J. Neurochem. (2012) 121, 903–914.
Proper development of neuronal networks relies on the polarization of the neurons, thus the establishment of two compartments, axons and dendrites, whose formation depends on cytoskeletal rearrangements. Rnd proteins are regulators of actin organization and they are important players in several aspects of brain development as neurite formation, axon guidance and neuron migration. We have recently demonstrated that mice lacking RhoE/Rnd3 expression die shortly after birth and have neuromotor impairment and neuromuscular alterations, indicating an abnormal development of the nervous system. In this study, we have further investigated the specific role played by RhoE in several aspects of neuronal development by using hippocampal neuron cultures. Our findings show that neurons from a mice lacking RhoE expression exhibit a decrease in the number and the total length of the neurites. We also show that RhoE-deficient neurons display a reduction in axon outgrowth and a delay in the process of neuronal polarization. In addition, our results suggest an involvement of the RHOA/ROCK/LIMK/COFILIN signaling pathway in the neuronal alterations induced by the lack of RhoE. These findings support our previous report revealing the important role of RhoE in the normal development of the nervous system and may provide novel therapeutic targets in neurodegenerative disorders.
rho kinase-I (p160ROCK)
- SMI 31
Monoclonal antibody to neurofilaments, phosphorylated epitope
Neurons are highly polarized cells and typically develop a single axon and several highly branched dendrites. The establishment of a polarized morphology and the specialization of axonal and dendritic compartments are essential steps in the differentiation of neurons (Schwamborn et al. 2006). To study neuronal polarization in vitro, cultures of hippocampal neurons have been extensively used. As neurons develop, they form several neurites, but at some point they begin to polarize so that one neurite becomes an axon whereas the remaining neurites become dendrites (Arimura and Kaibuchi 2007). This asymmetric neurite formation is regulated by molecules that control cytoskeletal rearrangement and protein trafficking.
Rho proteins act as key transducers of extracellular signals to the actin cytoskeleton (Jaffe and Hall 2005). There are six atypical members: Rnd1, Rnd2,Rnd3, RhoH, RhoBTB1 and RhoBTB2 (Riou et al. 2010). RhoE/Rnd3 is a member of the Rnd subfamily that acts antagonistically to RhoA (Riento et al. 2005) and induces stress fiber disassembly (Guasch et al. 1998; Nobes et al. 1998). RhoE binds and inhibits ROCK-I from phosphorylating its downstream target, myosin light-chain phosphatase (Riento et al. 2003), and RhoE itself is regulated by phosphorylation by ROCK-I on multiple sites (Riento et al. 2005). ROCK can also activate LIM kinase (LIMK), a kinase involved in the inactivation of Cofilin, which results in the reorganization of the actin cytoskeleton. LIMK has been demonstrated to regulate neuronal morphology and neuritogenesis (Beckloff et al. 2007).
In the nervous system, Rho family of GTPases play major roles in several aspects of neuronal development, including neurite outgrowth and differentiation, axon pathfinding, and dendritic spine formation and maintenance (Govek et al. 2005). Rho activation causes the collapse of the growth cone and neurite retraction, and activation of Rac and Cdc42 induces the formation of lamellipodia and filopodia of the growth cone (Negishi and Katoh 2002), regulating dendritic branch dynamics and growth. Other Rho GTPases, such as Rnd1 and 2, play a role in neurite formation and retraction. Rnd1 is involved in dendritic spine formation (Ishikawa et al. 2003) and Rnd2 promotes dendrite branching in PC12 cells (Fujita et al. 2002). Rnd2 and RhoE have been recently shown to be involved in cortical neuronal migration (Heng et al. 2008; Pacary et al. 2011). Rnd proteins can also be involved in axon guidance through their interaction with plexins (Puschel 2007). Our previous results with RhoE demonstrated that this protein is involved in neuritogenesis by inducing neurite outgrowth in PC12 cells (Talens-Visconti et al. 2010).
We have recently shown that mice lacking RhoE expression have neuromotor impairment and neuromuscular alterations, indicating an abnormal development of the nervous system (Mocholi et al. 2011). In this work, we have examined the role of RhoE in several aspects of neuronal development. Our results demonstrate that RhoE deficiency leads to a decrease in both neurite and axon outgrowth and also to a delay in the process of neuronal polarization. In addition, we have found that the RHOA/ROCK/LIMK/COFILIN signaling pathway is involved in those alterations.
Material and methods
Wild-type (+/+) and RhoE gene-trap (gt/gt) mice were described previously (Mocholi et al. 2011). RhoE +/gt mice were generated at Lexicon Pharmaceuticals by using a gene-trapping method, based on a tagged random mutagenesis. Wild-type and RhoE gt/gt animals were derived by breeding the heterozygous mice on a hybrid genetic background (129SvEvBrd-C57Bl/6J). Experiments were performed by studying the progeny resulting from the heterozygous mice crossing. As no significant differences were detected between male and female mice of each genotype, results were combined. Animal care and work protocols were approved and carried out according to the regulations set forth by the European Communities Council Directive (86/609/ECC) and by the Spanish Royal Decree 1201/2005. The animal experiments were also approved by the Ethical Committee of Animal Experimentation of the Prince Felipe Research Center (Valencia, Spain). Same sex littermates were group-housed (four animals per cage) under a 12-h light/dark schedule in controlled environmental conditions of humidity and temperature with food and water supplied ad libitum.
Hippocampal neurons were prepared from P0 mouse brains. Each hippocampus was treated individually to avoid mixing genotypes. Briefly, mice were decapitated and the skin and skull were removed. The two brain hemispheres were separated and the hippocampus was dissected from the ventral part of the cortex. Cell suspensions were prepared by trypsin treatment of hippocampi and then trituration using two fire-polished Pasteur pipettes of different diameters. Neurons were plated onto poly-l-lysine coated glass coverslips and maintained with minimum essential medium plus 10% foetal bovine serum, 2% glucose, 1% penicillin–streptomycin and 1% glutamine for 2–3 h and cultured under humified conditions in 95% air and 5% CO2 at 37°C. Once cells attach to the substrate (2–3 h) the medium was replaced with neurobasal medium, 2% B-27 suplemment, 0,1% penicillin–streptomycin, 1% glutamine, and 25 μM glutamate. Half of the medium was change twice a week. Culture media, foetal bovine serum and other supplements were obtained from Gibco-Invitrogen (Paisley, UK).
The change in cell number was determined by counting the number of cells that excluded trypan blue. Neurons from both genotypes were trypsinized and incubated in the presence of 0.4% trypan blue solution in phosphate-buffered saline (Ca2+ and Mg2+-free), and the number of viable neurons was quantified. Cell counts were performed every 2 days. All counts represent the mean of five different experiments.
Hippocampi were washed with phosphate-buffered saline and lysed for 30 min on ice in cold lysis buffer 1% NP40, 20 mM Tris–HCl, pH 8, 130 mM NaCl, 10 mM NaF, 10 μg/ml aprotinin, 40 μM leupeptin, 1 mM dithiothreitol, 1 mM Na3VO4 and 10 mM phenylmethylsulfonyl fluoride). Total cell lysates were clarified by centrifugation at 16 000 g for 15 min at 4°C. Next, equal amounts of proteins were resolved on sodium dodecyl sulfate–polyacrylamide gel electrophoresis 12% gels, transferred to polyvinylidene fluoride membranes (Immobilon-P; Millipore, Billerica, MA, USA) and blocked for 1 h with non-fat milk 5% in Tris-buffered saline containing 0.1% Tween 20 to prevent non-specific binding. After blocking, blots were incubated for 1 h at 22°C or at 4°C overnight with the primary antibodies diluted in blocking solution. The following antibodies were used: anti-Cdc42 (1 : 250; Santa Cruz Biotechnologies, Santa Cruz, CA, USA), anti-RhoE (1 : 500; Upstate, Lake Placed, NY, USA), anti-ROCK-I (1 : 100; G-6, Santa Cruz, sc-17794), anti-ROCK-II (1 : 2000; BD Biosciences, Lexington, KY, USA), anti p190-B RhoGAP (1 : 1000; Cell Signalling), anti-pLIMK 1 (1 : 500; Abcam, Cambridge, UK), anti-pLIMK 2 (1 : 500; Abcam), anti-LIMK 1 (1 : 100; BD Biosciences), anti-LIMK 2 (1 : 100; Santa Cruz Biotechnologies), anti-pCofilin (1 : 1000; Santa Cruz Biotechnologies), anti-Cofilin (1 : 1000; BD Biosciences). After three washes with Tris-buffered saline containing 0.1% Tween 20, blots were incubated for 1 h at 22°C with the following horseradish peroxidase conjugated anti-IgG antibodies: anti-mouse (1 : 1000; Santa Cruz Biotechnologies) for RhoE, ROCK-I, ROCK-II, LIMK-1 and Cofilin, anti-rabbit (1 : 20 000; Sigma-Aldrich, St. Louis, MO, USA) for Cdc42, p190-B RhoGAP, pLIMK-1, pLIMK-2 and pCofilin, and anti-goat (1 : 20 000; Sigma-Aldrich) for LIMK-2. Blots were developed using the enhanced chemiluminescent system ECL Plus (GE Healthcare Life Sciences, Buckinghamshire, UK).
RhoA activity assay
The Rho activation assay kit (Upstate) was used according to the manufacturer’s instructions. Hippocampal lysates were incubated with 25 μg of Rho Assay Reagent (Rhotekin Rho-Binding Domain, agarose). Proteins retained on the beads were resolved on 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels. Bound Rho proteins were detected by western blot using the anti-RhoA antibody (1 : 500; Santa Cruz Biotechnologies). The amount of GTP-bound protein was normalized relative to the total amount of RhoA in the lysates.
Hippocampal neurons were seeded on polylysine-coated glass coverslips (15 mm) in 12-well plates and fixed and permeabilised as previously described (Guasch et al. 2003). Anti-microtubule-associate protein 2 (MAP-2) (1 : 200; Sigma-Aldrich) was used as a marker for neurons followed by incubation with Cy3-conjugated donkey anti-mouse IgG (1 : 200; Invitrogen). Monoclonal antibody to Neurofilaments, Phosphorylated epitope SMI 31 (1 : 500; Covance, NJ, USA) was used as axon marker followed by incubation with Oregon Green 488-conjugated anti-mouse IgG (1 : 200; Invitrogen).
Neurite outgrowth analysis
For quantification of various parameters of neurite outgrowth, fluorescent images from a Zeiss microscope (Axioskop 2) were taken, and morphological measurements were performed using the plugin NeuronJ from the ImageJ software. Neurites from 3 day neurons were stained with MAP-2 and were counted to obtain the number of neurites per neuron. Total length of neurites is defined as the sum of the neurite length per cell. Axons that were positive for the SMI-31R marker were analyzed by using the ImageJ software. At least 100 cells were counted in each experiment, and the experiments were repeated 4–6 times, therefore, no less than 400 cells were analyzed in each experiment.
Time-lapse imaging for analysis of neurite growth rates was performed in neurons cultured in 35-mm-coated dishes. After 24 h of culture, images were acquired every 10 min and the number of elongations and retractions were counted during a 5-h period. Cells were filmed on a Leica BL-TLB 058 (Leica Microsystems Heidelberg GmbH, Mannheim, Germany) microscope with an incubator at 37°C, 5% CO2.
Animals were sacrificed by decapitation (P0), or overdose of pentobarbital (P15). Brains were extracted and post-fixed overnight in 4% paraformaldehyde, dehydrated in increasing concentrations of ethanol, embedded in paraffin, serially sectioned (5 μm) in a HM 310 Microm microtome and collected on polylysine-coated slides. Antigen retrieval of deparaffined and rehydrated sections was performed by heating at 100°C in a water-bath for 15 min in citrate buffer (10 mM pH 8). Then, sections were washed three times in 0.1 M phosphate buffer with 0.2% Triton X-100 and incubated with 3% H2O2 in methanol for 40 min to quench endogenous peroxidase activity. Non-specific binding was blocked with 10% normal horse serum in 3% bovine serum albumin (BSA). Immunohistochemistry was performed using the immunoperoxidase procedure of Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA, USA). Briefly, sections were incubated with the primary antibody anti-MAP-2 (1 : 100; Sigma-Aldrich) overnight at 4°C in a humidified chamber. Then they were incubated 1 h with a secondary anti-mouse biotinyladed antibody, amplified with the avidin-peroxidase complex and finally revealed by diaminobenzidine tetrahydroclorhide stain. Sections were analyzed with an Olympus BX40 microscope.
Data are shown as the mean ± standard error of the mean. Student’s t-test for unpaired variables was used to test for differences elicited by the lack of RhoE expression and differences were considered statistically significant at p < 0.05.
Hippocampal neurons lacking RhoE expression exhibit a decrease in neurite outgrowth
As we have recently shown that RhoE deficiency induces neurodevelopmental alterations (Mocholi et al. 2011), we asked for the specific role of this protein in neuronal polarization. For this purpose, and given that RhoE is expressed in newborn hippocampus (Ballester-Lurbe et al. 2009), we first examined the architecture of the hippocampus in mice lacking RhoE expression. These animals were generated by a gene-trap method. The presence of the gene-trap (gt) allele was confirmed by PCR and by the lack of RhoE protein expression in RhoE gt/gt mice was confirmed by western blot (Mocholi et al. 2011). Histological analysis of the hippocampus did not show major differences between wild-type and RhoE gt/gt mice, as revealed by cresyl violet staining (Fig. 1a). The architecture of the dentate gyrus and the CA fields were similar in RhoE gt/gt and wild-type mice, both in the newborn and at postnatal day 15. MAP-2 immunostaining did not show noticeable differences in the distribution of the neurites between both genotypes. These results suggested that RhoE had not a major effect on the cell layer formation, but whether this protein has a role on the processes formation remained unknown. We then analyzed the neuritogenesis process in primary cultures of hippocampal neurons from mice lacking RhoE expression.
As illustrated in Fig. 1b, neurons that do not express RhoE (gt/gt) and heterozygous (+/gt) exhibited a significant decrease in the number of neurites (26% reduction in gt/gt neurons and 14% in +/gt neurons). This reduction in neuritogenesis was even more pronounced when the total length of the neurites was compared, with about 50% reduction in RhoE gt/gt neurons respect to the wild-type ones. We also observed that RhoE deficiency induced less branched dendritic arbors. Indeed, quantification of the experiments (Fig. 1c) showed a decrease in the number of primary (18% reduction in gt/gt neurons), secondary (55% reduction in gt/gt neurons and 44% in +/gt neurons) and tertiary processes (50% reduction in gt/gt neurons and 25% in +/gt neurons). There were no major alterations between wt and mutant neurons when actin and tubulin were examined by microscopy (Fig. 1c). These results indicate that the lack of RhoE decreases processes branching.
To further investigate the effect of RhoE deficiency on neurite growth rate we performed time-lapse imaging (see movies as supporting information). Neurons were filmed over a 5-h period starting 24 h after plating. Our observations revealed that whereas neurites of the wild-type neurons were very active, extension and retraction of the gt/gt neurites were slower. Quantification of the results showed a significant decrease in the extension rate and to a lesser degree in the process retraction (Fig. 1d). These results suggest that RhoE deficiency may mainly affect the process of neurite elongation.
RhoE-deficient mice show a delay in neuronal polarization
Given the effect of RhoE on neurite formation we decided to investigate whether the absence of RhoE could affect the process of neuronal polarization. For this purpose, we first characterized the development of neuronal polarity in our culture conditions. Based on Dotti’s description (Dotti et al. 1988) neuronal stages were established considering that crucial steps in neuronal differentiation were taking place (Fig. 2a). Shortly after plating (5 h) round spheres attached to the substrate and spread a lamellipodium around the cell body (stage 1). At 20 h, several undifferentiated processes were formed (stage 2) and at about 30 h one of these neurites was selected to become the axon and extended rapidly (stage 3). On day 3, neurites started to elongate and differentiate into dendrites (stage 4) followed by the maturation of axonal and dendritic processes (7 days, stage 5). The number of neurons at each stage (5 h, 20 h, 30 h, 72 h and 7 days) were counted and illustrated in Fig. 2b.
Neuronal polarization was also studied in RhoE gt/gt neurons and compared with +/+ cells. At stage 1, no significant differences were observed between the number of RhoE gt/gt neurons and the control ones (Fig. 2c). However, whereas most of the RhoE +/+ neurons reach stage 2 by 20 h, a shift in the RhoE gt/gt curve was observed (Fig. 2d), indicating that it takes longer time (30 h) for the majority of RhoE gt/gt neurons to develop to the stage 2. These results suggest that it takes longer time for RhoE-deficient neurons to extend their neurites.
Looking at Fig. 2d and e we can observe that at 30 h most of control neurons are in developmental stage 3 whereas the majority of RhoE gt/gt neurons are at stage 2. These observations suggest that it takes longer time for RhoE-deficient neurons to break the neurite symmetry and acquire the polarized morphology, which takes place at 72 h, as illustrated in Fig. 2e. At stages 4 and 5, we can observe that the number of mature RhoE gt/gt neurons was lower when compared with control neurons (Fig. 2f and g).
To further investigate the effect of RhoE deficiency on neuron maturation we also analyzed the sequence of morphological changes in heterozygous neurons (Fig. 2c–g). A delay in heterozygous neuronal polarization was also detected, which is significantly different from 30 h of culture. These differences between RhoE +/gt and +/+ neurons were found at stages 3, 4 and 5, from which morphological changes leading to neuronal polarity takes place. Analysis of the data showed that about 75% of both RhoE +/gt and gt/gt populations were at stages 1 and 2 at 30 h, a time where the axon has not yet been formed. At this time, about 50% of the RhoE +/+ population was already at stage 3 and, therefore, starting already the polarization event. At 72 h control neurons are mainly in stage 4 whereas the majority of heterozygous neurons were at stage 3. These results suggest that RhoE may have an important role in the process of axonal establishment because even heterozygous neurons displayed a developmental delay during that cellular event.
We also analyzed the expression of RhoE along the process of polarization in control hippocampal neurons. Our results showed an important increase in RhoE levels at 30 h after plating (Fig. 2h), a time where we observed a delay in the maturation of RhoE-deficient neurons. At longer time points (72 h and 7 days) we also detected high RhoE levels in wt neurons compared with 20 h. The increase in RhoE levels at 30 h, 72 h and 7 days support our observations related to a delay in polarization in neurons lacking RhoE expression.
We also observed that this delay in neuronal maturation lasts for longer time, as illustrated in Fig. 2i. However, in our hands is not possible to properly measure the length of dendrites after day 7 because of the complexity of the dendritic arborization. To further study whether these alterations in neuronal maturation affected cell survival, neurons from both genotypes were cultured and cells were counted at different time points. Our results showed no significant differences in the number of viable neurons between wt and mutant genotypes, at least during 10 days of culture (Fig. 2j).
RhoE deficiency reduces axon outgrowth
Given the effect of RhoE deficiency on processes formation we decided to investigate whether the lack of RhoE expression could have an effect on axon outgrowth. For that purpose, we first quantified both the longest and the shortest process in hippocampal neurons from the three genotypes. As illustrated in Fig. 3a, whereas no differences in the shortest process were observed, statistically significant differences were found in the longest process between control and RhoE-deficient neurons (both in RhoE +/gt and gt/gt). Furthermore, shorter axons were observed not only at 3 days in culture but also at a time as long as 9 days, suggesting that RhoE could have an effect on axon formation (Fig. 3a).
However, as morphology should not be used as the only criteria to study axon formation (Arimura and Kaibuchi 2007) we decided to use a specific molecular marker for axons as SMI-31R. The length of the SMI-positive processes was quantified and Fig. 3a shows that axons in RhoE-deficient neurons were shorter than in wild-type cells. Our results also show that 78% of control neurons polarized and formed a positive process for the axonal marker SMI-31R, whereas only 66% of RhoE gt/gt neurons generated an axon (Fig. 3b). These observations indicate that there is a lower number of RhoE gt/gt neurons bearing an axon and also that those processes are shorter than in control neurons. Overall, these results support the conclusion that RhoE deficiency decreases axon outgrowth.
The RHOA/ROCK/LIMK/COFILIN pathway participates in the neuronal alterations induced by the lack of RhoE expression
As we observed that RhoE deficiency induced a delay in neuronal maturation, affecting both neurite and axon outgrowth, we aimed to study the signaling pathway involved in those alterations. Considering that RhoE antagonizes RhoA (Chardin 2006) we first analyzed the activation of RhoA in the hippocampus by using a pull-down assay. Our results showed an increase in RhoA activation in RhoE-deficient mice (Fig. 4a), in agreement with our previous reports (Guasch et al. 2007; Talens-Visconti et al. 2010). Furthermore, mutant mice displayed an important decrease in the p190GAP levels, an exchange factor by which RhoE antagonizes RhoA (Wennerberg et al. 2003) (Fig. 4a), supporting the involvement of RhoA in the observed neuronal alterations. We also examined whether other Rho GTPases were also implicated, as previous studies indicate with Cdc42 (Talens-Visconti et al. 2010). Mice lacking RhoE expression showed an increase in the levels of Cdc42, as detected by western blot analysis (Fig. 4b).
As RhoE inhibits ROCK-I (Riento et al. 2003) and ROCK-I inhibition is required for RhoE-induced neurite outgrowth in PC12 cells (Talens-Visconti et al. 2010), we would expect an activation of ROCK-I in RhoE-deficient neurons. Indeed, mutant mice displayed an increase in the levels of ROCK-I, as revealed by western blot analysis of hippocampus lysates (Fig. 4c). To further confirm this hypothesis, hippocampal neurons were treated with the ROCK inhibitor Y-27632, and we investigated whether the RhoE-induced alterations on neuronal polarization could be rescued. Our results show that the ROCK inhibitor was able to rescue the effect of the absence of RhoE on neurites. Thus, RhoE gt/gt neurons treated with the ROCK inhibitor showed similar length and numbers of neurites as +/+ neurons (Fig. 4d). We also observed a recovery of the longest process length when Y-27632 was used. Wild-type neurons were also treated with the ROCK inhibitor and a significant increase was found in the total length of the neurites (Fig. 4d).
As ROCK-II plays a critical role in the regulation of neuronal actin and spine morphology (Zhou et al. 2009) we wondered whether this ROCK isoform was involved in the neuronal alterations induced by the lack of RhoE. Western blot analysis showed that mutant mice exhibited a decrease in the levels of ROCK-II (Fig. 4c), suggesting a compensatory effect between both isoforms. All of these results suggest that ROCK is involved in the alterations of both neurite and axon formation induced by the lack of RhoE protein expression.
It is known that ROCK is involved in neurite retraction by phosphorylating and activating LIMK, which in turn inactivates Cofilin, an actin depolymerising factor (Govek et al. 2005). Therefore, we investigated whether this signaling pathway participates in the alterations seen in RhoE gt/gt neurons. However, as it has been previously demonstrated a close relationship between LIMK-1 and LIMK-2 (Sumi et al. 2001) we decided to study the participation of each specific isoform. We examined the phosphorylated levels of LIMK-1 and LIMK-2 by western blot analysis. Because of the small amount of protein obtained from neurons we decided to use lysates from hippocampus at postnatal day 15 (PD15), a time where we have previously observed the highest RhoE expression in the mouse CNS (Ballester-Lurbe et al. 2009). Our results show an important increase in the phosphorylation levels of both isoforms LIMK-1 and LIMK-2 in RhoE gt/gt hippocampus indicating an activation of LIMK1/2 (Fig. 4e). To further study the involvement of this signaling pathway we also examined the possible alterations on Cofilin levels. We found an increase in the phosphorylated level of Cofilin, corresponding to the inactive form (Fig. 4f). These results were obtained at PD15 as well as at PD5. Altogether these results indicate that the RHOA/ROCK/LIMK/COFILIN signaling pathway is involved in the neuronal polarization alterations induced by the absence of RhoE.
Rho proteins are a family of signaling molecules involved in regulating cytoskeletal dynamics (Jaffe and Hall 2005). Rnd proteins are atypical Rho members, highly expressed in the brain, which regulate several aspects of neuronal function (Riou et al. 2010). It has been shown that Rnd1 is involved in dendritic spine formation (Ishikawa et al. 2003) and Rnd2 is essential for brain development (Heng et al. 2008). Rnd proteins can also play a role on axon guidance through their interaction with plexins (Puschel 2007). We have recently demonstrated that RhoE is widely expressed in the central nervous system, especially in the early postnatal period (Ballester-Lurbe et al. 2009) and that mice lacking RhoE expression have neuromotor impairment and neuromuscular alterations, indicating an abnormal development of the nervous system (Mocholi et al. 2011). To get more insight into the specific role played by RhoE in neuronal development we have used cultured hippocampal neurons from mice lacking RhoE expression. Here, we show that RhoE-deficient neurons exhibit a decrease in both neurite and axon outgrowth and also a delay in the process of neuronal polarization. In addition, we have found that the RHOA/ROCK/LIMK/COFILIN signaling pathway participates in those alterations.
It is well known that rearrangements of the actin cytoskeleton and microtubules are crucial for the initial establishment of polarity (Tahirovic and Bradke 2009). Therefore, it is not surprising that Rho proteins, the main regulators of the cellular cytoskeleton, are implicated in several aspects of neuronal development. We have observed that the absence of RhoE in mice induces not only a decrease in the number of neurites but also in the total length of these processes. These effects were also observed in heterozygous animals. In addition, we have also noticed that the deficiency in the expression of RhoE protein (in RhoE gt/gt and +/gt neurons) induces a reduction in the dendritic branching, further supporting the involvement of RhoE in neurite outgrowth. In this regard, we have not observed major alterations when actin and tubulin from mutant neurons were compared with wild types. However, we believe that further studies will be necessary to investigate in more detail some structures of the actin cytoskeleton as the growth cone, and spines, where it is likely to find differences between wt and RhoE-deficient neurons.
These results are not surprising considering that RhoA has a negative role in neurite formation (Govek et al. 2005) and that RhoE acts antagonistically to RhoA (Chardin 2006). These findings are also supported by our previous finding in PC12 cells where we demonstrated that RhoE over-expression induced neurite-like outgrowth (Talens-Visconti et al. 2010). Moreover, our time-lapse imaging data suggest that RhoE may mostly participate in the process of neurite elongation rather than in the retraction. In this regard, a recent report shows that RhoE/Rnd3 promotes neuronal migration and that Rnd3-silenced neurons display an increase in the number of processes (Pacary et al. 2011). The apparent discrepancy between both studies could be explained by the different approaches used. Pacary et al. used an acute loss-of-function approach (RNA interference) and we have used a chronic method (genetically engineered mice lacking RhoE expression). We think that both situations could have different consequences and therefore, other pathways could be involved. In agreement with this interpretation, we have found a decrease in the levels of ROCK-II that may reflect a compensatory effect to the increased ROCK-I activity observed in mutant mice.
We have observed that RhoE is involved in neuronal polarization because it takes longer time for RhoE-deficient neurons to develop in comparison to control neurons. Moreover, our results suggest a gene-dosage effect because heterozygous neurons also exhibit a delay in neuronal maturation, although it seems a milder effect. These developmental differences between control and heterozygous neurons were observed at stages 3, 4 and 5, which are the initiation and progression of the polarization events. These findings suggest that RhoE could have an important role in the establishment of axon formation.
To further support the abovementioned hypothesis we have investigated the axon outgrowth in RhoE gt/gt neurons. Our findings reveal that the absence of RhoE expression reduces the axon outgrowth in cultured hippocampal neurons. Furthermore, as we demonstrate by morphological and molecular criteria, the lack of RhoE expression not only induces a shortening of axon length but also a decrease in the number of neurons with axon.
Because of the intrinsic limitations of a cell culture we cannot exclude the possibility that RhoE-deficient neurons would finally reach the same developmental stage as control neurons. However, our recent studies with RhoE null mice have demonstrated that these animals die shortly after birth and they also show severe alterations in the nervous system (Mocholi et al. 2011). Besides, our observations at longer culture times indicate a severe developmental delay in RhoE gt/gt neurons. Therefore, our results strongly support our previous hypothesis that RhoE is important for neuronal development.
It has been previously demonstrated that RhoA has a negative role in neurite formation (Govek et al. 2005) and our findings show a RhoA activation in the hippocampus of mice lacking RhoE expression. Moreover, mutant mice also display a decrease in the levels of p190GAP, a molecule by which Rnd proteins function as antagonists of RhoA (Wennerberg et al. 2003). These findings are in agreement with our previous results where RhoA inactivation was involved in the RhoE-induced neurite outgrowth in PC12 cells (Talens-Visconti et al. 2010). Altogether, these results suggest that RhoA is implicated in the neuronal alterations induced by the lack of RhoE expression. In addition, and in line with our previous report (Talens-Visconti et al. 2010), Cdc42 seems to be implicated in the observed neuronal alterations, as suggested by the increased protein levels seen in RhoE mutant mice.
RhoE controls the actin cytoskeleton through ROCK-I inhibition (Riento et al. 2003), a kinase involved in Rho-mediated neurite retraction (Amano et al. 1998; Hirose et al. 1998). Therefore, we would expect a ROCK-I activation in RhoE gt/gt neurons that would lead to a reduction in neurite outgrowth. Our results show an increase in the levels of ROCK-I together with a decrease in the case of ROCK-II. Given that the biological functions of both isoforms are in many cases redundant (Thumkeo et al. 2005), these findings suggest a compensatory effect between these two isoforms. Importantly, our findings demonstrate a rescue of the neurite shortening induced by the absence of RhoE when RhoE gt/gt neurons were treated with the ROCK inhibitor Y-27632. Moreover, ROCK inhibition not only induces an increase in the number of neurites but also in the length of both the neurites and the axon. These findings indicate that ROCK is involved in the shortening of the neurites and the axon induced by the lack of RhoE. In agreement with these findings, we have recently demonstrated that the neurite outgrowth induced by RhoE over-expression in PC12 cells leads to ROCK-I inhibition (Talens-Visconti et al. 2010). Our data are also in agreement with other studies where ROCK inhibition rescued the stimulation of neurite elongation in hippocampal neurons (Da Silva et al. 2003). Importantly, there is an increasing interest in using ROCK inhibition as a strategy to induce neurite outgrowth and stem cell neuronal differentiation (Dottori et al. 2008; Pacary et al. 2008). In addition, ROCK inhibition prolongs the lifespan and improves the maturation of neuromuscular synapses of a mouse model of spinal muscular atrophy (Bowerman et al. 2010) suggesting that molecules in the RhoA-ROCK pathway can serve as therapeutic targets in diseases of the nervous system.
It is known that RhoE binds to ROCK-I and activates myosin light-chain phosphatase (Riento et al. 2003). However, we could not detect any alteration in its phosphorylation level (data not shown), suggesting that this phosphatase does not participate in the neuronal alterations observed in the absence of RhoE. Therefore, as ROCK-I can also phosphorylate LIMK which in turn inactivates Cofilin, an actin severing protein involved in actin depolymerization, we decided to investigate this pathway. Our findings show that the absence of RhoE induces an increase in LIMK phosphorylation and therefore in its activation.
There is increasing evidence showing that each LIMK member may be subject to distinct regulatory pathways and may contribute to both different and overlapping cellular and developmental functions (Scott and Olson 2007). For example, LIMK-1 is recruited to the tips of neurites and thus involved in neuritogenesis (Wen et al. 2007). On the other hand, LIMK-2 binds Par-3 (Chen and Macara 2006) a polarity protein which plays critical roles in axon specification. Therefore, we were also interested in studying the effect of RhoE deficiency on each LIMK member. Our findings demonstrated an increase in both LIMK-1 and LIMK-2 phosphorylation levels in RhoE-deficient mice. Therefore, our results suggest that RhoE can interfere with the regulation of both LIMK-1 and LIMK-2 proteins which would finally contribute to the observed alterations in neuronal polarization.
To take a further step in the signaling pathway, we studied the possible alterations of Cofilin and observed that mice lacking RhoE show an increase in phosphorylated Cofilin, which in turn inactivates the protein. These findings agree with previous studies where increased activity of Cofilin causes an increase in neurite extension (Meberg 2000). Moreover, our previous studies demonstrated that RhoE over-expression induced a decrease in the phosphorylated Cofilin in PC12 cells. Taking into account all these findings, we can conclude that the RHOA/ROCK/LIMK/COFILIN signaling pathway is involved in the neuronal alterations induced by the absence of RhoE.
It has been demonstrated that the use of dominant-negative and constitutively active mutants of Rho proteins can lead to non-specific or nonphysiological effects. Therefore, it is very important to study the in vivo function of Rho proteins by using genetically modified mice. In this study, we have used hippocampal neurons from mice lacking RhoE expression. Our studies show that RhoE protein plays an important role in neuronal development because RhoE deficiency induces a decrease in both neurite and axon outgrowth and also a delay in the process of neuronal polarization. Importantly, we find that the RHOA/ROCK/LIMK/COFILIN signaling pathway is involved in the alterations induced by the lack of RhoE expression and its modulation may be very valuable to address neurodegenerative therapy.
This research was supported by grants from the Instituto de Salud Carlos III (PI10/01686, FEDER co-funding), Universidad CEU Cardenal Herrera (PRUCH and Santander-Copernicus), MCYT (SAF 2006-02178), Consellería de Educación de la Generalitat Valenciana (GVPROMETEO-2009/011) and Red RETICS (RD06/0001/0019 and RD06/0010/0022). BP is a fellow from the Conselleria de Sanidad.
Conflict of interest
The authors declare no conflict of interest.