Vitronectin promotes the progress of the initial differentiation stage in cerebellar granule cells

Molecular and Cellular Neuroscience 70 (2016) 76–85

Contents lists available at ScienceDirect

Molecular and Cellular Neuroscience journal homepage: www.elsevier.com/locate/ymcne

Vitronectin promotes the progress of the initial differentiation stage in cerebellar granule cells Kei Hashimoto, Fumi Sakane, Natsumi Ikeda, Ayumi Akiyama, Miyaka Sugahara, Yasunori Miyamoto ⁎ Graduate School of Humanities and Sciences, Ochanomizu University, Otsuka 2-1-1, Bunkyo-ku, Tokyo 112-8610, Japan

a r t i c l e

i n f o

Article history: Received 17 June 2015 Revised 16 November 2015 Accepted 27 November 2015 Available online 27 November 2015 Keywords: Vitronectin Development of the cerebellum Cerebellar granule cell Differentiation Migration

a b s t r a c t Vitronectin (VN), which is an extracellular matrix protein, is known to be involved in the proliferation and differentiation of primary cultured cerebellar granule cell precursors (CGCPs); however, the effect of VN is not fully understood. In this study, we analyzed the effects of VN loss on the proliferation and differentiation of CGCPs in VN knockout (VNKO) mice in vivo. First, immunohistochemistry showed that VN was distributed in the region from the inner external granule layer (iEGL) through the internal granule layer (IGL) in wild-type (WT) mice. Next, we observed the formation of the cerebellar cortex using sagittal sections of VNKO mice at postnatal days (P) 5, 8 and 11. Loss of VN suppressed the ratio of NeuN, a neuronal differentiation marker, to positive cerebellar granule cells (CGCs) in the external granule layer (EGL) and the ratio of CGCs in the IGL at P8, indicating that the loss of VN suppresses the differentiation into CGCs. However, the loss of VN did not significantly affect the proliferation of CGCPs. Next, the effect of VN loss on the initial differentiation stage of CGCPs was examined. The loss of VN increased the expression levels of Transient axonal glycoprotein 1 (TAG1), a marker of neurons in the initial differentiation stage, in the cerebella of VNKO mice at P5 and 8 and increased the ratio of TAG1-positive cells in the primary culture of VNKO-derived CGCPs, indicating that the loss of VN accumulates the CGCPs in the initial differentiation stage. Taken together, these results demonstrate that VN promotes the progress of the initial differentiation stage of CGCPs. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Cerebellar granule cells (CGCs) constitute the vast majority of neurons in the cerebellum. Therefore, determining the mechanisms underlying the proliferation of cerebellar granule cell precursors (CGCPs) and the differentiation of CGCPs into CGCs is important for understanding cerebellar development. The developing cerebellar cortex is formed from four layers: the external granule layer (EGL), the molecular layer (ML), the Purkinje cell layer (PcL), and the internal granule layer (IGL). The proliferation of CGCPs is active in the outer external granule layer (oEGL) (Miale and Sidman, 1961). After this proliferation stage, the CGCPs stop the cell cycle in the inner external granule layer (iEGL), initiate differentiation into CGCs, and migrate through the layers

Abbreviations: VN, vitronectin; CGC, cerebellar granule cell; CGCP, cerebellar granule cell precursor; WT, wild- type; VNKO, VN knockout; iEGL, inner external granule layer; IGL, internal granule layer; P, postnatal day; EGL, external granule layer; TAG1, transient axonal glycoprotein 1; ML, molecular layer; PcL, Purkinje cell layer; oEGL, outer external granule layer; LN, laminin; Shh, Sonic hedgehog; CREB, cyclic-AMP responsive element binding protein; PBS, phosphate-buffered saline; BrdU, bromodeoxyuridine; PFA, paraformaldehyde; PHH3, phospho-histone H3; DIG, digoxigenin; DEPC, diethylpyrocarbonate; GFAP, glial fibrillary acidic protein; DAPI, 4′, 6-diamidino-2-phenylindole dihydrochloride. ⁎ Corresponding author. E-mail address: [email protected] (Y. Miyamoto).

http://dx.doi.org/10.1016/j.mcn.2015.11.013 1044-7431/© 2015 Elsevier Inc. All rights reserved.

to form the iEGL, the ML, and the PcL. Additionally, the parallel fibers of CGCPs are elongating and the CGCPs are differentiating into CGCs (Miale and Sidman, 1961; Rakic, 1971; Hatten, 1999). Differentiation into CGCs is completed when the cell bodies of CGCs reach the IGL. Proliferation of CGCPs occurs actively at postnatal days (P) 5–8 and subsequently ends at approximately P15 in mice (Hatten et al., 1997). Additionally, the differentiation of CGCPs into CGCs is completed in the IGL at approximately P20 (Hatten et al., 1997). Extracellular matrix proteins, such as laminin (LN), CCN3/NOV, and heparan sulfate proteoglycan, regulate the proliferation, differentiation and migration of CGCs (Hatten, 1999). During cerebellar development, LN is expressed in the oEGL, ML and PcL and promotes the proliferation of CGCPs through its receptor β1 integrin (Pons et al., 2001). CCN3/NOV, a member of the CCN family, reduces Sonic hedgehog (Shh)- induced proliferation of CGCPs by stimulating GSK3β through αvβ3 integrin (Le Dreau et al., 2009). Additionally, heparan sulfate proteoglycan promotes the proliferation of CGCPs through an interaction between the heparan sulfate chains and Shh (Rubin et al., 2002). Vitronectin (VN) is a cell adhesion protein that can be detected in plasma and the extracellular matrix. It is known that VN acts as a scaffold during cell anchoring and is involved in the control of blood coagulation, fibrinolysis, complement activity, and cell migration (Schvartz et al., 1999). Furthermore, it was revealed that VN regulates the proliferation and differentiation of various types of neuron precursors during

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

development. For example, VN is detected in the developing chick retina and sustains both proliferation and differentiation of cultured neuroepithelial cells from embryonic day 5 retinas (Martinez-Morales et al., 1995). Additionally, VN expression is increased with the generation of motor neuron differentiation in the floor plate of the embryonic chick neural tube, and anti-VN antibodies reduce the number of motor neurons (Martinez-Morales et al., 1997; Pons and Marti, 2000). VN is also involved in the development of CGCPs during cerebellar formation. VN suppresses Shh-induced proliferation and promotes the differentiation of CGCPs in primary cultures (Murase and Hayashi, 1998; Pons et al., 2001; Katic et al., 2014). Further analyses have suggested that the suppression of CGCP proliferation is regulated by VN-induced phosphorylation of the cyclic-AMP responsive element binding protein (CREB) in primary cultures of CGCPs from rats (Pons et al., 2001). Moreover, blocking VN binding to CGCPs inhibits the centrosome positioning at the base of the axons in primary cultures of CGCPs from mice (Gupta et al., 2010). This centrosome positioning dictates the orientation of axon elongation. Therefore, VN plays an important role in the regulation of the proliferation and differentiation of cultured CGCPs and in the orientation of axon elongation of CGCs in culture. However, the roles of VN in the developing CGCPs have not been understood fully in the developing cerebellum. In this study, we analyzed the effects of VN on the proliferation and the differentiation of CGCPs during cerebellar development using VN-knockout (KO) mice. In the cerebellar cortex of mice at P8, the loss of VN reduced the number of differentiated CGCs but did not significantly affect the proliferation of CGCPs. Additionally, the loss of VN increased the expression level of Transient axonal glycoprotein 1 (TAG1), which is a marker of neurons in the initial differentiation stage, at P5 and 8. The loss of VN also increased the ratio of CGCPs in the initial differentiation stage in primary cultures. These findings demonstrate that VN plays an important role in the progress of the initial differentiation stage of CGCPs.

2. Material and methods 2.1. Animals The C57/BL6J mice were obtained from the Charles River lab in Japan (Yokohama, Japan), and the VNKO mice were generously provided by Dr. David Ginsburg (University of Michigan). All experiments were approved by the Institutional Animal Care and Use Committee of Ochanomizu University, Japan (animal study protocols 12003, 13003, and 14005) and followed the guidelines established by the Ministry of Education, Science and Culture in Japan. The C57/BL6J and VNKO mice were kept on a 12 h light–12 h dark cycle at 22 °C.

2.2. Western blotting Mouse cerebella at P2, 5, 8, 11, 14 and 20 were collected in phosphatebuffered saline (PBS) and lysed in sample buffer (25 mM Tris–HCl, 5% glycerol, 1% SDS, 0.05% bromophenol blue, 1% 2-mercaptoethanol). The supernatants of these sample solutions were separated by SDS-PAGE and blotted on polyvinylidene difluoride membranes (ATTO, Tokyo, Japan). The membranes, which were blocked with 0.3% skim milk overnight, were incubated with primary antibodies for 1 h and horseradish peroxidase secondary antibodies for 1 h. The primary antibodies in this study included: anti-VN (1:5000; a gift from Dr. Masao Hayashi, Ochanomizu University, Tokyo, Japan), anti-β-actin (1:10,000; G043; Applied Biological Materials, Richmond, BC) and anti-TAG1 (1:1000; C9C5; Cell signaling Technology, Beverly, MA). The bands were visualized with an enhanced luminescent reagent (ATTO) and quantified using LAS3000 and Multi Gauge-ver. 2.2 (GE Healthcare UK Ltd., Little Chalfont, England).

77

2.3. Immunohistochemistry Immunohistochemistry was performed using modifications of the procedure described by Sakane and Miyamoto (Sakane and Miyamoto, 2013). Newborn mice were injected intraperitoneally with 50 μg/g bromodeoxyuridine (BrdU) (BD Biosciences, San Jose, CA) at P3, 6 and 9 and sacrificed 48 h later. Mice at P5, 8 and 11 were perfused and fixed with 4% paraformaldehyde (PFA), and the brains were fixed with 4% PFA overnight. Serial sagittal sections (20 μm thick) through the cerebella were obtained using a cryostat (Leica CM1850; Leica Microsystems, Wetzlar, Germany). The sections were incubated with primary antibodies overnight and secondary antibodies for 1 h. The primary antibodies in this study included: anti-VN (1:500; LSL-LB-2096; Cosmo Bio, San Diego, CA), anti-Calbindin (1:3000; D28K; SigmaAldrich, St. Louis, MO), anti-BrdU (1:500; MAB3222; Merck Millipore, Birellica, MA), anti-Ki67 (1:200; RM9106-S0; Thermo Fisher Scientific, San Jose, CA), anti-phospho-histone H3 (PHH3) (1:200; 06–570; Merck Millipore), anti-NeuN (1:400; MABN140, Merck Millipore), anti-TAG1 (1:50; 4D7; Developmental Studies Hybridoma Bank, Iowa City, IA) and anti-active Caspase3 (1:200; AF835; R&D systems, Minneapolis, MN). Confocal images with a single plane in the immunostained samples were captured at magnification 40 × with a confocal microscope (LSM710, Carl Zeiss, Germany). 2.4. In situ hybridization In situ hybridizations were performed as described previously (Sakane and Miyamoto, 2013). The RNA antisense probes were prepared using plasmid cDNA clones for VN transcribed with SP6 polymerase using digoxigenin (DIG)- labeling reagents and a DIG RNA labeling kit (Roche Diagnostic Corporation, Indianapolis, IN). The brains were fixed overnight at room temperature in 4% PFA in diethylpyrocarbonate (DEPC)-treated PBS, cryoprotected in 15 and 30% sucrose in DEPC-treated PBS, and embedded in OCT compound. The frozen brains were processed into 20 μm sections with a cryostat. The sections were prehybridized with hybridization buffer (Amresco, Solon, OH) for 2 h before hybridization buffer containing DIG-labeled riboprobes (200–400 ng/ml) was applied at 65 °C overnight. To visualize the results, we used an alkali phosphataselabeled DIG antibody and BM purple (Roche). The images were captured using a microscope (FSX100, Olympus, Japan). 2.5. Granule cell culture Primary culture of CGCPs was performed using a modified version of the procedure described by Weber and Schachner (Weber and Schachner, 1984). P6 mouse cerebella were aseptically isolated and the meninges were removed. The cerebella were dissociated with 1% trypsin and DNase I (Roche Diagnostics) in Hanks balanced salt solution (135.8 mM NaCl, 5.36 mM KCl, 0.34 mM Na2HPO4, 0.44 mM KH2PO4, 1 mg/ml glucose, NaHCO3) for 13 min at room temperature. The cells were dispersed by pipetting and passed through a nylon net with a mesh size of 70 μm. The dissociated cells were suspended in serum-free Neurobasal Medium (Life Technologies, Carlsbad, CA) with 500 μML-glutamine, 2% B-27 Supplement (Life Technologies), 25 mM KCl, 50 unit/ml penicillin, 50 μl/ml streptomycin and 10 nM smoothened agonist (a gift from Curis Inc., Lexington, MA). The cell density was quantified and adjusted to 6.0 × 105 cells/ml. Next, the suspensions were plated on poly-L-lysine-coated cover slips in 24-well plates and cultured in 5% CO2 at 37 °C for 72 h. At 24 h after plating, all culture media were replaced with fresh media. To label the cells with BrdU, 20 μM BrdU was added to the culture media at 24 h prior to fixation with 4% PFA. When we added VN, we added 2.5 μg/ml or 5.0 μg/ml VN to all culture media. Immunofluorescence staining using the primary cultures was performed with primary antibodies overnight and secondary antibodies for 1 h. The primary antibodies in this study included: anti-glial fibrillary acidic protein (GFAP) (1:200; G9269; Sigma-Aldrich) as well as

78

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

antibodies mentioned in previous sections. Finally, the cell nuclei were stained with 0.4 μg/ml 4′, 6-diamidino-2-phenylindole dihydrochloride (DAPI) (Roche) in PBS. The images were captured using a microscope (FSX100, Olympus, Japan). 2.6. Purification of VN from human plasma VN was purified from human plasma using heparin sepharose-4B (a gift from Dr. H. Ogawa, Ochanomizu University, Tokyo, Japan) as described previously (Yatohgo et al., 1988).

primary culture experiments, we performed a least three independent experiments and analyzed two coverslips (10 different fields (330 μm × 440 μm)/coverslip) for each culture. The data were analyzed with a one-tailed Student's t-test. The values were expressed as the mean ± standard error. Changes were considered to be significant if the p value from the Student's t-test was less than 0.05, *: p b 0.05, **: p b 0.01, and ***: p b 0.001. 3. Results

2.7. Statistical analyses

3.1. VN is expressed from the iEGL to the IGL in the developing mouse cerebellum

In the experiments using tissue sections, we used pairs of VNKO (n = 3) and WT (n = 3) littermate mice at each postnatal day, and 4 sections from each mouse were analyzed in the in vivo studies. For

Because VN has been found in the cerebella of rat and chick embryos (Murase and Hayashi, 1998; Walker and McGeer, 1998; Pons et al., 2001; Bouslama-Oueghlani et al., 2012), it is likely that developing

Fig. 1. Expression of VN in the developing mouse cerebellum. (A, B) Western blot analysis of the expression of VN in mice cerebella at P2, 5, 8, 11, 14 and 20. (A) Profile of VN expression in the developing mice cerebella from the Western blot analysis. Mouse plasma was used as a positive control for the VN protein. The molecular size of VN in all development stages is 70 kDa, which is the same size as VN in the plasma, and the 70 kDa band was not detected in P8 VNKO mouse cerebellum. (B) Relative expression levels of VN during the cerebellar development of postnatal mice. The raw value of the VN level was normalized against the β-actin level, and the mean values from three animals are normalized against the VN level at P2. The expression level of VN increased at P2-8 and peaked at P11. (C) Pattern of VN localization in mice cerebella at P8. Double immunostaining of the sagittal sections obtained from WT mice at P8. VN (green), calbindin, a marker of Purkinje cells, (red), and DAPI (blue). Calbindin is used for specification of the ML and PcL. VN expression is observed from the iEGL through the PcL. (D) Pattern of VN mRNA localization in WT mice cerebella at P8. Signals are not detected from the sections binding the sense probe. Scale bar: 50 μm. Abbrevs.: oEGL, outer external granule layer; iEGL, inner external granule layer; ML, molecular layers; PcL, Purkinje cell layer; IGL, internal granule layer.

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

79

Fig. 2. The effect of VN loss on the differentiation of CGCPS to CGCs in the developing cerebellar cortex. (A–E) Analysis of the differentiation efficiency. Immunostaining of the sagittal sections was performed on sections from the WT and the VNKO mice at P5, 8 and 11 that were injected with BrdU at 48 h prior to sacrifice (n = 3 mice). (A) Immunostained cryosections from the WT and the VNKO mice at P8. The sections were stained with anti-NeuN (green), anti-BrdU (red) and DAPI (blue). Scale bar: 50 μm. (B) Analysis of the differentiation efficiency. Loss of VN reduced the ratio of the NeuN+;BrdU+ cell number to the BrdU+ cell number for the last 48 h in the EGL at P8. (C) Analysis of the CGC number in IGL. Loss of VN reduced the number of CGCs in the IGL at P8. (D) The number of cells in EGL. Loss of VN increased the number of cells in the EGL at P8. (E) Analysis of the proliferating or differentiating CGCPs in the EGL. Loss of VN increased the ratio of the NeuN-;BrdU+ cell number to the total cell number within the last 48 h in the EGL at P8 and 11. *: p b 0.05, **: p b 0.01, ***: p b 0.001, t-tests. Abbrevs: oEGL, outer external granule layer; iEGL, inner external granule layer; ML, molecular layer; IGL, internal granule layer.

80

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

mouse cerebella also express VN. To confirm the expression of VN in the developing mouse cerebella, Western blotting was performed on the cerebella of WT mice at P2, 5, 8, 11, 14 and 20. The results showed that 70 kDa VN, which is the same size as VN in the mouse plasma, was present at all the development stages of WT mice, but not detected in P8 cerebellum from VNKO mouse (Fig. 1A). Additionally, the expression level of VN dramatically increased after P8 (Fig. 1B). Immunohistochemistry was performed on the frozen sagittal sections of P8 WT mice cerebella using the anti-VN antibody to examine the VN distribution in the developing mouse cerebellar cortex. The results showed that VN was distributed in the region from the iEGL through the IGL, and the highest expression was in the PcL and the ML (Fig. 1C). Additionally, in situ hybridization was performed on the sagittal sections of P8 mouse cerebella to examine where VN is expressed in

the developing cerebellar cortex. VN mRNA was expressed throughout the four layers of the cerebellum (Fig. 1D). This result suggests that VN protein is secreted not only from Purkinje cells but also from CGCPs in the EGL in the developing mouse cerebellum.

3.2. VN promotes the differentiation of CGCPs into CGCs in vivo In previous studies, it was shown that VN coating on culture dishes reduces the number of proliferating cells and increases the differentiated cells in the primary culture of CGCPs (Pons et al., 2001). In this study, to examine the VN functions, we examined the proliferation and differentiation of CGCPs in vivo in the developing VNKO mouse cerebellar cortex. We focused especially on P5, 8 and 11, which is the time when CGCPs are proliferated actively.

Fig. 3. The effect of VN loss on the differentiation of CGCPs to CGCs in the primary cultures. (A-C) Analysis of the differentiation efficiency for the primary cultures of CGCPs. The cells from P6 WT and VNKO mice were labeled with BrdU for 24 h before fixation (n = 3). (A) Analysis of the differentiation efficiency. Loss of VN reduced the ratio of the NeuN+;BrdU+ cell number to the BrdU+ cell number for the last 24 h in the cultured CGCPs. (B) Analysis of proliferating or differentiating CGCPs in the VNKO-derived primary culture. Loss of VN increased the ratio of the NeuN-;BrdU+ cell number to the total cell number within the last 24 h in the cultured CGCPs. (C) Immunostaining of primary cultured CGCPs. The CGCPs derived from the newborn WT and VNKO mice were immunostained with anti-NeuN (green), anti-BrdU (red) and DAPI (blue). (D) A rescue experiment to confirm the effect of VN on the differentiation to the CGCs. The differentiation was promoted depending on the density of VN (n = 3). *: p b 0.05, **: p b 0.01, ***: p b 0.001, t-tests.

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

First, to determine where to analyze the role of VN in the developing cerebellum, we chose 200 μm width areas which are identified by the white box 1 (the middle part close against the sixth lobule in the fourth/fifth lobule), the box 2 (the caudal region in the eighth lobule), and the box 3 (the depth region of the ditch which there is between the eighth and ninth lobule in the ninth lobule), and analyzed the efficiency of the differentiation of CGCPs into CGCs in the areas marked with white frames (Fig. S1A). The results showed that the loss of VN decreased the efficiency of differentiation in the all frames. Specifically, the loss of VN reduced the ratio of the neuronal marker NeuN- and BrdU-double positive cells [NeuN +;BrdU +/BrdU + ratio] most in the fourth/fifth lobule (white box 1) at P8 (Fig. S1B). Then, we decided to analyze the fourth/fifth lobule in this study. The effect of VN loss on the efficiency of differentiation was examined in the fourth/fifth lobule at P5, 8 and 11 (Fig. 2A, B). The loss of VN reduced the ratio of the differentiated, NeuN- and BrdU-double positive CGC number to the BrdU-positive cell number in the EGL at P8 (Fig. 2B) but did not affect the ratio of the 48 h-labeled BrdU-positive cell number to the whole cell number in the EGL at P8 (data not shown). In other words, the loss of VN reduced the efficiency of the differentiation of CGCPs in the EGL and did not affect the proliferation. Additionally, the loss of VN reduced the number of CGCs in the IGL at P8 (Fig. 2C; P8, WT, 306 ± 17 cells; VNKO, 260 ± 14 cells; p = 0.001; NeuN positive cells). These results demonstrate that VN promotes the neuronal differentiation of CGCPs into CGCs in the developing cerebellum. Interestingly, loss of VN reduced the ratio of the differentiated CGCPs in the EGL (Fig. 2B) but increased the number of total cells, which are almost CGCPs or CGCs, in the EGL of VNKO mice at P8 compared to the WT mice (Fig. 2D). We speculated that this increase in the total cell number in the EGL is related to the reduction of apoptosis or an increase in the ratio of proliferating or differentiating NeuN-negative CGCPs in EGL. First, we analyzed the number of active-caspase 3-positive cells in the cerebellar cortex. The result indicated that the loss of VN did not significantly affect apoptosis in the cerebellar cortex, and the numbers of apoptotic cells were very small in the cerebellar cortex of VNKO mice at P5, 8 and 11 (P8 WT, 0.2 ± 0.1%; P8 VNKO, 0.3 ± 0.1%; p = 0.14; activecaspase 3-positive cells, data not shown). Therefore, the effect of VN loss on the ratio of proliferating or differentiating CGCPs, specifically NeuN-negative and 48 h-labeled BrdU-positive cells [NeuN−;BrdU+/ DAPI ratio], in the EGL was examined. The loss of VN increased the ratio of proliferating or differentiating CGCPs in the EGL at P8 and 11 (Fig. 2E). These results demonstrate that the loss of VN increased the number of CGCPs in particular stages, especially the proliferation or the initial differentiation stage. To confirm the above in vivo data, the effect of VN loss on the proliferation or differentiation of VNKO-derived cultured CGCPs was examined (Fig. 3A − C). Before the experiment, we checked the ratio of glial cell populations. There was no significant difference between the percentage of glial cells in the primary cultures from the WT mice and the VNKO mice (WT, 2.5 ± 0.6%; VNKO, 2.9 ± 0.2%; p = 0.31; GFAP positive cells). Consistent with the in vivo data, the loss of VN reduced the efficiency of terminal differentiation (Fig. 3A) and increased the ratio of proliferating or differentiating cells (Fig. 3B) compared to the WT samples. To conduct a rescue experiment, VN was added to the VNKO-derived cultured CGCPs. VN addition rescued the effect of VN loss and increased the ratio of the differentiated CGC [NeuN +;BrdU+/BrdU+ ratio] depending on the density of VN (Fig. 3D). These data support the notion that VN promotes the differentiation of CGCPs in the developing cerebellum. 3.3. VN does not affect the proliferation of CGCPs in the developing cerebellum Next, we examined whether the proliferation or differentiation stages are affected by the loss of VN. To test whether the loss of VN

81

affects the proliferation phase of CGCPs in the developing cerebellum, we examined the ratio of cells in the cell cycle [Ki67 +;BrdU +/ BrdU+ ratio] within 48 h in P5, 8 and 11 cerebella (Fig. 4A, B). The results showed that the loss of VN did not significantly affect the ratios of cells in the cell cycle (Fig. 4B). Additionally, we did not observe any significant differences in other cell cycle analyses, including the ratios of 2 h-labeled BrdU-positive cells, which marked cells in the S-phase, and the PHH3-positive cells, which marked cells in the M-phase, in the EGL between the VNKO and the WT mice at P5, 8 and 11 (data not shown). These data demonstrate that the loss of VN did not affect the proliferation stage of CGCPs in the developing cerebellum. To confirm our above in vivo data, the proliferation stage we analyzed using VNKO-derived cultured CGCPs (Fig. 4C, D). Similar to the in vivo data, the results demonstrated that the loss of VN did not affect the ratio of cells in the cell cycle (Fig. 4D), 2 hlabeled BrdU-positive cells (Fig. S2A, B) and PHH3-positive cells (Fig. S2C, D). Furthermore, we examined the effect of exogenous VN on the proliferation of VNKO-derived cultured CGCPs. Interestingly, the presence of 5.0 μg/ml VN suppressed the ratio of the CGCP proliferation [Ki67+;BrdU+/BrdU+ ratio] within 24 h, whereas 2.5 μg/ml VN did not affect the ratio (Fig. 4E). Additionally, the presence of 5.0 μg/ml VN significantly reduced the ratios of 24 h-labeled BrdU- and PHH3-positive cells (data not shown). The result of this high concentration of VN is similar to that as shown in a previous study (Pons et al., 2001). However, as shown in Fig. 4, the suppression of the CGCP proliferation in VNKO cerebellum was not observed, suggesting that endogenous VN expression levels in the developing cerebellum are not sufficiently high to suppress the proliferation of CGCPs. 3.4. VN promotes the progress of the initial differentiation stage of CGCPs Based on our results, we hypothesized that the loss of VN suppresses the progress of the initial differentiation stage of CGCPs and accumulates them in the initial differentiation stage in the EGL. To check our hypothesis, the effect of VN loss on the distribution of TAG1, a marker of neurons in the initial differentiation stage (Bizzoca et al., 2003; Xenaki et al., 2011), and the expression levels of TAG1 were examined at P5, 8 and 11. In the cerebellar cortex of the WT and the VNKO mice at P5, 8 and 11, TAG1 was expressed in the iEGL (Fig. 5A), and the loss of VN significantly increased the expression level of TAG1 in the cerebella at P5 and 8 compared to the WT mice (Fig. 5B, C). To clarify whether these results were due to the increase in TAG1-positive cells or the TAG1 expression level per cell, we analyzed the ratio of cells in the initial differentiation stage [TAG1 +;BrdU +/BrdU + ratio] within 24 h using primary cultured CGCPs (Fig. 5D, E). The results showed that the loss of VN increased the ratio of cells in the initial differentiation stage in the cultured CGCPs (Fig. 5E), which indicates that the loss of VN increases the number of TAG1-positive CGCPs in the developing cerebella. To conduct a rescue experiment, VN was added to the VNKO-derived culture. In the rescue effect of VN, 2.5 μg/ml VN reduced the ratio of the TAG1 +;BrdU + cell number to the BrdU + cell number in the initial differentiation stage of CGCPs (Fig. 5F). However, 5.0 μg/ml VN did not rescue the effect of VN loss (Fig. 5F). This difference may be because the excess VN not only rescues the accumulation of CGCPs in the initial differentiation stage but also promotes the differentiation of CGCPs, similar to the increase in NeuN + cells shown in Fig. 3D. These results support the idea that VN promotes the progress of the initial differentiation stage of CGCPs. We speculated that the loss of VN suppresses the migration of CGCPs from the EGL to the ML and IGL because TAG-1 positive CGCPs in the initial differentiation stage do not enter the ML and IGL (Fig. 5A). To trace the migration of CGCPs in vivo, we examined the effect of VN loss on the distribution of 48-h labeled BrdU-positive cells in the EGL, ML and IGL at P5, 8 and 11 (Fig. 6A, B). The results showed that the BrdU-positive cells

82

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

in the cerebellar cortex of VNKO mice at P8 and 11 remained in the EGL compared to the WT mice (Fig. 6B). These results support our speculation and suggest that VN promotes the migration of CGCPs from the EGL to the ML in vivo.

4. Discussion In this study, we researched the roles of VN in the development of the mouse cerebellum. The results of our experiments show that the

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

83

Fig. 5. The effect of VN loss on TAG1 expression. TAG1 is a marker of CGCPs in the initial differentiation stage. (A-C) The effect of VN loss on the expression of TAG1 in the mouse cerebella at P5, 8 and 11. (A) Pattern of TAG1 localization in the developing cerebellar cortex. Immunostaining of the sagittal sections obtained from the WT and the VNKO mice at P5, 8 and 11 was performed using anti-TAG1 (green) and DAPI (blue). Scale bar: 50 μm. (B) Western blot analysis of the expression of TAG1 in the VNKO mice cerebella at P5, 8 and 11. (C) The quantification of expression levels of TAG1 in VNKO mice at P5, 8 and 11. The TAG1 expression level was normalized against its β-actin level, and the mean value from three animals was normalized against the level of WT mice. The expression levels of TAG1 increased significantly in the VNKO mice at P5 and 8 compared to the WT. (D, E) Analysis of the TAG1 positive CGCPs using primary cultures of CGCPs. The CGCPs from the WT and the VNKO mice at P6 were labeled with BrdU for the last 24 h before fixation (n = 4). (D) Analysis of the TAG1-positive cells in the primary cultures. Loss of VN increased the ratio of the TAG1+;BrdU+ cell number to the BrdU+ cell number within the last 24 h. (E) Immunostaining of the primary cultured CGCPs. The CGCPs derived from the newborn WT and VNKO mice were immunostained with anti-TAG1 (green), anti-BrdU (red) and DAPI (blue). (F) A rescue experiment to confirm the effect of VN on the early differentiation stage of CGCPs in the primary cultures. VN (2.5 μg/ml) addition reduced the ratio of the TAG1+;BrdU+ cell number to the BrdU+ cell number, but VN (5.0 μg/ml) did not affect the ratio (n = 3) *: p b 0.05, **: p b 0.01, t-tests. Abbrevs.: oEGL, outer external granule layer; iEGL, inner external granule layer.

loss of VN does not affect the proliferation of CGCPs in vivo. Additionally, we found that VN promotes differentiation into CGCs, especially the progress of the initial differentiation stage. 4.1. Endogenous VN is not enough to affect CGCPs proliferation In this study, the loss of VN did not influence the proliferation stage of CGCPs in the developing cerebellum. However, it was previously reported that VN suppresses Shh-induced proliferation and promotes the differentiation of CGCPs into CGCs through the phosphorylation of CREB in cultures of CGCPs (Pons et al., 2001). Our results are

inconsistent with this previous report. Two possibilities may explain this discrepancy. A possible reason is that the amount of endogenous VN in vivo might not be sufficient to suppress the proliferation of CGCPs. This idea is supported by our findings that the addition of 5.0 μg/ml VN suppressed the proliferation of VNKO-derived cultured CGCPs, yet the addition of 2.5 μg/ml VN did not affect the proliferation (Fig. 5E). In contrast, in the previous study, WT-derived CGCPs were cultured on 10 μg/ml VN-coated plate (Pons et al., 2001), and therefore, the excess amount of VN might suppress the proliferation of CGCPs. Furthermore, we observed that VN is localized in the iEGL (Fig. 1C), where CGCPs exit the cell cycle, and our results in Fig. 4 suggest that

Fig. 4. The effect of VN loss on the proliferation of CGCPs in the developing cerebellar cortex. (A, B) Analysis of the proliferation of CGCPs. Immunostaining of the sagittal sections from the WT and the VNKO mice at P5, 8 and 11 that were injected with BrdU at 48 h prior to sacrifice (n = 3 mice). (A) Immunostaining of the developing cerebellum at P8. The cryosections were stained with anti-Ki67 (green), anti-BrdU (red) and DAPI (blue). Scale bar: 50 μm. (B) Analysis of the cell cycle progression of CGCPs. Loss of VN did not affect the ratio of the Ki67+;BrdU+ cell number to the BrdU+ cell number in the EGL within the last 48 h before being sacrificed. (C, D) Analysis of the proliferation of CGCPs using primary cultures of CGCPs. The CGCPs from the P6 WT and VNKO mice labeled with BrdU for the last 24 h before fixation (n = 4). (C) Immunostaining of the cultured CGCPs. The CGCPs derived from the postnatal WT and VNKO mice were immunostained with anti-Ki67 (green), anti-BrdU (red) and DAPI (blue). (D) Analysis of the cell cycle progression of the cultured CGCPs. Loss of VN did not affect the ratio of the Ki67+;BrdU+ cell number to the BrdU+ cell number within the last 24 h before fixation. (E) A rescue experiment to confirm the effect of VN on the cell cycle progression of the CGCPs in the primary cultures. VN (2.5 μg/ml) did not affect the ratio of the Ki67+;BrdU+ cell number to the BrdU+ cell number, but VN (5.0 μg/ml) reduced the ratio (n = 3). ***: p b 0.001, t-tests. Abbrevs.: oEGL, outer external granule layer; iEGL, inner external granule layer; ML, molecular layer.

84

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

Fig. 6. Influence of the loss of VN on the migration of the CGCPs from the EGL to the IGL. (A, B) Analysis of the migration of CGCPs in the developing cerebellum. Immunostaining of the sagittal sections from WT and VNKO mice at P5, 8 and 11 injected with BrdU at 48 h prior to sacrifice (n = 3 mice). (A) Immunostaining of the developing cerebellum. The cryosections from the WT and the VNKO mice at P5, 8 and 11 were stained with anti-BrdU (white). (B) Distribution of the BrdU-positive cells in the EGL, ML and IGL. Loss of VN leads to a significant increase in the distributions of BrdU in the EGL at P8 and 11 and a significant decrease in BrdU in the IGL at P11. *: p b 0.05, t-tests. Abbrevs.: EGL, external granule layer; ML, molecular layer; IGL, internal granule layer.

the VN expression level in the iEGL is not sufficient to suppress the proliferation. Another possibility is that other factors compensate for the loss of VN. A possible compensation factor is CCN3/NOV or fibronectin (Pons et al., 2001; Le Dreau et al., 2009). CCN3/NOV is a ligand for αvβ3 integrin, which is also a VN receptor and has been reported to suppress CGCP proliferation through αvβ3 integrin (Le Dreau et al., 2009). 4.2. VN promotes the progress of the initial differentiation stage of CGCs We revealed that VN promotes the progress of the initial differentiation stage of CGCPs. The progress occurs in the iEGL, and in this zone, we observed the localization of VN and TAG1, which is a marker of initial differentiation in CGCPs, during the cerebellar development. We speculate that the increase in TAG1 expression is closely related to the suppression of the progress of the initial differentiation stage in CGCPs of VNKO mice. TAG1, a protein in the immunoglobulin superfamily, is expressed temporarily when CGCPs are forming axons in the iEGL and disappears before the cells migrate to the ML (Pickford et al., 1989; Bailly et al., 1996). Previous studies have suggested that TAG1 has various functions in CGCs, such as the promotion of axon formation and the determination of neuron polarity and cell migration

(Buttiglione et al., 1998; Baeriswyl and Stoeckli, 2008; Wang et al., 2010). In this study, we observed that the loss of VN increases TAG1 expression (Fig. 5B, C). Two possible explanations for the increases in TAG1 expression in the VNKO mice are considered. A possible reason is that VN controls TAG1 expression directly. However, because the mechanism to control TAG1 expression is not fully understood, we need to analyze the mechanism further. Another possible explanation for the increase in TAG1 expression is that the suppression of progress in the differentiation stage of CGCPs by the loss of VN causes the increase in TAG1 expression. This explanation is supported by our unpublished data that the loss of VN suppressed the axon formation of CGCPs. Our results show that VN promotes the differentiation without the suppression of the proliferation and provide new insight that VN induces the differentiation of CGCPs into CGCs by promoting the progress of the initial differentiation stage. In the floor plate of the embryonic chick neural plate, VN induces differentiation of cells into motor neurons without suppressing proliferation (Pons and Marti, 2000). This result is consistent with our findings. Therefore, motor neurons may be induced to differentiate by VN promoting the progress of the initial differentiation stage. We revealed that the effective amount of VN for the proliferation of CGCPs is higher than the amount for differentiation of CGCPs. This

K. Hashimoto et al. / Molecular and Cellular Neuroscience 70 (2016) 76–85

difference in the response for CGCPs may be caused by differences in the affinity of receptor–VN interactions on CGCPs. Currently, αvβ3 and αvβ5 integrins have been identified as VN receptors (Pytela et al., 1985; Cheresh and Spiro, 1987; Smith et al., 1990; Felding-Habermann and Cheresh, 1993). Our recent studies have revealed that αvβ3 integrin suppresses the proliferation and αvβ5 integrin promotes the progress of the initial differentiation stage of CGCPs (unpublished data). Additionally, as mentioned above, αvβ3 integrin has been reported to be responsible for the suppression of CGCP proliferation (Le Dreau et al., 2009). These findings support the notion that the effect of VN on the proliferation and differentiation of CGCPs is dependent on αvβ3 and αvβ5 integrins as a receptor for VN, respectively. 4.3. VN is necessary for normal development of cerebellum It was reported that VN is not essential for normal development because previous observations showed that VNKO mice have normal development, survival and fertility (Zheng et al., 1995). However, this study did not analyze the development of the cerebellum in VNKO mice at the cellular level. In this study, we performed a cellular level analysis for the developing cerebellum and primary cultures from VNKO mice. The analyses show that the loss of VN leads to significant differences, such as decreases in the efficiency of differentiation, increases in the ratio of CGCPs in the initial differentiation stage, and suppression of cell migration during the development of cerebellum. Therefore, our findings suggest that VN is necessary for normal development of the cerebellum. 5. Conclusions We revealed that VN promotes the progress of the initial differentiation stage of CGCPs during cerebellar development. However, how VN promotes the progress has not been thoroughly elucidated. Additionally, the mechanism of the regulation of TAG1 expression by loss of VN needs to be analyzed because TAG1 may be involved in the progress of the initial differentiation stage of CGCPs. We hope that this study helps to elucidate the mechanisms underlying cerebellum formation. Further studies are warranted to obtain a complete picture of VN functions in cerebellar formation. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.mcn.2015.11.013. Acknowledgments The authors wish to thank Dr. Haruko Ogawa for the heparin sepharose, Dr. Masao Hayashi for the anti-VN antibody, Dr. David Ginsburg for the VNKO mice, and Curis Inc. for the Smoothened agonist. The authors gratefully acknowledge the work of past and present members of the Miyamoto laboratory. References Baeriswyl, T., Stoeckli, E.T., 2008. Axonin-1/TAG-1 is required for pathfinding of granule cell axons in the developing cerebellum. Neural Dev. 3, 7. Bailly, Y., Kyriakopoulou, K., Delhaye-Bouchaud, N., Mariani, J., Karagogeos, D., 1996. Cerebellar granule cell differentiation in mutant and X-irradiated rodents revealed by the neural adhesion molecule TAG-1. J. Comp. Neurol. 369, 150–161. Bizzoca, A., Virgintino, D., Lorusso, L., Buttiglione, M., Yoshida, L., Polizzi, A., Tattoli, M., Cagiano, R., Rossi, F., Kozlov, S., Furley, A., Gennarini, G., 2003. Transgenic mice expressing F3/contactin from the TAG-1 promoter exhibit developmentally regulated changes in the differentiation of cerebellar neurons. Development 130, 29–43. Bouslama-Oueghlani, L., Wehrle, R., Doulazmi, M., Chen, X.R., Jaudon, F., LemaigreDubreuil, Y., Rivals, I., Sotelo, C., Dusart, I., 2012. Purkinje cell maturation participates

85

in the control of oligodendrocyte differentiation: role of sonic hedgehog and vitronectin. PLoS One 7, e49015. Buttiglione, M., Revest, J.M., Pavlou, O., Karagogeos, D., Furley, A., Rougon, G., FaivreSarrailh, C., 1998. A functional interaction between the neuronal adhesion molecules TAG-1 and F3 modulates neurite outgrowth and fasciculation of cerebellar granule cells. J. Neurosci. 18, 6853–6870. Cheresh, D.A., Spiro, R.C., 1987. Biosynthetic and functional properties of an Arg–Gly–Aspdirected receptor involved in human melanoma cell attachment to vitronectin, fibrinogen, and von Willebrand factor. J. Biol. Chem. 262, 17703–17711. Felding-Habermann, B., Cheresh, D.A., 1993. Vitronectin and its receptors. Curr. Opin. Cell Biol. 5, 864–868. Gupta, S.K., Meiri, K.F., Mahfooz, K., Bharti, U., Mani, S., 2010. Coordination between extrinsic extracellular matrix cues and intrinsic responses to orient the centrosome in polarizing cerebellar granule neurons. J. Neurosci. 30, 2755–2766. Hatten, M.E., 1999. Central nervous system neuronal migration. Annu. Rev. Neurosci. 22, 511–539. Hatten, M.E., Alder, J., Zimmerman, K., Heintz, N., 1997. Genes involved in cerebellar cell specification and differentiation. Curr. Opin. Neurobiol. 7, 40–47. Katic, J., Loers, G., Kleene, R., Karl, N., Schmidt, C., Buck, F., Zmijewski, J.W., Jakovcevski, I., Preissner, K.T., Schachner, M., 2014. Interaction of the cell adhesion molecule CHL1 with vitronectin, integrins, and the plasminogen activator inhibitor-2 promotes CHL1-induced neurite outgrowth and neuronal migration. J. Neurosci. 34, 14606–14623. Le Dreau, G., Nicot, A., Benard, M., Thibout, H., Vaudry, D., Martinerie, C., Laurent, M., 2009. NOV/CCN3 promotes maturation of cerebellar granule neuron precursors. Mol. Cell. Neurosci. 43, 60–71. Martinez-Morales, J.R., Marti, E., Frade, J.M., Rodriguez-Tebar, A., 1995. Developmentally regulated vitronectin influences cell differentiation, neuron survival and process outgrowth in the developing chicken retina. Neuroscience 68, 245–253. Martinez-Morales, J.R., Barbas, J.A., Marti, E., Bovolenta, P., Edgar, D., Rodriguez-Tebar, A., 1997. Vitronectin is expressed in the ventral region of the neural tube and promotes the differentiation of motor neurons. Development 124, 5139–5147. Miale, I.L., Sidman, R.L., 1961. An autoradiographic analysis of histogenesis in the mouse cerebellum. Exp. Neurol. 4, 277–296. Murase, S., Hayashi, Y., 1998. Concomitant expression of genes encoding integrin alpha v beta 5 heterodimer and vitronectin in growing parallel fibers of postnatal rat cerebellum: a possible role as mediators of parallel fiber elongation. J. Comp. Neurol. 397, 199–212. Pickford, L.B., Mayer, D.N., Bolin, L.M., Rouse, R.V., 1989. Transiently expressed, neuralspecific molecule associated with premigratory granule cells in postnatal mouse cerebellum. J. Neurocytol. 18, 465–478. Pons, S., Marti, E., 2000. Sonic hedgehog synergizes with the extracellular matrix protein vitronectin to induce spinal motor neuron differentiation. Development 127, 333–342. Pons, S., Trejo, J.L., Martinez-Morales, J.R., Marti, E., 2001. Vitronectin regulates sonic hedgehog activity during cerebellum development through CREB phosphorylation. Development 128, 1481–1492. Pytela, R., Pierschbacher, M.D., Ruoslahti, E., 1985. A 125/115-kDa cell surface receptor specific for vitronectin interacts with the arginine–glycine–aspartic acid adhesion sequence derived from fibronectin. Proc. Natl. Acad. Sci. U. S. A. 82, 5766–5770. Rakic, P., 1971. Neuron-glia relationship during granule cell migration in developing cerebellar cortex. A Golgi and electronmicroscopic study in macacus rhesus. J. Comp. Neurol. 141, 283–312. Rubin, J.B., Choi, Y., Segal, R.A., 2002. Cerebellar proteoglycans regulate sonic hedgehog responses during development. Development 129, 2223–2232. Sakane, F., Miyamoto, Y., 2013. N-cadherin regulates the proliferation and differentiation of ventral midbrain dopaminergic progenitors. Dev. Neurobiol. 73, 518–529. Schvartz, I., Seger, D., Shaltiel, S., 1999. Vitronectin. Int. J. Biochem. Cell Biol. 31, 539–544. Smith, J.W., Vestal, D.J., Irwin, S.V., Burke, T.A., Cheresh, D.A., 1990. Purification and functional characterization of integrin alpha v beta 5. An adhesion receptor for vitronectin. J. Biol. Chem. 265, 11008–11013. Walker, D.G., McGeer, P.L., 1998. Vitronectin expression in Purkinje cells in the human cerebellum. Neurosci. Lett. 251, 109–112. Wang, W., Karagogeos, D., Kilpatrick, D.L., 2010. The effects of Tag-1 on the maturation of mouse cerebellar granule neurons. Cell. Mol. Neurobiol. 31, 351–356. Weber, A., Schachner, M., 1984. Maintenance of immunocytologically identified Purkinje cells from mouse cerebellum in monolayer culture. Brain Res. 311, 119–130. Xenaki, D., Martin, I.B., Yoshida, L., Ohyama, K., Gennarini, G., Grumet, M., Sakurai, T., Furley, A.J., 2011. F3/contactin and TAG1 play antagonistic roles in the regulation of sonic hedgehog-induced cerebellar granule neuron progenitor proliferation. Development 138, 519–529. Yatohgo, T., Izumi, M., Kashiwagi, H., Hayashi, M., 1988. Novel purification of vitronectin from human plasma by heparin affinity chromatography. Cell Struct. Funct. 13, 281–292. Zheng, X., Saunders, T.L., Camper, S.A., Samuelson, L.C., Ginsburg, D., 1995. Vitronectin is not essential for normal mammalian development and fertility. Proc. Natl. Acad. Sci. U. S. A. 92, 12426–12430.