Wiskott–Aldrich syndrome protein (WASP) –interacting protein (WIP) is an actin-binding protein involved in the regulation of actin polymerization in cells, such as fibroblasts and lymphocytes. Despite its recognized function in non-neuronal cells, the role of WIP in the central nervous system has not been examined previously. We used WIP-deficient mice to examine WIP function both in vivo and in vitro. We report here that WIP−/− hippocampal neurons exhibit enlargement of somas as well as overgrowth of neuritic and dendritic branches that are more evident in early developmental stages. Dendritic arborization and synaptogenesis, which includes generation of postsynaptic dendritic spines, are actin-dependent processes that occur in parallel at later stages. WIP deficiency also increases the amplitude and frequency of miniature excitatory postsynaptic currents, suggesting that WIP−/− neurons have more mature synapses than wild-type neurons. These findings reveal WIP as a previously unreported regulator of neuronal maturation and synaptic activity.
Neuronal cytoarchitecture is first established through neuritogenesis, a process in which neurons extend their neurites to form a functional network during neuronal development (de Curtis 2007). Neuron morphology greatly determines the final complexity of the nervous system and is essential for the signal flow that underlies information integration and processing. It is therefore important that neuritogenesis occurs at the right place and time for correct establishment of synaptic contacts with proper targets (de Curtis 2007). Several environmental cues converge on common coordinated intracellular pathways to modulate neuritogenesis. Such intracellular events involve signaling transduction, exocytic and endocytic mechanisms related to membrane trafficking and cytoskeletal rearrangements.
Neurite initiation and outgrowth are based on the capacity of the neuronal cytoskeleton, constituted mainly of actin microfilaments (MF) and tubulin microtubules (MT), to assemble and disassemble in response to extracellular signals (Luo 2002; Conde and Caceres 2009). The polarized growth of neurites requires the initial depolymerization of actin MF (Bradke and Dotti 1999), stabilization of MT (Ferreira and Caceres 1989), and accumulation of a number of specific proteins (Wiggin et al. 2005). Actin polymerization is controlled by the actin-related protein (Arp2/3) complex and by the action of actin-binding proteins and nucleation-promoting factors (NPF), such as neural Wiskott-Aldrich syndrome protein (N-WASP). The Arp2/3 complex nucleates actin, inducing branching and elongation, and with N-WASP, it mediates neurite elongation (Suetsugu, Hattori, et al. 2002; Pinyol et al. 2007) and neurite branching (Kakimoto et al. 2004). N-WASP interacts with WASP-interacting protein (WIP), a broadly expressed proline-rich protein that regulates N-WASP function as NPF and whose deficiency modifies actin polymerization kinetics and the density of the subcortical actin network (Anton et al. 2007). Through WASP/N-WASP–dependent or –independent mechanisms, WIP participates in a wide variety of cellular functions, including signaling, endocytosis, and actin cytoskeleton remodeling (Anton et al. 2007). WIP deficiency in mice alters the immune response, reducing T and mast cell activity and increasing B cell function (Anton et al. 2002; Kettner et al. 2004). Moreover, WIP null mice have a progressive immunological disorder of autoimmune nature, with evident ulcerative colitis, interstitial pneumonitis, glomerular nephropathy with IgA deposits, autoantibodies, and joint inflammation that lead, all together, to premature death (Curcio et al. 2007). Although molecular details of WIP-WASP/N-WASP interaction have been studied extensively (Volkman et al. 2002; Ho et al. 2004; Dong et al. 2007; Peterson et al. 2007), few data are available on its functional impact and even fewer regarding the central nervous system, where the role of WIP has not been previously addressed.
Using the WIP knockout mouse as a tool, here, we describe that loss of this protein impacts neurite and dendrite dynamics and morphology, both in early and in late developmental stages, in vitro and in vivo. Gross examination of WIP−/− brain revealed changes in forebrain and hippocampal size. Extensive analysis of WIP−/− hippocampal neuron development showed premature neuritogenesis. Finally, electrophysiological and immunocytochemical analyses demonstrated modified synaptic activity of WIP−/− mature neurons. These studies show that WIP is an essential negative regulator in the control of the cytoskeletal events that underlie neuronal and synaptic development.
Materials and Methods
Wild-type (WT) and WIP KO SV129/BL6 mice (Anton et al. 2002) were housed in specific pathogen-free conditions at the animal facility of the Centro de Biología Molecular “Severo Ochoa,” Madrid, Spain. The mouse colony was maintained by continuous mating of heterozygous females with heterozygous males for more than 20 generations. To obtain control or WIP−/− embryos/litters, we mate control male and female or WIP−/− male and female mice. Handling of mice and all manipulations were carried out in accordance with national and European Community guidelines and were reviewed and approved by the institutional committee for animal welfare. All quantification was conducted in a genotype-blind manner.
Brain Lysates and Western Blot
Control or WIP−/− brains were homogenized in lysis buffer (20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, pH 7.4, 100 mM NaCl, 5 mM ethylenediaminetetraacetic acid, 1% Triton X-100, 100 mM NaF, 1 mM Na3VO4, and the Complete Protease Inhibitor Cocktail, Roche Diagnostics), and soluble extracts were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis after determination of protein concentration by Bradford analysis (BioRad). Proteins were then transferred to nitrocellulose filters, which were blocked and incubated with a mouse monoclonal antibody (mAb) specific to WIP (1/1000, 3D10; a generous gift of Prof. R. Geha, Children’s Hospital, Boston, MA). After exposure to a specific secondary antibody, antibody binding was visualized by enhanced chemiluminiscence substrate (Amersham Biosciences).
Primary Hippocampal Cultures
Primary hippocampal cultures were prepared as described (Dotti et al. 1988; Kaech and Banker 2006). Briefly, hippocampi from E18 mouse embryos (WT and WIP−/−) were washed and digested with 0.25% trypsin (15 min, 37 °C). The tissue was then dissociated, resuspended in minimun essential medium with 10% horse serum, and plated on poly-L-lysine–coated coverslips (1 mg/ml) at a density of 6 × 103 cells/cm2 for imaging at early times (up to 24 h after plating) and at a density of 4 × 103 cells/cm2 for electrophysiological recordings. In some experiments, neurons were allowed to adhere to the substrate (1 h) and then incubated with 2 or 5 μM wiskostatin (BIOMOL International) in dimethyl sulfoxide (DMSO). In all cases, after 3 h, plating medium was replaced with neurobasal medium supplemented with B27 (Gibco). For long-term culture, at this time, neuron-including coverslips were transferred into dishes containing an astrocyte monolayer, with neurons oriented facing the glia but without contacting them.
DNA Constructs and Transient Transfection
pLVWIP-GFP was obtained by digestion of pcDNA3WIP-GFP (kindly provided by Prof. N. Ramesh, Children’s Hospital, Boston, MA) and by cloning the insert into pLV; control pLVGFP was generated in a similar manner. Suspended WT (cortical and hippocampal) or WIP−/− (cortical) neurons were nucleofected using a 6 μg DNA/100 μl suspension (Amaxa pulser; Lonza, Germany). Cells were maintained in suspension (4 h) to permit exogenous gene expression and then plated for 24 h before fixation for immunofluorescence analysis.
At 3 h, 1 and 22 DIV (days in vitro) postplating, cells were fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) (pH 7.4; 20 min, room temperature). Cells were then permeablilized with 0.1% Triton X-100 and labeled with phalloidin-tetramethylrhodamine isothiocyanate (Sigma) and antibodies to tyrosinated α-tubulin (1/400; T9028, Sigma), MAP2 (MT-associated protein, 1/400; 514, Sanchez Martin et al. 1998), PSD-95 (5 μg/ml; 75-028, NeuroMab), or GFP (1/100, 11814460; Roche); secondary fluorescent antibody or biotinylated antibody and labeled streptavidin were then added. Cells were excluded from analysis if they showed obvious features of toxicity, such as neurite fragmentation/blebbing or vacuoles in the cell body.
Confocal images were acquired digitally on a confocal LSM510 Meta microscope (Zeiss) coupled to an inverted Axiovert 200 microscope (Zeiss). Image stacks (logical size 1024 × 1024 pixels) consisted of 10 image planes acquired through a 40× (numerical aperture (NA), 1.3) or a 63× oil-immersion lens (NA 1.4).
Time-lapse images of phase-contrast fields were captured on an inverted Axiovert 200 microscope (Zeiss) equipped with a monochrome CCD camera and ultrafast filter change. Image stacks (logical size 1024 × 1024 pixels) consisted of 6 image planes acquired through a 40× oil-immersion lens (NA, 1.3). Metamorph 6.2r6 (Universal Imaging) software was used to process the time-lapse captured images.
Image stacks (physical size 76.9 × 76.9 μm) were imported to the confocal module of Neurolucida 7.1 (MicroBrightfield, Inc., Williston, VT), and neuronal dendritic trees were traced by drawing the dendrites and the bifurcation points. Sholl analysis was performed for each traced neuron by automatically calculating the number of dendritic intersections and the dendritic length at 10-μm interval starting from the soma. Total dendritic length and total number of intersections and branches for each neuron were also calculated as an index of dendritic complexity. Soma area was determined by drawing soma contours while tracing cells.
Miniature Excitatory Postsynaptic Currents
Miniature excitatory postsynaptic currents (mEPSC) were recorded from dissociated hippocampal neurons bathed in artificial cerebrospinal fluid (containing 119 mM NaCl, 2.5 mM KCl, 1 mM NaH2PO4, 11 mM glucose, 26 mM NaHCO3, 2.5 mM CaCl2, and 1.3 mM MgCl2) in the presence of 1 μM tetrodotoxin and 100 μM picrotoxin (at 29 °C). Spontaneous activity was recorded for 3 min for each cell. mEPSC were identified using pClamp software and corrected by eye on the basis of their kinetics.
Morphology and Unbiased Stereology
Three-month-old male mice (WT, n = 6; WIP−/−, n = 6) were anesthetized with pentobarbital (0.04 mg/kg) and perfused transcardially with PBS (20 ml) followed by 100 ml 4% PFA (pH 7.4) in the same buffer. Brains were postfixed in the same fixative (24 h).
Brains were sectioned coronally at a thickness of 50 μm to facilitate measurements. Strict morphological criteria were used in all mice to determine the boundaries of these brain regions. Briefly, hippocampal outlines encompassed the CA1–3 fields of Ammon’s horn and the subiculum but not the presubiculum or fimbria hippocampus. The dentate gyrus was measured separately. Starting with one of the sections, randomly selected across brains, 1 of every 6 sections was analyzed through the extent of a hemisphere of the brain. One hemisphere of each brain was analyzed. Using this sampling strategy, 7–10 histological sections per brain were analyzed. All volumetric quantifications were performed with an Olympus BX51 with a 1.25× objective, a motorized XYZ axis computer-controlled stage (Prior Scientific, Houston, TX), a digital video camera (JVC), and Stereo-Investigator, a stereology software package (version 8.03, MicroBrightField). When calculating hippocampus volume, the boundaries were defined and the volumes determined with Stereo-Investigator software.
Intracellular Injection of Lucifer Yellow
Coronal sections (150 μm) were cut on a vibratome. Sections were prelabeled with 4,6-diamidino-2-phenylindole, and a continuous current used to inject individual cells with lucifer yellow (8% in 0.1 M Tris buffer, pH 7.4). Dye was injected into neurons in the dentate gyrus until the individual dendrites of each cell could be traced to an abrupt end at the distal tips and the dendritic spines were readily visible, indicating that the dendrites were completely filled. The sections were then processed, first with rabbit anti-lucifer yellow (Cajal Institute, 1/400 000 in stock solution of 2% Bovine serum albumin, 1% Triton X-100, 5% sucrose in PBS) and then with Alexa 488-conjugated secondary antibody (Molecular Probes, 1/1000, 4 h). Sections were mounted on a glass slide in fresh ProLong Gold antifade reagent (Invitrogen, Eugene, OR; 24 h, room temperature in the dark) and sealed with nail polish.
Imaging was performed on a Leica laser scanning multispectral confocal microscope (TCS SP5) using an argon laser. Image stacks (physical size 76.9 × 76.9 μm, logical size 1024 × 1024 pixels) consisted of 100–350 image planes acquired through a 63× glycerol-immersion lens (NA, 1.3; working distance, 280 μm; refraction index, 1.45) with a calculated optimal zoom factor of 3.2 and a z-step of 0.14 μm (voxel size, 75.1 × 75.1 × 136.4 nm). These settings and optics represent the highest resolution currently possible with confocal microscopy. For each neuron (5 neurons per mouse, 55 neurons total), 1–5 randomly selected dendrites were scanned from soma to tip, and stacks were processed with a 3D blind deconvolution algorithm (Autodeblur; Autoquant, Media Cybernetics) for 10 iterations to reduce the out-of-focus light.
Quantification was performed in a blind fashion by independent researchers. For statistical analyses, we used the two-tailed Student t-test to compare means, the chi-square test to compare nominal variables, and the two-way analysis of variance to compare Sholl analysis or the Kolmogorov–Smirnov test to compare cumulative frequency analysis. Results are shown as mean ± standard error of the mean.
WIP Is Expressed in the WT Mouse Brain
The expression pattern of WIP messenger RNA (mRNA) suggests that WIP is ubiquitously expressed in all mouse tissues, including brain (Ramesh et al. 1997; Tsuboi 2006). To date, however, the levels of endogenous WIP protein have only been analyzed in fibroblasts, myoblasts, and hematopoietic-derived cells (reviewed in Anton et al. 2007). To study WIP expression in brain, we probed western blots of cell lysates from 3 regions of the adult mouse brain (cortex, hippocampus, and olfactory bulb), using the WIP-specific mAb 3D10 (Koduru et al. 2007). This mAb recognized 2 protein bands in cell lysates from each of the regions examined (Fig. 1A), which could represent 2 WIP isoforms and/or be the result of posttranslational modifications of WIP. Neither band was present in WIP-deficient tissue. These data demonstrate WIP expression in murine brain and corroborate previous mRNA studies (Ramesh et al. 1997; Tsuboi 2006) as well as the data included in the Allen atlas (http://www.brain-map.org/).
Brain Hypertrophy in WIP-Deficient Mice
To assess general effects of WIP deficiency on the mouse brain, we determined volumes of the hippocampus and of the rest of the murine forebrain (3-month-old male mice). Contours of the structures of interest were drawn on Nissl-stained serial sections with the aid of Stereo-Investigator software to yield the volume of each. There was a significant increase (30%) in forebrain volume in WIP-deficient brains (n = 5) compared with WT littermates (n = 6, Fig. 1B–D and Supplementary Fig. 1). The hippocampus proper (without the dentate gyrus) and the dentate gyrus itself were also hypertrophied in WIP−/− brain (Fig. 1E,F). WIP deficiency thus causes general macrocephaly that includes hippocampal hypertrophy.
Enhanced Early Neuronal Development in WIP −/− Neurons
The enlarged brain observed in WIP−/− mice relative to their WT littermates raised the possibility that the greater brain volume could be due to increased neuritic branching since dendrites and axon collaterals account for most of the brain volume (Acebes and Ferrús 2000). To test this hypothesis at the cellular level, we examined the effect of WIP deficiency on early neuronal development. Primary hippocampal neurons from control and WIP−/− embryos were grown at very low density so that their neurite arbors did not overlap to avoid confusion about which neurites protruded from each cell body. Cells were labeled with anti-tyrosinated α-tubulin and fluorescent phalloidin (Fig. 2A,B). As a measure of neuronal development, we quantified the fraction of neurons at each developmental stage at 3 h postplating. At stage 1, the MT-containing soma is surrounded by flattened actin-rich lamellipodia, while in stage 2, the lamellipodia are transformed into equidistant neurites (Dotti et al. 1988). Stage 1+ is defined as cells with incipient tubulin-rich small protrusions. We found a clear difference in the distribution of these populations (Fig. 2C); there was a significant decrease (35%) in the frequency of WIP−/− neurons classified as developmental stage 1 compared with WT neurons (n = 4 cultures and n = 150 neurons/genotype, Fig. 2C). Conversely, there was an increase (45%) in the percentage of WIP-deficient neurons at stage 2. As another measure of neuronal development, we quantified the number of neurites arising from each neuron at 3 h postplating and found a significant increase (36%) in the average number of neurites arising from WIP−/− (n = 172) compared with WT neurons (n = 151, Fig. 2D). This finding was confirmed in time-lapse images made with phase-contrast microscopy for 3 h after plating (Fig. 2E and Supplementary Videos 1 and 2). These data suggest that, at 3 h after plating, WIP−/− neurons are in a more advanced developmental stage than WT neurons. This implies enhanced early neuronal development in the absence of WIP.
N-WASP Activity and Fine-Tuned WIP Levels Regulate Neuronal Development
To test whether this accelerated neuronal development continues at later developmental times, we quantified the fraction of neurons that remains at developmental stage 1 at 24 h postplating. Cultured hippocampal neurons were fixed and labeled with anti-tyrosinated α-tubulin antibody and fluorescent phalloidin and imaged by confocal microscopy. As observed at 3 h postplating, we found a decrease of 48% in the frequency of WIP−/− neurons at developmental stage 1 (Fig. 3A left and Fig. 3B), implying that WIP−/− neurons still show accelerated development at this time.
To determine whether this phenotype depends on the activity of N-WASP, an actin NPF with WIP-binding capacity, we incubated cultures (23 h) with wiskostatin, a selective reversible N-WASP inhibitor (Peterson et al. 2004). Wiskostatin is a cell-permeable N-alkylated carbazole derivative that selectively blocks actin filament assembly by binding to N-WASP, stabilizing the autoinhibited conformation, and preventing activation of the Arp2/3 complex (Wegner et al. 2008). Wiskostatin produced a dose-dependent increase in the fraction of stage 1 neurons, indicating impairment in neuronal development after blockade of N-WASP-dependent filament assembly (Fig. 3A,B) (da Silva and Dotti 2002; Dent et al. 2007). Inhibition of N-WASP activity with 2 μM wiskostatin induced a phenotype in WIP−/− neurons resembling that of untreated control neurons, and with 5 μM wiskostatin, WIP−/− neurons no longer showed a lower frequency of stage 1 neurons (Fig. 3A,B). This finding suggests that N-WASP activity is essential in mediating the accelerated development of WIP−/− neurons.
To further evaluate the role of WIP in neurite sprouting, WT neurons were nucleofected with lentiviral constructs for overexpression of WIP-GFP (pLVWIP-GFP) or GFP as control (pLVGFP; Fig. 3C). At 24 h postplating, while 31% of GFP-expressing neurons were in stage 1, the frequency of stage 1 WIP-GFP–expressing neurons was 80% (Fig. 3D). The developmental delay at 24 h induced by WIP overexpression was confirmed in cortical embryonic neurons (Supplementary Fig. 2). This effect was sustained over time, since 48 h after plating, 22.6% of neurons with increased WIP levels remained at stage 1 compared with none of the GFP-expressing neurons.
We used rescue experiments to confirm the WIP contribution to neuron development. WIP−/− neurons were nucleofected with a lentiviral-based vector coding for WIP-GFP or control GFP, and 24 h after plating, the fraction of cells in stage 1 or 2 was quantified (Fig. 3E,F; WIP-GFP–expressing WIP−/− neurons). The population distribution of rescued neurons was indistinguishable from that of WT neurons and differed significantly from that of GFP-expressing WIP−/− neurons. WIP-GFP expression in WIP−/− neurons also reversed the phenotype, as the mean number of primary neurites (Fig. 3G and Supplementary Fig. 3) and neuritic bifurcations (Supplementary Fig. 3) were similar to those of GFP-positive WT neurons.
These findings indicate that WIP acts as a negative regulator of neuronal differentiation and that neuronal development can be modulated bidirectionally by WIP levels.
Enhanced Dendritic Maturation in WIP −/− Neurons
To test whether the WIP regulatory role persists in differentiated neurons, we also imaged cultured hippocampal neurons at 22 DIV (Fig. 4A). For quantitative analysis of the branching pattern of the neuritic or dendritic tree at the distinct neuronal developmental stages, we traced neurons (1 and 22 DIV) with Neurolucida software. Sholl analysis was used to measure the number of neurites/dendrites crossing circles at various radial distances from the soma (Sholl 1953). Both neuronal types showed addition of dendritic branches over time (Fig. 4B,D), although the number of crossings was significantly higher in WIP−/− neurons compared with WT neurons at both time points. This finding implies that WIP−/− neurons show higher ramification of their neuritic (1 DIV) and dendritic arbor (22 DIV, Fig. 4D). Nevertheless, the difference in the number of crossings between phenotypes, which was 82% at 1 DIV, decreased to 45% at 22 DIV (Supplementary Fig. 4A).
The increased ramification was reflected in greater total neuritic or dendritic length of WIP−/− neurons at both time points (Fig. 4C,E), a difference that was also reduced in mature neurons (from 109% to 40%; Supplementary Fig. 4B). This difference was even smaller in neurons of adult mice (see below).
Soma size determines the number of dendrites and associated branches in neurons (Kernell and Zwaagstra 1989; Kollins and Davenport 2005). To analyze whether WIP deficiency alters soma area, we quantified this parameter in WT and WIP−/− neurons. Similar to our observations for the dendritic arbor, soma area was significantly larger in WIP−/− neurons (Fig. 4F,G), but this difference attenuated from 40% to 27% with time (Supplementary Fig. 4C). The effect of WIP deficiency on promoting neuritic or dendritic arborization and on soma size is thus greater in early developmental stages.
Enhanced Neuronal Ramification in Adult WIP −/− Mice
To test whether WIP deficiency produces a long-lasting effect on neuronal maturation in vivo, we evaluated dendritic architecture in adult (3-month-old) mice after injection of lucifer yellow into hippocampal neurons of fixed brain sections (Fig. 5A) and studied the structure of granule cells in the dentate gyrus. Sholl analysis showed significantly more crossings in the distal part of the dendrite in WIP−/− mice (n = 59 neurons from 6 mice) compared with WT mice (n = 39 neurons from 5 mice, Fig. 5B). This increased number of crossings reflects a larger number of ramifications at the outer molecular layer of the dentate gyrus. As a result of the higher dendritic ramification, the dendritic length in this layer is also increased in WIP−/− mice (Fig. 5C), leading to greater total dendritic length (18%) in WIP−/− mice (Fig. 5D). Cumulative frequency analysis indicated that WIP−/− mice have a subpopulation of highly ramified neurons not detected in WT mice (Fig. 5E). WIP deficiency thus leads to enhanced dendritic ramification in the adult mouse. Nonetheless, the magnitude of change in total dendritic length of the adult WIP−/− mouse is smaller than in early developmental stages.
Enhanced Synaptic Maturation in WIP −/− Neurons
During later stages of neuronal development, dendritic arborization and synaptogenesis occur in parallel (Cantallops et al. 2000; Cline 2001). The actin cytoskeleton has a pivotal role in both processes. In the context of synaptogenesis, actin dynamics control the morphogenesis and function of the dendritic spines, small actin-rich protrusions from dendritic shafts (Ethell and Pasquale 2005). As we found that WIP modulates dendritic arborization, we examined the possibility that WIP deficiency fosters synaptic maturation. We measured the effects of WIP deficiency on spontaneous miniature (action-potential independent) postsynaptic currents (mEPSC) in dissociated hippocampal neurons (22 DIV) in the presence of 1 μM tetrodotoxin (Fig. 6A). Both the amplitude and the frequency of mEPSC were significantly increased in WIP−/− compared with WT neurons (n = 17 and 15 neurons, respectively; Fig. 6B,C). mEPSC amplitude is indicative of the postsynaptic strength of individual functional synapses, whereas mEPSC frequency depends on synapse number and presynaptic properties (Turrigiano et al. 1998; Han and Stevens 2009). WIP deficiency functionally increases both the strength and possibly the number of individual synapses, suggesting that at this time, WIP−/− neurons show either a more mature phenotype or more abundant and/or enlarged dendritic spines.
To examine whether WIP deficiency resulted in an increased number of structural synapses, we stained dissociated hippocampal neurons at the same developmental stage (22 DIV) for the postsynaptic marker PSD-95, known to drive maturation of glutamatergic synapses (El-Husseini et al. 2000). WIP−/− neurons showed increased fluorescence intensity of individual PSD-95 puncta, indicating greater PSD-95 accumulation in spines in the absence of WIP (WT, n = 117 spines from 6 neurons; WIP−/−, n = 168 spines from 7 neurons, Fig. 6D,E). Frequency analysis showed that in WIP−/− neurons, staining of some PSD-95–positive spines is particularly intense, whereas such a population was not found in WT neurons (Fig. 6F). Positive spine density was nonetheless similar in WIP−/− and WT neurons (Fig. 6G). As WIP−/− neurons are substantially more branched than WT neurons (and their total dendritic length is therefore increased), the total number of PSD-95–positive spines is significantly increased in WIP−/− neurons (Fig. 6H). These neuronal cultures were maintained over WT astrocytes, excluding possible astrocyte involvement in the synaptic modulation derived from the lack of WIP. These results indicate that WIP deficiency increases the number and size of structural and functional synapses.
Here, we present the first report of a role for WIP in the brain, showing a specific role for WIP as a negative regulator of murine neuronal development. We found that, in hippocampal primary neurons, loss of WIP accelerates the onset of neuritogenesis and increases neuritic and dendritic branching. Moreover, lack of WIP enhances PSD-95 accumulation in hippocampal dendritic spines and increases neuronal synaptic activity. Conversely, WIP overexpression blocks the initiation of neuritogenesis, suggesting that WIP is a bidirectional regulator of neuronal development.
WIP Modulates Neuritic Branching During Neuronal Development
Marked dendritic elaboration normally begins in neuronal cultures 2–3 days after axonal outgrowth (Nowakowski and Rakic 1979; Dotti et al. 1988). Our experiments showed that immature WIP−/− neurons begin to emit neurites starting at the first day in culture, much earlier than control neurons. After cell cycle exit and before neuronal polarization, cortical postmitotic neurons make a transition through a multipolar stage, when multiple neurites emerge rapidly from the cell body (Barnes and Polleux 2009). This morphological transition is also controlled by WIP during brain development, as GFP in utero electroporation shows a notable increase in the percentage of multipolar cells in WIP−/− embryos, in parallel to a reduction in the fraction of round and unipolar neurons (not shown). In addition, neurites in immature WIP−/− neurons are more branched than neurites of WT neurons. Dendritic branching is also increased in WIP−/− mature neurons, although to a lesser extent than at early neuronal developmental stages (Fig. 4 and Supplementary Fig. 4). Sholl analysis demonstrated that the initial morphological features acquired by immature WIP−/− neurons determine the pattern of dendritic branching of the more differentiated cells. We therefore suggest that WIP acts as a negative modulator that controls the precise timing of the correct onset of dendritic development. Our results point to the possibility that some pathologies detected in mature neurons might have their origin, yet unknown, in altered processes that occur during early neuronal development.
WIP versus Other Negative Regulators of Dendritic Branching
Numerous molecules are involved in the positive control of neuritic or dendritic outgrowth and branching, many of them related to the control of cytoskeleton dynamics (McAllister 2000; Urbanska et al. 2008). Only a minority of these molecules, such as PTEN (Kwon et al. 2001; Jaworski et al. 2005) and RhoA (Negishi and Katoh 2002; da Silva et al. 2003), have been identified as negative regulators of these processes. Similarly to our findings, conditional adult Pten mutant mice show enlargement of cortex and hippocampus associated with dendritic hypertrophy and increased soma size (Kwon et al. 2001, 2006). At difference from the WIP−/− phenotype observed here, PTEN is not necessary for initiation of neuritogenesis (Lachyankar et al. 2000). The most striking difference between WIP- and PTEN-mediated negative regulation thus appears to depend on the temporal window of neuronal development in which both proteins operate. PTEN does not control early stages of neuritogenesis, and its influence on neuronal structure increases progressively throughout the animal’s life. In contrast, WIP modulates neuronal development and synaptogenesis early in postnatal life, and its effects are attenuated as neuronal maturation progresses.
Similarly to the absence of WIP, RhoA inhibition increases both the number of primary neurites and total neuritic length in immature dissociated hippocampal neurons, whereas expression of a constitutively active form of RhoA arrests cells in stage 1 (da Silva et al. 2003). In mature neurons, transfection of a dominant-negative RhoA construct does not affect dendritic morphology (Nakayama et al. 2000), whereas WIP−/− neurons still show enhanced dendritic arborization. The effect of WIP on neuronal maturation thus appears to be longer lasting than that of other negative modulators of neuritic and dendritic outgrowth and branching.
Possible Mechanisms for WIP-Mediated Neurite Sprouting and Branching
Sprouting of primary neurites and interstitial branching from neuritic or dendritic shafts follows the same sequence as cortical cytoskeletal rearrangements (Wu et al. 1999; Dent et al. 2007). For the extension of a single filopodium, which is subsequently stabilized by MT invasion, F-actin and MT must undergo local depolymerization triggered by extracellular signals (Acebes and Ferrús 2000; da Silva and Dotti 2002; Luo 2002). WIP inhibits F-actin depolymerization (Martinez-Quiles et al. 2001), which could explain the neuritogenesis arrest found in WIP-overexpressing neurons.
Our data indicate that WIP modulates neuritogenesis through its previously described ability to maintain N-WASP in its autoinhibited state (Martinez-Quiles et al. 2001). N-WASP promotes neurite outgrowth and regulates neurite branching in hippocampal neurons (Suetsugu, Hattori, et al. 2002; Abe et al. 2003; Pinyol et al. 2007). At the molecular level, N-WASP acts as a signal integration device that can precisely target actin polymerization to membrane sites at which PI(4,5)P2 and activated Src kinases and Cdc42 are located (Prehoda et al. 2000; Suetsugu, Miki, et al. 2002). Cdc42 stimulates the actin-polymerizing activity of N-WASP, creating free barbed ends from which actin polymerization can take place (Miki, Suetsugu, et al. 1998), leading to filopodium formation (Miki, Sasaki, et al. 1998). In the absence of WIP, N-WASP would be more easily released from inhibition and be hyperactivated to initiate premature filopodium formation. This interpretation is supported by pharmacological inhibition of N-WASP by wiskostatin, which blocked the accelerated neuritogenesis seen in WIP-deficient neurons (Fig. 3).
WIP Modulates Synaptic Activity
We report here that mEPSC amplitude and frequency are increased in WIP−/− neurons, suggesting both stronger AMPA (alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid) receptor–mediated synaptic currents and more functional synapses in these cells (Fig. 6). In addition, we describe an increase in PSD-95 accumulation at individual spines in WIP−/− neurons. These 2 events are probably linked since PSD-95 drives the maturation of excitatory synapses with the incorporation of AMPA receptors (El-Husseini et al. 2000). Indeed, the contribution of AMPA receptors to synaptic transmission increases gradually with neuronal maturation (Mammen et al. 1997; Petralia et al. 1999). The increase in PSD-95 accumulation at individual WIP−/− spines might suggest that the PSD (and therefore the synapse) is enlarged in the absence of WIP. Given that the synaptic area correlates positively with the number of synaptic AMPA receptors (Nusser et al. 1998), our electrophysiological findings are strengthened by the morphological analysis of PSD-95 in spines.
In conclusion, this study sheds light on the molecular events that control early neuronal development by presenting WIP as a previously undescribed regulator that prevents premature dendritic and synaptic maturation.
Grants from Consejo Superior de Investigaciones Científicas-Comunidad de Madrid (CCG08-CSIC/SAL-3471), CSIC (PIE200720I002), and the Spanish Ministry of Education and Science (BFU2007-64144 and BFU2010-21374) to I.M.A., from Centro de Investigación Biomédica en Red Enfermedades Neurodegenerativas (Instituto de Salud Carlos III), the Plan Nacional DGCYT (SAF2009-12249-C02-01) to F.W. and SAF2010-15676 to S.K., and by an institutional grant from the Fundación Ramón Areces. A.F. was a recipient of an FPU MEC fellowship (AP2005-3405), I.B. held a contract from the Comunidad Autónoma de Madrid and S.K., a Ramón y Cajal contract.
We thank Chiara Ragazzini and Sonia Pérez for their excellent technical assistance and Javier de Felipe for his contribution to the morphometric analysis with NeuroLucida Software. We are grateful to Raif Geha and Narayanaswamy Ramesh for the 3D10 mAb and to Lola Ledesma and Carlos Dotti for helpful advice on the manuscript. We acknowledge Daniel Gallego for electrophysiological signal processing, José Ramón Valverde for statistical assistance and Catherine Mark for editorial assistance. Conflict of Interest : None declared.