle (hindlimb, E14 and E18), but not in myoblasts (MB) and in adult muscle (Ad, gastrocnemius). mRNAs of the Dcx homologs Dclk1 and Dclk2 are present during all steps of in vitro and in vivo development and persist in adult muscle. As a control, reverse transcriptase was omitted from PCR reactions (-RT). (C) Immunoblot analysis of Dcx in lysates of MLCL cells in culture (upper panel). Dcx is detected in myotubes (T2) but not in myoblasts (MB); GAPDH was used as a loading control. The specificity of the antibody (lower panel) was confirmed by detection of a 40 kDa protein in embryonic brain extracts of wild type (WT) but not mutant mice lacking Dcx (Dcx-/Y). (D) Immunofluorescence localization of Dcx and α-tubulin in T2 myotubes. The subcellular distribution of Dcx partially overlaps with microtubules (arrows). Bars, 10 µm.

two proteins shows that Dcx exhibits a localization pattern that overlaps with the microtubule cytoskeleton. 3.2. Dcx is localized along the nerve trunk, axonal branches and axon terminals of motor neurons Previous data have localized Dcx transcripts in neuronal somata of the spinal cord ventral horn, suggesting that Dcx is expressed in motor neurons [9]. We thus stained diaphragms at E16.5 using whole-mount preparations with antibodies to Dcx to continue to explore the localization of this protein in axons and muscle fibers (Fig. 2). The diaphragm is innervated by the right and left phrenic nerves that enter at the middle of muscle and then develop toward the ventral and dorsal part of the diaphragm. Axons of each phrenic nerve form the nerve trunk that runs along the surface of the muscle, and individual axons exit this bundle along the length of the muscle and form the primary branches. These primary axonal branches then form secondary and tertiary branches that terminate on postsynaptic

areas forming NNJs. The branching pattern of the two nerves is asymmetric, with the right phrenic nerve forming long primary branches, whereas the left nerve forms branches close to the postsynaptic area (see Fig. 3D). To visualize axon terminals, we used antibodies against Dcx, neurofilament, a neuronal cytoskeletal protein to stain axons, and synaptophysin, a component of synaptic vesicles. The use of the latter two antibodies allows the visualization of axons along their entire length including the terminal element. As shown in Fig. 2, Dcx was revealed in the nerve trunk and axonal branches (Fig. 2A). At higher magnification of the synapses, Dcx was observed in the motor axon terminal and co-labeled with synaptophysin (Fig. 2B). A fainter Dcx staining was observed along muscle fibers (Fig. 2C), but no obvious accumulation of Dcx was observed on the muscular postsynaptic side of the NMJ marked by AChR staining (Fig. 2D). Thus, Dcx is mainly and strongly expressed in the motor neuron, including its terminals, aligning with the postsynaptic AChR clusters.

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

A

Dcx

NF

Merge

B

Dcx

Syn

Merge

C

Dcx

AChR

Merge

D

Dcx

AChR

Merge

465

Fig. 2. Dcx is localized at the neuromuscular junction (NMJ) of E16.5 mouse diaphragm. Dcx is present at the presynaptic element of the NMJ, where it partially codistributes with phrenic nerve trunk, branches and nerve terminals as revealed by double labeling for neurofilament (NF; A) and synaptophysin (Syn; B) respectively. Little expression is observed along muscle fibers (arrows; C) with no extensive codistribution with α-BGT-labeled AChRs (C and D). Dcx labeling is mostly adjacent to AChR clusters, consistent with an essentially presynaptic Dcx localization (D). Bars, 20 µm.

3.3. Dcx mutants show wider AChR bands in the embryonic diaphragm To explore the role of Dcx in the motor system, we looked at the phenotype of Dcx-/Y mutant mice [24]. No gross

morphological difference between WT and Dcx-deficient muscles was observed at E18.5 concerning neither the diameter of the muscles, nor the diameter of the myofibers (not shown), nor the ultrastructure of the muscle fiber using electron microscopy (suppl. Fig. S1). The structure of the sarcomeres

466

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

Dcx

AChR

NF/Syn

AChR

Merge

Dcx -/Y

WT

A

Merge

Dcx -/Y

WT

B

C

WT

Dcx -/Y

WT

Dcx -/Y

D

Synaptic band width (µm)

WT Dcx -/Y 200 150

*** *

100

50 0

Left diaphragm

Left

Right

Right diaphragm

Fig. 3. Synaptic phenotype in Dcx-/Y mice in embryonic diaphragm. (A) Localization of Dcx (in green) and AChRs (in red) in whole-mount left diaphragm of wild-type (WT) and mutant Dcx-/Y mouse embryos (E16.5). No Dcx staining is observed in phrenic nerve and muscle fibers in Dcx-/Y mutants. Postsynaptic differentiation is still present in the absence of Dcx, as indicated by the presence of AChRs in both WT and mutant diaphragms. Longitudinal muscle fibers are labeled by anti-Dcx antibody in WT (arrows), but not in mutant Dcx-/Y diaphragm. Bars, 20 µm. (B–D) The synaptic band is larger in Dcx-/Y as compared to WT diaphragms (E16.5). (B) Nerve trunk and axon terminals were labeled with antibodies to neurofilaments and synaptophysin (NF/Syn, green) and postsynaptic AChRs were stained with α-BGT (red). Synaptic areas are delimited by white dotted lines. Bars, 20 µm. (C) Low magnification images of right and left WT and mutant diaphragms labeled for neurofilaments and synaptophysin (green) and AChRs (red). Bars, 0.4 mm. (D) Quantification of synaptic band width on low magnification images shows a significant increase in mutants (ANOVA test; *, p < 0.05 and ***, p < 0.001 for left and right diaphragms, respectively). WT n = 4, Dcx-/Y n = 8.

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473 Table 1 Quantitative analysis of AChR clusters in the left hemidiaphragm of wild type (WT) and Dcx -/Y E16.5 mice.

WT Dcx -/Y

Number of AChR clusters

% of non-innervated AChR clusters

Size of AChR clusters (µm3)

Density of AChR clustersa

73.6 ± 3.57 63.0 ± 2.13*

1.7 ± 0.31 1.8 ± 0.27

98.5 ± 2.82 96.4 ± 3.02

100.0 ± 1.80 102.1 ± 2.50

WT n = 4 and Dcx -/Y n = 8, ANOVA test, * p < 0.05. a In % of WT.

was normal in mutant muscle compared to normal muscle. The analysis of the Dcx-/Y mutant phenotype was performed by quantitative immunohistochemistry of diaphragms. Dcx staining was absent from the nerve and the muscle of mutant embryos (Fig. 3A), confirming the specificity of the Dcx antibody used. At E16.5, clusters of AChR were still present (Fig. 3A and B) indicating that the loss of Dcx did not impair postsynaptic differentiation. However, the distribution of AChR clusters seemed to encompass a larger area of the muscle, as seen in the left diaphragm (Fig. 3B and C). We quantified the width of the synaptic band along its entire length on epifluorescent images of the left and right diaphragms (Fig. 3C and D). In both diaphragms, and especially in the right diaphragm, this band was larger in Dcx-/Y mice compared to WT mice. We also quantified several other parameters characterizing AChR clusters in confocal images such as total number, percentage of innervation, size and density of AChRs within individual clusters (Table 1). For these quantifications, we used the same area in WT and mutant mice where the left phrenic nerve enters the diaphragm. This area matures first as innervation progresses from its entry point in the middle of the diaphragm to both anterior and posterior regions of the muscle. A slight but significant decrease (–14%) in the total number of clusters, as estimated from 3D-rendering confocal images, was found in the mutant. Decrease should concern the innervated clusters, since the number of non-innervated ones was found similar between WT and mutants (Table 1). Size and density of AChR clusters (as measured by the average fluorescence intensity/µm3/cluster) were similar for WT and Dcx-deficient diaphragms. 3.4. Absence of Dcx affects the nerve branching pattern and leads to axonal overgrowth at the NMJ In the brains of Dcx-/Y mutant mice, a defect in branching was observed during migration, with neurons exhibiting shorter neuritic processes but an increased number of branches compared to WT cells [24,29]. To determine if the function of Dcx is conserved in motor neuronal processes, we focused on the branching pattern of the phrenic nerve. The presynaptic arborization was revealed in the right diaphragm by the use of neurite and axon terminal markers (NF and Syn respectively) at two embryonic stages (E14.5 and E16.5) (Fig. 4). At both stages, the diameter of the nerve trunk was similar in WT and mutant mice, suggesting that the number of motor neurons remained unchanged. However, more nerve branches were

467

observed in the Dcx-/Y mutant at the two stages. Indeed, at E14.5, the number of primary branches was not significantly affected but a 75% increase was observed in secondary branches. At E16.5, there was a 42% and 95% increase in primary and secondary branches, respectively, in the mutant compared to the WT. Although increased branching was observed, the number of innervated AChR clusters was slightly lower in Dcx-/Y mutants (see Table 1 and section 3.3). The fact that we did not observe axonal branches not contacting AChR clusters suggests that the number of tertiary branches should be lower in mutant mice compared to WT. However, the number of tertiary branches cannot be accurately quantified. The main observation is that Dcx deficiency increases primary and secondary axonal branches from the nerve trunk, a phenotype potentially related to that observed in migrating neurons in the brain. The terminal axonal branches and their relationship with AChR clusters was further analyzed in WT and Dcx-/Y mutant diaphragms at E16.5, by comparing confocal images. In mutants, neurites innervate AChR clusters as in WT; however, they often extend beyond them (Fig. 5A and C), a phenomenon which only rarely occurs in WT at this age. The number of axons extending beyond AChR clusters was increased by 2.5 fold in the mutant compared to WT (Fig. 5B). At P8, the morphologies of the NMJs were similar in WT and Dcx-/Y mice with no occurrence of neurite overshooting the NMJ (suppl Fig. S2). Thus, inactivation of Dcx in the mouse leads to a motor neuron overshooting phenotype at E16.5, which is likely to be transitory, although the mechanisms leading to the reversibility of this defect currently remain unknown. 3.5. NMJ phenotype in a patient with SBH: a case report Next we investigated the morphology of NMJs from a patient with SBH caused by a DCX mutation (p.R59H falling in the N-DC domain). At last evaluation (17.5 years), neither motor signs nor fatigability were reported by the parents, although no specific exercise tests were performed. Only mild facial hypomimia with excessive drooling was noted, but without ophtalmoplegia or ptosis. Muscular testing was normal as revealed by stimulated single fiber electromyography (StimSFEMG) of orbicularis oculi muscle (data not shown), with normal jitter measurement as per automated analysis by the single fiber program and normal average mean consecutive difference (MCD) (normal ≤32 µs in at least 90% of samples). Data were recorded using concentric needle electrodes at a stimulus frequency of 10 Hz, using a keypoint electromyography as in Tidswell and Pitt [30]. The number of fibers recorded by StimSFEMG was 111. We analyzed a biopsy that was taken from the paravertebral muscles at the level of lumbar vertebrae following scoliosis surgery of this patient at 16 years of age. Control biopsies were taken at the same localization from two 14 to 16-yearold female scoliotic patients with no neurodevelopmental abnormalities and showing the same severity of scoliosis. A predominance of type-1 fibers was observed in the SBH patient biopsy by immunohistochemistry using antibodies

468

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

A

E14.5

WT

Dcx -/Y

II

II

I

I

E16.5

NT

NT

II

I NT

II

NT

B

I

C 20

WT Dcx -/Y

15

NS

25

*

10 5

Number of nerve branches (E16.5)

25

Number of nerve branches (E14.5)

NT

20

WT Dcx -/Y

***

15 10

*

5 0

0 Primary

Secondary

Primary

Secondary

Fig. 4. The nerve branching pattern is affected in Dcx-/Y mouse embryonic diaphragm during synaptogenesis. (A) Low magnification epifluorescence images of the right diaphragm of WT and Dcx-/Y mouse embryos at E14.5 and E16.5 simultaneously labeled by antibodies against neurofilaments and synaptophysin. Mutant mice show more extensive primary (I) and secondary (II) axonal branching as compared to WT. NT, nerve trunk. Bars, 100 µm. (B and C) Quantification of primary and secondary branching using images of the right diaphragms as shown in (A) at E14.5 (B; WT n = 2, Dcx-/Y n = 2) and E16.5 (C; WT n = 4, Dcx-/Y n = 8) shows a significant increase in E16.5 primary branches and in both E14.5 and E16.5 secondary branches (Student’s t-test; *, p < 0.05 and ***, p < 0.001) in mutants (black columns) as compared to WT (white columns).

against fast and slow myosins (not shown), but no other abnormalities in the structure of the tissue were detected. The morphology of five synapses from the biopsies of the two control patients was compared with the eleven synapses present in the SBH patient sample (Fig. 6). Immunostaining

of axons using neurofilament and synaptophysin antibodies and labeling of AChR using fluorescent α-BGT revealed profound presynaptic as well as postsynaptic defects. Whereas in control patients NMJs formed a fork-shaped presynaptic structure with the nerve terminal superimposing with AChR

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

469

Fig. 5. Axon terminals extend beyond AChR clusters in the NMJ of Dcx-/Y mice. (A) Left diaphragms from WT and Dcx-/Y E16.5 mice were labeled with antibodies to neurofilaments and synaptophysin (NF/Syn, in green) and with α-BGT (AChR in red). Bars, 20 µm. (B) Quantification of axonal extensions beyond AChR clusters on images of left diaphragms as shown in (A) (WT n = 4, Dcx-/Y n = 8). A 2.5 fold increase in overshooting extensions is seen in the mutant (ANOVA test; **, p < 0.01). (C) 3D rendering images showing axonal processes (arrowheads) stopping or extending beyond innervated AChRs (asterisks) in WT and mutant NMJ respectively. Bars, 20 µm.

clusters (Fig. 6A), various abnormalities were observed in the SBH patient. AChR clusters were consistently found to be partially or fully denervated, fragmented or abnormal in their shapes (Fig. 6B–D). The nerve either did not fill the entire postsynaptic domain (Fig. 6E and F) or formed extensive loops around postsynaptic domains (Fig. 6G). Although the nerve appeared fragmented in Fig. 6G, confocal images indicated that this was not the case. A number of pre- and postsynaptic elements were not apposed. In summary, none of the observed SBH patient NMJs appeared normal, although all were normal in the controls. Thus, DCX may be critical for NMJ formation and maintenance in humans.

4. Discussion Most of the previous studies on Dcx have focused on its role in neuronal migration and differentiation in the brain. Here, we reveal new roles for Dcx in the motor system and at the NMJ, as revealed by morphological defects in Dcx-/Y mutant mice and importantly in an SBH patient. 4.1. Dcx is expressed in the motor system and is developmentally regulated We show that Dcx is expressed in both muscle cells and motor neurons during development. The presence of Dcx in

470

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

Control biopsy

A

SBH biopsy

B

C

D

E

F

G

AChR NF

Fig. 6. Abnormal NMJs are observed in a muscle-biopsy from a patient with double cortex. Muscle fibers from a 16-year-old patient were stained for AChR (red) and for axon terminals with antibodies to neurofilaments (green). NMJ morphologies were observed by confocal microscopy. (A) Control NMJ (control patient). The axonal branch ends as a fork and innervates the postsynaptic structure. (B–G) Abnormal NMJs from DCX SBH patient muscle biopsy. (B–D) No axon is visualized and fragmented postsynaptic structures are observed. (E,F) The axon branch innervates some but not all of the fragmented postsynaptic structures. (G) The axon terminal envelops the AChR cluster but does not terminate on it. Bars, 10 µm.

motor axons is consistent with the staining of Dcx mRNA previously observed in the spinal cord [9]. Besides its expression in mice, Dcx has also been shown to be expressed in muscle of human fetuses [31], suggesting that Dcx expression in muscle is conserved among mammals. We found that Dcx is only present in postmitotic muscle cells and in embryonic but not in adult muscle. Both in vivo and in vitro, Dcx is distributed along the muscle fibers without any noticeable accumulation at the postsynaptic domain. Its role in these cells, along with its regulation of microtubules, remains to be elucidated. The homologues of Dcx, Dclk1 and Dclk2 are expressed across a wider time scale from myoblasts to adult muscles. Thus, the timing of expression in the motor system for the members of the Dcx family is remarkably similar to the brain in which Dcx is mainly expressed in immature postmitotic neurons [9], whereas Dclk1 and 2 are clearly expressed in neuroprogenitors and persist in mature adult neurons [28,32–34]. Thus, Dcx family members may play varied roles during muscle development and function. The importance of Dcx expression in immature neurons during embryonic brain formation suggested that Dcx might play a role in regulating motor neuron growth and NMJ formation. From the early stages of NMJ formation, Dcx localizes in motor neurons along the axon and also at the terminal synapse. Few previous studies have as clearly revealed

a synaptic localization for Dcx, nor as strongly suggested that Dcx plays a role in synaptogenesis. In fitting with this, at E14.5, the time at which phenotypes are observed in Dcx mutants, it should be noted that most of the motor neuron axons have already reached their muscle targets and hence are no longer growing. 4.2. Consequences of Dcx deficiency on NMJ structure To address the role of Dcx in NMJ formation, we analyzed the phenotype of the NMJ at different embryonic stages in Dcx-/Y mice. As in embryonic mouse mutant brains [18,19], the NMJ phenotype still remains mild and young adult mice present no obvious motor defects (data not shown). However, motor activity was found to be perturbed in a different model of Dcx mutant mice [18]. When comparing WT and heterozygous mutant mice in a mixed (129/SvJ × NIH Black Swiss) genetic background, significant strength defects were observed. Compensation between the members of the Dcx family have been reported in mice [21,22] that could potentially explain the minor impact of Dcx deficiency on brain organization and motor function in the Dcx mouse model studied here, which was bred on the pure C57BL6 background. Taken together, there are nevertheless a number of consistent signs which point to abnormal NMJ development due to DCX/Dcx mutation.

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

Interestingly, the lack of Dcx in the mouse had no major consequence on postsynaptic differentiation, except for an increase in the width of the synaptic band and a slight reduction in the number of innervated AChR clusters. However, AChR clusters were similar in WT and Dcx deficient mice in terms of size and density. They form at the right time and the axon recognized the clusters of AChR that are located in the middle of the muscle fibers. Moreover, almost all of the clusters are innervated even if substantially more axons initially bypassed AChR clusters. Therefore, Dcx is dispensable for the normal localization, timing and clustering of postsynaptic AChRs. These observations imply that Dcx deficiency does not greatly perturb the signaling pathways leading to postsynaptic differentiation. However, in the SBH patient, AChR clusters are noticeably abnormal, confirming a wider pre- and postsynaptic role for DCX at the paravertebral NMJ in humans. Again, the greater severity of the NMJ phenotypes in humans compared to mice is comparable to what is observed in human brain versus mouse brain. Concerning the presynaptic domain in the mouse, striking differences between WT and Dcx mutants were identified. First, excessive axonal branching was observed at different stages, a feature that is consistent with the increased width of the synaptic band. This is reminiscent of previous observations where Dcx-deficient cells migrating from brain explants and in brain slices exhibit more branching than WT cells [22,24]. This phenotype is likely to fit with the main function of Dcx in stabilizing microtubules and therefore limiting neuronal branching. The second characteristic of the phenotype that is specific to the NMJ is the overgrowth of neurites that bypass AChR clusters at E16.5 on a short distance. The fact that all axons reach AChR clusters indicate that Dcx is not involved in the initial axonal guidance and targeting mechanism. However, neurites continue beyond the NMJ. Here, it is important to note that in WT mice, aneural AChR clusters are organized in the middle of the muscle fiber at E14 before they are contacted by the nerve terminal. The growth cone then normally recognizes these clusters and stops growing while apposing to the postsynaptic domain, although some short overshooting is still observed at low frequency until the neonatal period. In the absence of Dcx, the frequency of these short overshooting events is much higher. This is different from other mutant phenotypes in which neurites grow extensively through the muscle after contacting the NMJs. Such phenotypes have been observed in a number of mutants, including for example HSALRP4 or Cdk5 mutant mice [35,36]. Interestingly, Dcx is one of the targets phosphorylated by Cdk5 [37], but this kinase is quite ubiquitous and phosphorylates many proteins, a function that may explain its complex phenotype. The Dcx mutant overshooting phenotype suggests that Dcx is involved in the stop signal machinery, although a delay in the normal process of elimination of bypassing neuritic extensions in the absence of Dcx cannot be excluded. So far, the nature of the stop signal at the NMJ has not been elucidated. Two candidates, agrin and s-laminin, have been proposed to participate in this process. Agrin is a heparan sulfate proteoglycan that is released as several cell-type-specific

471

isoforms by motor neurons, glial and muscle cells. The neuronal isoform is the only active form involved in AChR clustering [38], which binds LRP4, the co-receptor of MuSK [39,40]. However, the evidence supporting agrin as a stop signal mostly relies on in vitro studies. In agrin mutants lacking all forms of agrin or the neuronal form of agrin, axons grow aberrantly, contacting only some of AChR clusters [41,42], hence differing from the Dcx mutant phenotype. The other candidate, S-laminin, is an extracellular matrix protein that is enriched in synaptic basal lamina. Like agrin, it plays an inhibitory role on neurite outgrowth, a function that has been assigned to the tripeptide LRE [43]. However, no aberrant neuritic outgrowth was observed in a mutant lacking s-laminin and most of the AChR clusters were not contacted by nerve terminals [44]. Dcx is hence a new molecule contributing in different ways to stop signal mechanisms. One possibility is that the lack of Dcx impairs the delivery or recycling of a molecule at the synapse that contributes to the stop signal and /or to the retraction of bypassing neurites. In support of this hypothesis, Dcx has been involved in vesicle transport and sorting based on specific partners. These include the µ subunits of clathrin adaptor complexes [45] and two kinases, Cdk5 and Jnk [37,46]. Dcx has also been shown to interact with and regulate endocytosis and surface distribution of the phosphorylated neurofascin adhesion molecule [47,48]. Interestingly, this function of Dcx may not involve its microtubule-binding activity. Furthermore, the lack of Dcx and/or Dclk has been shown to selectively alter the transport of Vamp2, a component of the SNARE complex [21,49]. Analysis of the NMJ phenotype in the biopsy from the SBH patient revealed profound defects in the structure of the synapse and all of the eleven synapses present in the sample were affected. This abnormal morphology probably results from the mutation in DCX since the same paravertebral muscles were normal from age- and gender-matched control patients with scoliosis but without DCX mutations. Indeed, the SBH patient examined exhibits the p.R59H mutation, known to have an effect on microtubule function, as well as apparently perturbing the interaction with phosphorylated neurofascin [50]. In an attempt to correlate the structural defects in the NMJ with the physiology of the muscle, a single fiber EMG was recorded from a facial muscle, the orbicularis oculi, but the EMG was found to be normal. At this point, we cannot make a general conclusion on the muscle activity of this patient since distal and proximal muscles should be tested as well. Indeed, a number of neuromuscular disorders affect selectively specific muscles. However, because of the poor social interaction of the patient, it was not possible to perform the EMG in other muscles. Further opportunities to perform such tests may help clarify this point, particularly studies in male XLIS patients who have more severe brain abnormalities, and potentially more severe neuromuscular phenotypes, as suggested by their profound motor skill deficits [7]. The advantage of the Dcx mouse model is the presence of a relatively simple phenotype where the defects of overshooting and axonal branching represent the major abnormalities. It should thus in future studies be possible to further query the

472

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473

nature of the signals and downstream pathway controlling the termination of motor neuron neurite extension during development. Extrapolating from the mouse mutant data, we propose that motor neuron branching and extension are abnormal in DCX patients and that this defect could be potentially combined with postsynaptic defects. The clinical significance of this altered branching pattern of axon terminals at the NMJ is yet to be determined. Acknowledgements CL thanks AFM, CNRS, Université Paris Descartes and INSERM for financial support. FF is grateful for financial support from the Agence Nationale de la Recherche (ANR-08MNP-013), as well as from INSERM, including the Avenir program, the CNRS and UPMC, the Fondation Bettencourt Schueller, the Conseil Régional, Île-de-France, and the Fondation Jérôme Lejeune. MS is supported by the European Union Seventh Framework Programme FP7/2007–2013 under the project DESIRE (grant agreement no. 602531). We thank Marika Nosten-Bertrand and her colleagues for communicating information on motor tests previously performed in adult Dcx-/Y mouse mutants, and Richard Belvindrah for discussions on this work. We thank Sylvie Thomasseau for producing and genotyping the mice. We thank the IFM animal house staff and the CDTA, Orléans for help with mouse work. The team of FF and contributors from the IFM are associated with the BioPsy Labex project and the Ecole des Neurosciences de Paris Ile-deFrance network. Appendix: Supplementary material Supplementary data to this article can be found online at doi:10.1016/j.nmd.2015.01.012.

Abbreviations AChR BTX Cdk Dcx Dclk HSA LRP4 MAP MuSK NF NMJ PAF PBS SBH SNARE XLIS

acetylcholine receptor bungarotoxin cyclin-dependent kinase doublecortin doublecortin-like kinase human skeletal actin low density lipoprotein receptor-related protein 4 microtubule-associated protein muscle-specific kinase neurofilament neuromuscular junction paraformaldehyde phosphate buffered saline subcortical band heterotopia soluble NSF attachment protein receptor X-linked lissencephaly

References [1] Barkovich AJ, Kuzniecky RI, Jackson GD, Guerrini R, Dobyns WB. Classification system for malformations of cortical development: update. Neurology 2001;57(2001):2168–78. [2] Guerrini R, Marini C. Genetic malformations of cortical development. Exp Brain Res 2006;173:322–33.

[3] Gleeson JG, Allen KM, Fox JW, et al. Doublecortin, a brain-specific gene mutated in human X-linked lissencephaly and double cortex syndrome, encodes a putative signaling protein. Cell 1998;92:63–72. [4] des Portes V, Francis F, Pinard JM, et al. Doublecortin is the major gene causing X-linked subcortical laminar heterotopia (SCLH). Hum Mol Genet 1998;7:1063–70. [5] Bahi-Buisson N, Souville I, Fourniol FJ, et al. New insights into genotypephenotype correlations for the doublecortin-related lissencephaly spectrum. Brain 2013;136:223–44. [6] Quélin C, Saillour Y, Souville I, et al. Mosaic DCX deletion causes subcortical band heterotopia in males. Neurogenetics 2012;13:367–73. [7] Leger P-L, Souville I, Boddaert N, et al. The location of DCX mutations predicts malformation severity in X-linked lissencephaly. Neurogenetics 2008;9:277–85. [8] Barkovich AJ, Guerrini R, Battaglia G, et al. Band heterotopia: correlation of outcome with magnetic resonance imaging parameters. Ann Neurol 1994;36:609–17. [9] Francis F, Koulakoff A, Boucher D, et al. Doublecortin is a developmentally regulated, microtubule-associated protein expressed in migrating and differentiating neurons. Neuron 1999;23:247–56. [10] Gleeson JG, Lin PT, Flanagan LA, Walsh CA. Doublecortin is a microtubule-associated protein and is expressed widely by migrating neurons. Neuron 1999;23:257–71. [11] Horesh D, Sapir T, Francis F, et al. Doublecortin, a stabilizer of microtubules. Hum Mol Genet 1999;8:1599–610. [12] Moores CA, Perderiset M, Francis F, Chelly J, Houdusse A, Milligan RA. Mechanism of microtubule stabilization by doublecortin. Mol Cell 2004;14:833–9. [13] Moores CA, Perderiset M, Kappeler C, et al. Distinct roles of doublecortin modulating the microtubule cytoskeleton. EMBO J 2006;25:4448–57. [14] Taylor KR, Holzer AK, Bazan JF, Walsh CA, Gleeson JG. Patient mutations in doublecortin define a repeated tubulin-binding domain. J Biol Chem 2000;275:34442–50. [15] Sapir T, Horesh D, Caspi M, et al. Doublecortin mutations cluster in evolutionarily conserved functional domains. Hum Mol Genet 2000;9:703–12. [16] Burgess HA, Reiner O. Doublecortin-like kinase is associated with microtubules in neuronal growth cones. Mol Cell Neurosci 2000;16:529–41. [17] Reiner O, Coquelle FM, Peter B, et al. The evolving doublecortin (DCX) superfamily. BMC Genomics 2006;7:188. [18] Corbo JC, Deuel TA, Long JM, et al. Doublecortin is required in mice for lamination of the hippocampus but not the neocortex. J Neurosci 2002;22:7548–57. [19] Kappeler C, Dhenain M, Phan Dinh Tuy F, et al. Magnetic resonance imaging and histological studies of corpus callosal and hippocampal abnormalities linked to doublecortin deficiency. J Comp Neurol 2007;500:239–54. [20] Nosten-Bertrand M, Kappeler C, Dinocourt C, et al. Epilepsy in Dcx knockout mice associated with discrete lamination defects and enhanced excitability in the hippocampus. PLoS ONE 2008;3:e2473. [21] Deuel TAS, Liu JS, Corbo JC, Yoo S-Y, Rorke-Adams LB, Walsh CA. Genetic interactions between doublecortin and doublecortin-like kinase in neuronal migration and axon outgrowth. Neuron 2006;49:41–53. [22] Koizumi H, Tanaka T, Gleeson JG. Doublecortin-like kinase functions with doublecortin to mediate fiber tract decussation and neuronal migration. Neuron 2006;49:55–66. [23] Wu H, Xiong WC, Mei L. To build a synapse: signaling pathways in neuromuscular junction assembly. Development 2010;137:1017–33. [24] Kappeler C, Saillour Y, Baudoin J-P, et al. Branching and nucleokinesis defects in migrating interneurons derived from doublecortin knockout mice. Hum Mol Genet 2006;15:1387–400. [25] Cartaud A, Strochlic L, Guerra M, et al. MuSK is required for anchoring acetylcholinesterase at the neuromuscular junction. J Cell Biol 2004;165:505–15. [26] Sigoillot SM, Bourgeois F, Lambergeon M, Strochlic L, Legay C. ColQ controls postsynaptic differentiation at the neuromuscular junction. J Neurosci 2010;30:13–23.

F. Bourgeois et al. / Neuromuscular Disorders 25 (2015) 461–473 [27] Bolte S, Cordelières FP. A guided tour into subcellular colocalization analysis in light microscopy. J Microsc 2006;224:213–32. [28] Tuy FPD, Saillour Y, Kappeler C, Chelly J, Francis F. Alternative transcripts of Dclk1 and Dclk2 and their expression in doublecortin knockout mice. Dev Neurosci 2008;30:171–86. [29] Koizumi H, Higginbotham H, Poon T, Tanaka T, Brinkman BC, Gleeson JG. Doublecortin maintains bipolar shape and nuclear translocation during migration in the adult forebrain. Nat Neurosci 2006;9:779–86. doi:10.1038/nn1704. [30] Tidswell T, Pitt MC. A new analytical method to diagnose congenital myasthenia with stimulated single-fiber electromyography. Muscle Nerve 2007;35:107–10. [31] Mizuguchi M, Qin J, Yamada M, Ikeda K, Takashima S. High expression of doublecortin and KIAA0369 protein in fetal brain suggests their specific role in neuronal migration. Am J Pathol 1999;155:1713–21. [32] Burgess HA, Martinez S, Reiner O. KIAA0369, doublecortin-like kinase, is expressed during brain development. J Neurosci Res 1999;58:567–75. [33] Ohmae S, Takemoto-Kimura S, Okamura M, et al. Molecular identification and characterization of a family of kinases with homology to Ca2+/calmodulin-dependent protein kinases I/IV. J Biol Chem 2006;281:20427–39. [34] Omori Y, Suzuki M, Ozaki K, et al. Expression and chromosomal localization of KIAA0369, a putative kinase structurally related to Doublecortin. J Hum Genet 1998;43:169–77. [35] Fu AKY, Ip FCF, Fu W-Y, et al. Aberrant motor axon projection, acetylcholine receptor clustering, and neurotransmission in cyclin-dependent kinase 5 null mice. Proc Natl Acad Sci U S A 2005;102:15224–9. doi:10.1073/pnas.0507678102. [36] Wu H, Lu Y, Shen C, et al. Distinct roles of muscle and motoneuron LRP4 in neuromuscular junction formation. Neuron 2012;75:94–107. [37] Tanaka T, Serneo FF, Tseng HC, Kulkarni AB, Tsai LH, Gleeson JG. Cdk5 phosphorylation of doublecortin ser297 regulates its effect on neuronal migration. Neuron 2004;41:215–27. [38] Bezakova G, Ruegg MA. New insights into the roles of agrin. Nat Rev Mol Cell Biol 2003;4:295–308.

473

[39] Kim N, Stiegler AL, Cameron TO, et al. Lrp4 is a receptor for Agrin and forms a complex with MuSK. Cell 2008;135:334–42. [40] Zhang B, Luo S, Wang Q, Suzuki T, Xiong WC, Mei L. LRP4 serves as a coreceptor of agrin. Neuron 2008;60:285–97. [41] Burgess RW, Nguyen QT, Son YJ, Lichtman JW, Sanes JR. Alternatively spliced isoforms of nerve- and muscle-derived agrin: their roles at the neuromuscular junction. Neuron 1999;23:33–44. [42] Gautam M, Noakes PG, Moscoso L, et al. Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 1996;85:525–35. [43] Brandenberger R, Kammerer RA, Engel J, Chiquet M. Native chick laminin-4 containing the beta 2 chain (s-laminin) promotes motor axon growth. J Cell Biol 1996;135:1583–92. [44] Noakes PG, Gautam M, Mudd J, Sanes JR, Merlie JP. Aberrant differentiation of neuromuscular junctions in mice lacking s-laminin/ laminin beta 2. Nature 1995;374:258–62. doi:10.1038/374258a0. [45] Friocourt G, Chafey P, Billuart P, et al. Doublecortin interacts with mu subunits of clathrin adaptor complexes in the developing nervous system. Mol Cell Neurosci 2001;18:307–19. [46] Jin J, Suzuki H, Hirai S-I, Mikoshiba K, Ohshima T. JNK phosphorylates Ser332 of doublecortin and regulates its function in neurite extension and neuronal migration. Dev Neurobiol 2010;70:929–42. [47] Kizhatil K, Wu Y-X, Sen A, Bennett V. A new activity of doublecortin in recognition of the phospho-FIGQY tyrosine in the cytoplasmic domain of neurofascin. J Neurosci 2002;22:7948–58. [48] Yap CC, Vakulenko M, Kruczek K, et al. Doublecortin (DCX) mediates endocytosis of neurofascin independently of microtubule binding. J Neurosci 2012;32:7439–53. [49] Liu JS, Schubert CR, Fu X, et al. Molecular basis for specific regulation of neuronal kinesin-3 motors by doublecortin family proteins. Mol Cell 2012;47:707–21. [50] Reiner O, Coquelle FM. Missense mutations resulting in type 1 lissencephaly. Cell Mol Life Sci 2005;62:425–34.