Cyclin Dependent Kinase 5 Is Required for the Normal Development of Oligodendrocytes and Myelin Formation

Developmental Biology 378 (2013) 94–106

Contents lists available at SciVerse ScienceDirect

Developmental Biology

journal homepage: www.elsevier.com/locate/developmentalbiology

Cyclin dependent kinase 5 is required for the normal development of oligodendrocytes and myelin formation

Yan Yang a,c,1, Haibo Wang d,1, Jie Zhang e,2, Fucheng Luo b, Karl Herrup e,2, James A. Bibb f,g, Richard Lu d,nn, Robert H. Miller b,c,n a Department of Neurology, Case Western Reserve University, School of Medicine, 10900 Euclid, Ave., Cleveland, OH 44106, United States b Department of Neurosciences, Case Western Reserve University, School of Medicine, 10900 Euclid, Ave., Cleveland, OH 44106, United States c Center for Translational Neurosciences, Case Western Reserve University, School of Medicine, 10900 Euclid, Ave., Cleveland, OH 44106, United States d Department of Developmental Biology and Kent Waldrep Foundation Center for Basic Neuroscience Research on Nerve Growth and Regeneration, The University of Texas Southwestern Medical Center, Dallas, TX 75390, United States e Department of Cell Biology and Neuroscience, Rutgers University, 604 Allison Road, Piscataway, NJ 08854-8082, United States f Department of Psychiatry, The University of Texas Southwestern Medical Center, Dallas, TX 75390, United States g Department of Neurology and Neurotherapeutics, The University of Texas Southwestern Medical Center, Dallas, TX 75390, United States article info abstract

Article history: The development of oligodendrocytes, the myelinating cells of the vertebrate CNS, is regulated by a Received 15 October 2012 cohort of growth factors and transcription factors. Less is known about the signaling pathways that Received in revised form integrate extracellular signals with intracellular transcriptional regulators to control oligodendrocyte 2 March 2013 development. Cyclin dependent kinase 5 (Cdk5) and its co-activators play critical roles in the regulation Accepted 4 March 2013 of neuronal differentiation, cortical lamination, neuronal cell migration and axon outgrowth. Here we Available online 10 April 2013 demonstrate a previously unrecognized function of Cdk5 in regulating oligodendrocyte maturation and Keywords: myelination. During late embryonic development Cdk5 null animals displayed a reduction in the number Conditional Cdk5 knockout of MBP+ cells in the spinal cord, but no difference in the number of OPCs. To determine whether the OPC reduction of oligodendrocytes reflected a cell-intrinsic loss of Cdk5, it was selectively deleted from Olig1+ Oligodendrocytes oligodendrocyte lineage cells. In Olig1Cre/+;Cdk5fl/fl conditional mutants, reduced levels of expression of Differentiation Myelination MBP and PLP mRNA were observed throughout the CNS and ultrastructural analyses demonstrated a significant reduction in the proportion of myelinated axons in the optic nerve and spinal cord. Pharmacological inhibition or RNAi knockdown of Cdk5 in vitro resulted in the reduction in oligodendrocyte maturation, but had no effect on OPC cell proliferation. Conversely, over-expression of Cdk5 promoted oligodendrocyte maturation and enhanced process outgrowth. Consistent with this data, Cdk5−/− oligodendrocytes developed significantly fewer primary processes and branches than control cells. Together, these findings suggest that Cdk5 function as a signaling integrator to regulate oligodendrocyte maturation and myelination. & 2013 Elsevier Inc. All rights reserved.

Introduction before differentiating into oligodendrocytes and myelinating target axons (Miller, 1996; Pringle and Richardson, 1993; Sherman Oligodendrocytes (OLs) the myelin-forming cells of CNS and are and Brophy, 2005). Perturbed myelination during development or derived from oligodendrocyte precursor cells (OPCs) (Raff et al., failure of remyelination in demyelinating diseases such as multiple 1984; Rowitch, 2004). Oligodendrocyte precursors originate in sclerosis (MS) leads to axonal damage and irreversible functional distinct locations and subsequently migrate throughout CNS loss. Oligodendrocyte development is regulated by a wide range of growth factors (PDGF, neuregulin-1, FGF2), signaling pathways

n (Shh, LINGO-1, BMP, Wnt and notch signaling) and transcription Corresponding author at: Case Western Reserve University, Research & factors (Olig2, Nkx2.2, Sox10 and MRF) (Barres et al., 1992; Calver Technology Management, 10900 Euclid Avenue, 2040 Adelbert Road Suite 329, Cleveland, OH 44106-7026, United States. Fax: +1 216 368 4650. et al., 1998; Emery et al., 2009; Fuller et al., 2007; Lu et al., 2000; nn Corresponding author. McKinnon et al., 1990; McKinnon et al., 1993; Mi et al., 2008; E-mail addresses: [email protected] (R. Lu), Miller et al., 2004; Richardson et al., 1988; Shimizu et al., 2005; [email protected], [email protected] (R.H. Miller). Wang et al., 1998). How the roles of these different components 1 Contributed equally. are coordinated to regulate the conversion of proliferating OPCs to 2 Present address: Division of Life Sciences, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong. OLs and subsequent myelination is currently unclear.

Cyclin-dependent kinase 5 (Cdk5) is a serine/threonine kinase Olig1Cre/+; Cdk5fl/fl CKO mice were conducted at UT Southwestern. with significant homology to cell cycle related Cdks (Hellmich These mice were derived by crossing Cdk5fl/fl mice (Benavides et al., et al., 1992) but has not been implicated in regulating cell cycle 2007; Hawasli et al., 2007)withOlig1Cre/Cre mice to generate double progression (Vermeulen et al., 2003). Although Cdk5 is expressed transgenic Olig1cre/+; Cdk5fl/+ animals which were intercrossed to in virtually all tissues, its activity appears to be largely restricted to Cdk5fl/fl mice to produce a conditional knockout mouse of Olig1Cre/+; the nervous system since its critical co-activators, p35 and p39 are Cdk5fl/fl, in which Cdk5 is deleted exclusively in Olig1+ oligoden- highly expressed in the CNS (Lew et al., 1994; Tsai et al., 1994). drocyte lineage cells. Previous studies demonstrated that Olig1 was During neuronal development Cdk5 is important in regulation of initially detected in the murine spinal cord at E8.5 at the ventral cortical lamination, neuronal migration, differentiation and synap- midline in a cohort of cells that generate both to OPCs and motor togenesis (Gilmore and Herrup, 2001; Gilmore et al., 1998; neurons (Lu et al., 2002). All animals were genotyped by PCR as Ohshima et al., 1996; Patrick et al., 1998) and loss of Cdk5 leads described previously (Lu et al., 2002; Xin et al., 2005). Both Olig1+/+; to deficits of neuronal development and neuronal cell death Cdk5 fl/+ (WT; Cdk5 fl/+) and heterozygotes Olig1Cre/+;Cdk5fl/+ served (Alexander et al., 2004; Cicero and Herrup, 2005). The regulation as controls since the heterozygote of Olig1Cre/+;Cdk5fl/+ has no of Cdk5 activity is complex and may involve the calpain-mediated abnormal phenotype compared to Olig1+/+;Cdk5 fl/+ wild type proteolysis of its co-activators p35 and p39 (Kamei et al., 2007). animal. Alternatively, the non-receptor tyrosine kinases Abl and Fyn, For quantification of oligodendrocytes P7 and P14 Olig1Cre/+; members of the Src-family of tyrosine kinases, phosphorylate Cdk5fl/fl CKO and age-matched wild type littermate animals were Cdk5 at Tyr15, which may modulate its activity (Sasaki et al., anesthetized using ketamine/xylazine, the brains and the spinal 2002; Zukerberg et al., 2000). Not all neuronal cell types are cords were dissected and fixed in 4% paraformaldehyde at 4 1C dependent on Cdk5 during development and the spinal cord is less overnight followed in 20% sucrose in PBS overnight. Coronal affected than forebrain by the lack of Cdk5 (Yip et al., 2007). sections of brains and spinal cords were cryosectioned at 16 μm During OPC differentiation Cdk5 activity has been proposed to and prepared for in situ hybridization. increase and through the phosphorylation of paxillin (Miyamoto et al., 2007), as well as WAVE2 regulate OPC migration (Miyamoto BrdU injection and immunohistochemistry for frozen cryosections et al., 2008). In addition, Emx1-Cre mediated Cdk5 conditional mice, showed hypomyelination (He et al., 2011). It is unclear, Animals were injected with 5'-bromo-2'-deoxyuridine (BrdU) however whether hypomyelination in these animals reflects Cdk5 (Sigma, St. Louis, MO, 100 mg/g, i.p.) 2 h before sacrifice. Embryos loss in oligodendrocyte lineage cells or is secondary to its loss in were taken at embryonic day E18, genotyped and fixed in 4% neurons. paraformaldehyde at 4 1C overnight followed by equilibration in Here we show that Cdk5 expression in the oligodendrocyte 20% sucrose. Coronal sections of spinal cord were cut at 20 mmona lineage is required for the maturation of oligodendrocytes and Leica cryostat. For double labeling of BrdU and OPC or OL cell normal myelination. In Cdk5 knockout animals there is a reduction markers, sections were blocked with 5% normal goat serum (NGS) in the number and distribution of spinal cord MBP+ cells. Condi- in 0.3% triton in PBS for 1 h, followed by incubation in primary tional deletion of Cdk5 from Olig1+ cells resulted in a reduction of antibody, and subsequently secondary antibody conjugated with MBP and PLP mRNA expression in the CNS and a reduction in the Alexa 488 or 596. For BrdU labeling, sections were permeabilized number of myelinated axons in the optic nerve, while inhibition of with 2 N HCL for 30 min followed by incubation with anti-BrdU Cdk5 delayed OPC differentiation and over-expression promoted antibody and Alexa conjugated secondary antibodies. OPC differentiation in the absence of changes in cell proliferation. These data suggest that Cdk5 is required for the normal develop- RNA in situ hybridization ment of mature oligodendrocyte and myelination in the developing CNS and provide insights to speed the development of re- Digoxigenin-labeled riboprobes against Mbp, Plp1, PdgfαR were myelination strategies in demyelinating diseases such as MS. used to perform RNA in situ hybridization as described previously (Lu et al., 2002). For analysis of proliferation, the conjunction of PdgfαR in situ hybridization and immunostaining for BrdU were Materials and methods performed as described previously (Xin et al., 2005). Controls and Olig1 Cre/+;Cdk5fl/fl CKO at P7 and P14 were injected (i.p.) with Antibodies 100 mg/kg BrdU for 4 h before sacrifice of the animals. Sections were hybridized with the PdgfαR riboprobe (Lu et al., 2002). After The primary antibodies used in the study include: rabbit fixation with 4% paraformaldehyde for 15 min, sections were polyclonal Cdk5 (Santa Cruz, CA, SC19); mouse monoclonal anti- treated with 2 N HCl in PBS for 30 min at 37 1C and rinsed in bodies of A2B5, O4 and O1 as described previously (Gao et al., PBS. Sections were blocked with 0.1% NP-40 and 5% normal goat 2006; Tsai et al., 2006); rabbit polyclonal antibody anti-NG2 serum in PBS for 1 h at room temperature. After addition of rat (Millipore, AB5320); rabbit polyclonal Olig2 antibody (Millipore, polyclonal anti-BrdU antibody (1:100 dilution; Sigma, St. Louis, AB9610); A mouse monoclonal antibody against MBP (Covance, MO) at 4 1C overnight, the sections were treated using a Vectastain Princeton, NJ) and a polyclonal antibody against EGFP (Invitrogen, Elite ABC kit (Vector Laboratories, Burlingame, CA); HRP was A11122). detected with diaminobenzidine (Sigma, St. Louis, MO) according to the manufacturer's instructions. Animals Tissue preparation for EM analysis Animals were maintained in the Animal Resource Center of Case Western Reserve University School of Medicine using procedures Both spinal cord and optic nerve of Olig1Cre/+; Cdk5fl/fl CKO and approved by the Institutional Animal Care and Use Committee control mice were dissected at P14 and were fixed in glutaralde- of the University. Cdk5 heterozygous mice on a C57BL6/J genetic hyde (2%) and paraformaldehyde (4%) followed by 1% OsO4, block background were obtained from Dr. Herrup (Rutgers University) stained with saturated uranyl acetate, dehydrated through graded and individual offspring were genotyped by PCR as described alcohols and embedded in Epon 812. Sections were cut in a previously (Cicero and Herrup, 2005). Studies involving the plane parallel to the surface of the slice. To assess the extent of 96 Y. Yang et al. / Developmental Biology 378 (2013) 94–106 myelination and compact myelination, ultrathin sections were O1+ or MBP+) and the total number of DAPI stained cells were examined on a Jeol 100CX electron microscope at 80 kV. To calculated and compared between treated cultures and non- compare the relative thickness of myelin sheaths in the Cdk5 cKO treated controls. For quantification of MBP immunostaining in and WT animals the g ratios were calculated from at least 35–50 the tissue sections, the average of total numbers of MBP+ cells for axons in the anterior funiculus of the spinal cord. The proportion per spinal cord section was calculated and compared between of myelinated axons was quantified from the same region of the Cdk5−/− and wild type animals at E18. To quantify the proportion of spinal cord and optic nerve in control and transgenic animals and myelinated axons in the spinal cord, the anterior funiculus region a minimum of 35–50 axons were assayed. in ventral spinal cord was selected and the at least 200 axons assayed in three animals of each genotype. To quantify the effects Preparation and characterization of spinal cord cell cultures of loss of Cdk5 on the cytoarchitecture of oligodendrocytes the number of primary processes and the number of branches were Mixed brain or spinal cord cell cultures were prepared either from assayed in the mixed brain cell cultures of WT and Cdk5−/− cells E18 Cdk5 KO or postnatal day 2 (P2) Sprague-Dawley rat pups (SD, (E18+ 5DIV or 9DIV). For foot print analyses, the average footprint Harland Animal Resource, USA) as previously described (Kerstetter area of at least 50 individual O4+ cell or MBP+ cell was measured et al., 2009) and plated at a density of 2 104 cells/coverslip in DMEM by ImageJ from WT and Cdk5−/− cultures treated with PDGF. Paired supplemented with 1% FBS-c (HyClone), 10 ng/ml PDGF (PDGF-AA, 2-way t-tests were used for statistical analysis between treated

Sigma) and N2 supplement (Invitrogen). Cultures were maintained at and non-treated groups and the significance level determined 37 1C/5% CO2 and the growth medium was changed every second day. as Po0.05. Highly enriched populations A2B5+ or O4+ OPCs were prepared from spinal cord of P1 SD rat pups by immunopanning as previously described (Robinson and Miller, 1996). Results Spinal cord explant cultures were prepared from E18 Cdk5−/− and WT mouse embryos. Cervical and thoracic segments of the Reduction of MBP+ cells during development in Cdk5 knockout spinal cord were chopped into 1–2 mm tissue pieces and transferred animals to PLL-coated coverslips. BD MatrigelTM Basement Membrane Matrix, Growth Factor Reduced (Cat# 354230, BD Bioscience) was applied To determine whether the lack of Cdk5 influenced the devel- (50–100 μl/coverslip) followed by incubation at 37 1C(Fok-Seang opment of oligodendrocytes, sections through the spinal cord of −/− and Miller, 1994). Explant cultures were grown in standard media wild type and Cdk5 littermates were labeled with the antibodies with 1% FBS-c (HyClone), 10 ng/ml PDGF (PDGF-AA, Sigma) and N2 for OPCs (Olig2) and mature OLs (MBP) and the relative number of supplement (Invitrogen). In some experiments PDGF (50 ng/ml) cells assayed. The number of MBP+ cells was reduced in the spinal −/− was added. cord of Cdk5 animals compared to WT littermates at E18 Labeling with monoclonal antibodies A2B5 (1:3), O4 (1:4) and O1 (Fig. 1G, J). For example, the mean number of MBP+ cells/section (1:4) was performed on live cells as previously described (Nishiyama in the wild type spinal cord was approximately 5071.8 and this −/− et al., 1999; Zhang and Miller, 1996). Cells were incubated in primary number was reduced to 1871.6 in Cdk5 animals (Fig. 1M). The antibody for 30 min followed by fluorescence-conjugated Alexa distribution of MBP+ oligodendrocytes was also more restricted in −/− mouse IgM 488 or 596 secondary antibodies. Cells were then fixed Cdk5 animals compared to their wild type littermates. In Cdk5 in 4% paraformaldehyde for 20 min. For double labeling, second mutant animals MBP+ cells were primarily localized to ventral primary antibodies – rabbit Cdk5 (1:200), mouse MBP (1:500) or midline regions of the spinal cord, while in littermate controls rabbit EGFP (1:500) were applied to cells and incubated overnight at MBP+ cells were distributed more widely throughout the dorso- 4 1C followed by incubation with fluorescence-conjugated secondary ventral axis with a higher concentration in ventral regions. The −/− Alexa 488 or 596-conjugated antibody. reduction in the number of MBP+ cells in Cdk5 mice was not areflection of a reduction in progenitor cells since the number Transfection of OPCs in spinal cord cell cultures of Olig2+ OPCs was not significantly different between WT and Cdk5−/− animals. Previous studies suggested that Cdk5 regulated Cells were collected from dissociated spinal cord cultures and OPC migration (Miyamoto et al., 2008) in vitro, however no centrifuged at 100 rpm at room temperature. The cell pellet was significant difference in the distribution of Olig2+ OPCs was -/- re-suspended to a density of 5 106 cells/100 ml in OPC Nucleofec- detected between Cdk5 and WT animals. These data suggest tion solution (Amaxa rat oligodendrocytes Nucleofector kit, that in the absence of Cdk5 the generation of mature oligoden- Amaxa). 100 ml of Nucleofection solution containing 5 106 cells drocytes is compromised while the generation of OPCs is relatively was added to RNAi for Cdk5 (4 mg/ml) or Cdk5-EGFP plasmids less affected. The transition from OPCs to oligodendrocytes has (2 mg/ml) as well as control pEGFP-C1 plasmid or a scrambled been linked to the proliferative status of OPCs and to assess siRNA plasmid followed by electroporation using the Amaxa changes in cell division animals were pulsed with BrdU between nucleofection apparatus with the program O-17 (Amaxa) accord- E16-E18 and the average numbers of dividing OPCs in region ing to the manufacturer's instructions. Cells were plated on matched sections of spinal cord compared between WT and −/− PLL-coated coverslip at 2 104/coverslip and grown for 2– 4 days Cdk5 animals. No significant differences in OPC proliferation + −/− in growth media as described above. were apparent between WT (61 BrdU cells71.4) and Cdk5 (56 BrdU+ cells71.7) littermates (Fig. 1B and E; H and K) Quantification and Statistics suggesting the Cdk5 is not required for OPC proliferation and dispersal but is required for the timely development of mature Cell counts: Blinded cell counts of cultured cells and labeled oligodendrocytes. cells in age and location matched frozen sections were performed on at least three coverslips or sections per condition from three Deletion of Cdk5 in Olig1+ cells results in reduced expression of Mbp separate experiments using a Leica Fluorescence microscope with and Plp1 a20 or 40 objective lens and a reticule (10 10 grid). For each experiment, seven fields were randomly selected on each coverslip The Cdk5 null animals have multiple defects and die before or section. The total number of immunopositive cells (A2B5+,O4+, birth limiting their usefulness in the analyses of the later stages Y. Yang et al. / Developmental Biology 378 (2013) 94–106 97

Fig. 1. The maturation of spinal cord oligodendrocytes is compromised in the absence of Cdk5 although the proliferation of OPCs is not significantly affected. Fig. 1(A–F) At E 18 the relative number of Olig2+ cells is not significantly different in the spinal cords of Cdk5−/− animals and WT littermates. Total cell proliferation was also not significantly altered in Cdk5−/− animals. Fig. 1(G–L) The number of MBP+ cells was significantly reduced in the spinal cords of Cdk5-/- animals and their distribution was also more restricted to ventral midline regions. The extent of BrdU incorporation in Olig 2+ and MBP+ cells between WT and Cdk5−/− was not significantly different. Fig. 1(M) Quantitation of the relative numbers of Olig2+,MBP+ and double labeled cells in WT and Cdk5−/− spinal cord sections. The data is taken from at least 4 sections from 3–5 animals in 3 independent experiments L¼ lateral spinal cord, V¼ ventral spinal cord. **po0.05, Scale bar¼25 μm. of oligodendrocyte maturation and myelination. To determine by the loss of Cdk5. Consistent with the reduction in myelin gene whether the reduction in MBP+ cells in Cdk5−/− animals reflected expression, ultrastructural analysis revealed a significant reduction the loss of Cdk5 function in cells of the oligodendrocyte lineage, of myelinated axons and compact myelin in P14 spinal cord of Cdk5 was specifically deleted in Olig1+ cells using Cre-lox tech- Olig1Cre/+;Cdk5 cKO compared to littermate controls (Fig. 3C–F). nology (Fig. 2A). In contrast to Cdk5−/− animals, the Olig1Cre/+; The relative proportion of myelinated axons was reduced from Cdk5fl/fl (Cdk5 cKO) animals are viable and fertile. Comparison of 44.9%72.4 in control animals to 30.5%72.9 in the Olig1Cre/+;Cdk5 the mRNA expression of the myelin components PLP and MBP by cKO animals (Fig. 3I). Analyses of the g ratios between WT and in situ hybridization, however, showed a substantial reduction in Olig1Cre/+;Cdk5 cKO animals revealed a reproducible reduction in the expression of these myelin genes in Cdk5 cKO animals the thickness of myelin sheaths in Olig1Cre/+;Cdk5 cKO animals compared to littermate controls in the corpus callosum and over- (Fig. 3E,F, J). To determine whether the reduction in myelination lying cortex at postnatal day 7 (Fig. 2B–E). This phenotype was not was apparent in other white matter regions, the proportion of restricted to early postnatal development, rather comparable myelinated axons was compared in the optic nerves of wild type analyses at P14 showed the phenotype to be relatively more and Olig1Cre/+;Cdk5 cKO animals. Consistent with the findings in severe than at P7 (Fig. 2F–I). In the P7 spinal cord quantification the spinal cord in Olig1Cre/+;Cdk5 cKO animals, an apparent reduc- of plp mRNA positive cells revealed a significant reduction in Cdk5 tion in the proportion of myelinated axons (5.7%70.6) was seen in cKO animals (2972.8 in gray matter and 13876.3 in white the optic nerve compared to littermate WT (18.3%71.1) controls matter) compared to WT littermates animals (4973.0 in gray (Fig. 3G–J, K) and as in the spinal cord there was an increase in the matter and 20777.6 in white matter, Po0.001; Fig. 2J–L). As in g ratio of Olig1Cre/+;Cdk5 cKO optic nerve myelinated axons Cdk5−/− animals, quantitation of the number of PDGFaR+ OPCs and (Fig. 3L). Taken together these data suggest Cdk5 is important their relative level of proliferation revealed no significant differ- for the generation of the correct cohort of oligodendrocytes and ences in either the total number of OPCs or their incorporation of subsequent myelin formation throughout the CNS. BrdU in Cdk5 cKOs compared to littermate controls (Fig. 2M–Q). Cdk5 is expressed in the different stages of OPCs and oligodendrocytes Deletion of Cdk5 in Olig1+ cells results in a reduction of myelinated and selectively regulates oligodendrocyte maturation axons in Olig1Cre/+;Cdk5fl/fl cKO The reduction in mature oligodendrocytes and myelination The reduction in the number of mature oligodendrocytes may reflect a role for Cdk5 in multiple aspects of oligodendrocyte persisted during development, was widespread, and resulted in a development. Characterization of the expression of Cdk5 in cells of reduction in the level of myelination. In the spinal cord of Olig1Cre/+; the oligodendrocyte lineage in vitro revealed its expression Cdk5 cKO animals in situ hybridization at P14 showed reduced throughout oligodendrocyte development in the majority of cells. expression of MBP mRNA compared to littermate controls (Fig. 3A, B). In cultures of pan-purified rat A2B5+ cells virtually all cells This reduction was relatively uniform in all regions of the expressed Cdk5 in the perinuclear cytoplasm and in the nucleus spinal cord with clear reductions in both dorsal and ventral white (Fig. 4A–C) and this expression was maintained as the cells mater suggesting the majority of oligodendrocytes were affected developed into O4+ cells (Fig. 4D–F). Mature oligodendrocytes 98 Y. Yang et al. / Developmental Biology 378 (2013) 94–106

Fig. 2. The expression of Mbp and Plp1 mRNA is reduced during postnatal development in Olig1Cre/+; Cdk5fl/fl cKO animals compared to Olig1+/+; Cdk5fl/+ littermate controls. (A) An outline of the genetic approach of conditional deletion of Cdk5 in olig1+ cells used in the current study. Exons II-V encoding vital Cdk5 catalytic-domain components were flanked with loxP elements and excised by Olig1-Cre. (B–E) At P7, in situ hybridization reveals a reduction of Mbp and Plp1 mRNA in the corpus callosum and overlying cortex of Olig1Cre/+; Cdk5fl/fl cKO (C & E) compared to their WT littermates (B, D). F-I the reduced expression mRNA for Mbp and Plp1 is more pronounced at P14 in the corpus callosum of Olig1Cre/+; Cdk5fl/fl cKO (G, I) compared to their WT littermates (F, H). (J–L) at P7 spinal cord, the total number of plp1+ cells in both gray matter (GM) and white matter (WM) is significantly reduced in Olig1Cre/+; Cdk5fl/fl cKO revealed by in situ hybridization. No significant difference is seen in the number of pdgfαR+ OPCs following a 2 h pulse of BrdU by double labeling of BrdU+ (brown)/ pdgfαR+ (blue) cells (arrows) in the cortex (Ctx) and in corpus callosum (CC) between WT (M, O) and Olig1Cre/+; Cdk5fl/fl cKO (N, P). Quantitative data is shown in bar graphs Q. The data represent the mean7SD taken from at least 8 sections from 3 independent animals of each genotype. Scale bar¼25 μm. Y. Yang et al. / Developmental Biology 378 (2013) 94–106 99

Fig. 3. Conditional deletion of Cdk5 in Olig1+ cells results in myelination defects. Reduced myelination in the spinal cord and the optic nerve of Olig1Cre/+; Cdk5fl/fl cKO.(A–B) In situ hybridization of mRNA expression of Mbp shows reduced level in the spinal cord of Olig1Cre/+; Cdk5fl/fl cKO at P14 compared to their WT littermates. (C–J) At P14 Olig1Cre/+; Cdk5fl/fl cKO contained significantly reduced number of myelinated axons compared to their WT littermates in both spinal cord (C–F) and the optic nerve (G–J). The relative thickness of the myelin was also reduced in the Olig1Cre/+; Cdk5fl/fl cKO animals and g ratio for spinal cord and optical nerve axons shows a significant reduction in Olig1Cre/+; Cdk5fl/fl cKO (J, L). Quantitative analysis of the proportion of myelinated axons in the spinal cord and optic nerve at P14 reveals a significant reduction in the absence of Cdk5 in oligodendrocyte lineage cells (I, K). Po0.001. Scale bar¼25 μm. 100 Y. Yang et al. / Developmental Biology 378 (2013) 94–106

Fig. 4. Cdk5 is expressed in cells during multiple distinct stages of development of the oligodendrocyte lineage. (A–C) In purified primary A2B5+ cell cultures from rat spinal cord, Cdk5 is expressed in the perinuclear cytoplasm and nuclei (red) of virtually every A2B5+ cell (green). (D–F) Cdk5 (red) is expressed in the majority of O4+ cells (green) as well as in more mature (G–I) MBP+ cells. (J–L) The expression of Cdk5 in immature cells of the oligodendrocyte lineage is retained in vivo where Cdk5 (red) is expressed in NG2+ cells (green) in the spinal cord of adult mouse. Bar¼25 μm. identified through expression MBP+ also expressed Cdk5 predo- of mature oligodendrocytes in roscovitine treated cultures is not minantly in their cell bodies (Fig. 4G–I). The expression of Cdk5 a consequence of enhanced proliferation of OPC blocking their was not restricted to cells of the oligodendrocyte lineage during differentiation. Pulsing OPC cultures with BrdU for 7 h prior to development but as previously described (Yip et al., 2007)was analyses revealed and no significant changes in the proportion also expressed in neurons and NG2+ cells of adult rat spinal cord of NG2+ (control 21.2%72.1: experimental 20%72.3) or Olig2+ (Fig. 4J–L). (control 19.2%71.9: experimental 17.1%72.2) OPCs that incorpo- While the Olig1 conditional deletion studies strongly implicate rated BrdU compared to controls (Fig. 5S, n¼3) suggesting that a requirement for Cdk5 in the development of oligodendrocytes Cdk5 is not critical for OPC proliferation but, rather is important they provide limited insights into which specific biological pro- for the transition of newly formed oligodendrocytes to mature cesses are influenced. The preferential effect of Cdk5 loss on MBP+ oligodendrocytes. cells in the mutant animals suggests that Cdk5 is predominantly involved in later stages of oligodendrocyte maturation. Consistent Knockdown of endogenous Cdk5 delays oligodendrocyte maturation with this notion treatment of P1 pan-purified A2B5+ spinal cord while over-expression of Cdk5 promotes MBP+ cell appearance OPCs with the Cdk inhibitor roscovitine (10 mM) (Gray et al., 1999; Jorda et al., 2012) had a small but not significant effect on the The roscovitine studies suggest that kinase activity is important numbers of NG2+ cells (control 15.9%72.1; experimental 10.8% for the maturation of oligodendrocytes. To define the specific role 72.1) and Olig2+ (control 18.6%71.5; experimental 12.4% 71) of Cdk5 in oligodendrogenesis, Cdk5 was targeted by RNAi in cells that developed (Fig. 5A–L, T). By contrast, the number of MBP+ spinal cord mixed cell cultures via electroporation using Amaxa cells that developed was significantly reduced in roscovitine treated Neucleitransfector and the development of oligodendrocytes were cultures compared to controls (control 20.1%71.8: experimental assessed. In cultures transfected with Cdk5 RNAi endogenous Cdk5 8.1%71.7, po0.05) (Fig. 5M–R, T). Not only were the numbers was markedly down-regulated after 2 days (Fig. 6Ab-b″) while in of MBP+ cells reduced but the morphological complexity of the control cultures the majority of cells expressed Cdk5 (Fig. 6Aa-a″), residual MBP+ cells was also reduced following roscovitine indicating that the Cdk5 RNAi effectively suppresses endogenous treatment such that the number of MBP+ cells with long and Cdk5 expression. Analysis of the oligodendrocyte development in complex branched processes or sheets of membrane was decreased Cdk5 RNAi treated cultures demonstrated that the complexity of compared to controls (Fig. 5M–R). The reduction in the number the processes of O4+ cells lacking Cdk5 was reduced, resulting in Y. Yang et al. / Developmental Biology 378 (2013) 94–106 101

Fig. 5. Blocking Cdk5 with roscovitine inhibits oligodendrocyte maturation but does not affect OPC proliferation. (A–F) The number of NG2+ cells present in A2B5 pan purified cell cultures was only slightly reduced following inhibition of Cdk5 activity for 4 days by treatment with roscovitine and their proliferation as measured by incorporation of BrdU was not significantly different. (G–L) The number of Olig2+ cells present in A2B5 pan purified cell cultures was slightly reduced following inhibition of Cdk5 activity for 4 days by treatment with roscovitine and their proliferation as measured by incorporation of BrdU was not significantly different. (M–R) the number of MBP+ cells that developed in A2B5 pan purified cell cultures was significantly reduced following inhibition of Cdk5 activity for 4 days by treatment with roscovitine. The morphology of MBP+ cells in roscovitine treated cultures was notably less complex that in vehicle treated control cultures. The cells contained fewer processes and lacked characteristic membrane sheets of mature MBP+ cells. Quantitation is shown in histogram (S, T) and represents the mean7SD taken from at least 3 independent coverslips form 3 different experiments. Bar¼25 μm. 102 Y. Yang et al. / Developmental Biology 378 (2013) 94–106

Fig. 6. The maturation of oligodendrocytes is modulated by knockdown or overexpression of Cdk5 in spinal cord cultures. (A) In spinal cord cultures, cells transfected with RNAi for Cdk5. (Ab-b″) showed approximately 98% endogenous Cdk5 knockdown compared with cells transfected with empty-vector (EGFP-C3) as a control (Aa-a″). (B) The morphological complexity of cells transfected with control (Ba-a″) constructs was signiﬁcantly more pronounced than cells transfected with Cdk5 RNAi (Bb-b″) (green) and co-labeled with O4+ (red) and MBP+ (red) (Bc-d″). Cdk5 RNAi transfected cells had shorten processes and a less complex morphology compared to control cells. (Be-e″) Cells that escaped Cdk5 knockdown retained a complex morphology in experimental cultures demonstrating the speciﬁcity to the Cdk5 effect. (C) Over expression of Cdk5 results in increased complexity of cultured O1+ (Cb-b″) and MBP+ oligodendrocytes (Cd-d″). More than 6% of the MBP+ cells had an increased complexity of their processes in cultures driven to overexpress Cdk5-EGFP compared to control. Quantitative data were shown in graphics. Scale bar¼25 μm.

a cell body with relatively few short processes (Fig. 6Bb-b″). A transfected with Cdk5-EGFP (Fig. 6C) or a randomly scrambled similar reduction in the complexity of MBP+ cell processes was siRNA (not shown) and labeled with O1+ or anti-MBP+ antibodies. seen in MBP+ cells following Cdk5 RNAi treatment (Fig. 6Bc-c″). Overexpression of Cdk5 resulted in O1 and MBP+ cells with a more Oligodendrocytes in the cultures that escaped electroporation mature morphology including more highly branched processed retained Cdk5 expression and developed normal process arboriza- and sheets of membrane (Fig. 6Cb-b″ and Cd-d″) compared to tions (Fig. 6Be-e″). These findings are consistent with the effects cells transfected with a control construct containing only EGFP of pharmacological inhibition of Cdk5 and suggest that it plays a (Fig. 6Ca-a″ and Cc-c″). Quantitative analysis demonstrated that critical role in regulating oligodendrocytes development and approximately 10% of the MBP+ cells had complex morphology in maturation. Quantification of the proportion of Cdk5+/MBP+ cells the culture transfected with Cdk5-EGFP, while less than 4% of MBP+ demonstrated a significant decrease in cells subjected to Cdk5 cells had the mature phenotype in control cultures. knockdown compared to controls (Fig. 6D). While not all cells in the cultures had reduced expression of Cdk5 none of the The extent of oligodendrocyte process outgrowth and branching is cells that lacked Cdk5 had detectable expression of MBP, while significantly reduced in the absence of Cdk5 cells that retained Cdk5 exhibited normal MBP expression, suggesting reduced or delayed maturation of oligodendrocytes in the To assess in more detail the effects of Cdk5 on the elaboration of absence of Cdk5. a mature morphological oligodendrocyte phenotype, the morphol- To determine whether increasing the expression of Cdk5 ogy of WT and Cdk5−/− cells was compared in vitro. While there enhanced the maturation of oligodendrocytes, cultures were was no significant difference in the proportion of O4+ or O1+ Y. Yang et al. / Developmental Biology 378 (2013) 94–106 103

Fig. 7. Absence of Cdk5 results in a specific reduction in the extent of mature oligodendrocyte processes outgrowth and branching. (A) Brain mixed dissociated cell cultures from WT and E18 Cdk5 KO were labeled with O4, O1 and MBP antibodies at 2.5, 7 and 9 days in vitro (DIV) respectively. Although a slight reduction in the proportion O4+ or O1+ oligodendrocytes in Cdk5−/− cultures was seen it was not significant (7A a–d). By contrast, the number of MBP+ cells was significantly reduced in Cdk5−/− cultures (7A e–f). Quantification shows in (7A g). A slight reduction in the number of processes and the extent of branching in O4+ of Cdk5−/− cells but a robust reduction of the number of processes and the extent of branching in MBP+ cells that developed in Cdk5−/− cultures compared to WT (Fig. 7A h). nnnpo0.001. (B) The perturbed development of oligodendrocytes in the absence of Cdk5 is not rescued by addition of exogenous PDGF in spinal cord explant cultures. (a, b) Addition of exogenous PDGF (10 ng/ul) in spinal cord explant culture of E18 Cdk5−/− did not enhance the number or morphology of O4+ cells in Cdk5−/− cultures compared to Cdk5+/+. (c, d) Addition of exogenous PDGF did not significantly alter the morphology or the number of MBP+ cell that developed in explant cultures of E18 Cdk5−/− animals. Quantitative analysis of the average footprint for individual MBP+ cells from WT or Cdk5−/− animals treated with PDGF treatment showed a substantially smaller footprint from Cdk5−/− cells (Fig. 7B e). Scale bar¼25 μm in A and B. oligodendrocytes that developed in the CDK5−/− cultures (Figs. 7A PDGF does not rescue reduced oligodendrocyte maturation in Cdk5-/- (a–d)) the number of MBP+ cells was significantly reduced (Fig. 7A E18 spinal cord explant culture (e–f)). With maturation, oligodendrocytes acquire an increasingly complex morphology and while a slight reduction in the number One possible mechanism that might account for the perturba- of process and the extent of branching was apparent in O4+ cell tion of oligodendrocyte differentiation in the absence of Cdk5 is a lacking Cdk5 the effect was significantly more pronounced in limited response to PDGF, a major OPC mitogen and survival factor mature MBP+ cells (Fig. 7A(e–f)). For examples while control MBP for newly generated oligodendrocytes (Raff et al., 1988; Richardson + cells had a mean of 5.170.3 primary process and 19.871.5 et al., 1988). To determine if the lack of mature oligodendrocytes in branches, Cdk5−/− cells had a mean of 3.470.23 primary process Cdk5−/− animals could be reversed by exposure to PDGF, spinal and 9.470.75 branches. cord explants from WT and Cdk5−/− animals were grown in the 104 Y. Yang et al. / Developmental Biology 378 (2013) 94–106 presence of exogenous PDGF (10 ng/ml) for 7days. Addition of Cdk5 indicating that it is not essential for OPC dispersal in vivo. PDGF failed to rescue the perturbed development or morphology Rather we propose that a dysregulation of cell polarity and of oligodendrocytes in Cdk5−/− cultures. The number of O4+ and motility events underlie the disruption of maturation of oligoden- MBP+ cells was decreased in Cdk5−/− derived culture compared to drocytes devoid of Cdk5. Several lines of evidence are consistent WT (Fig. 7B(b and d)) and the MBP+ cells that did develop had few with this hypothesis. First in Cdk5 null animals the appearance of short processes (Fig. 7B(d)) compared to the morphology of WT MBP+ cells in the embryonic spinal cord is reduced and more MBP+ cells that possessed long highly branched processes (Fig. 7B restricted than in wild type littermates, implicating a maturational (c)). The quantitative analysis of average footprint for individual defect. Second, the reduction in the complexity of Cdk5 null MBP+ cells of WT or Cdk5−/− with PDGF treatment showed a oligodendrocytes and the reduced length of their processes substantial decrease with Cdk5−/− cells (Fig. 7B(e)). The reduction compared to wild type cells suggests that oligodendrocyte process in the number and morphological complexity of O4+ cells in these extension is compromised in the absence of Cdk5. This is remi- Cdk5−/− cultures likely reflects the overlapping expression of O4 niscent of the lack of axonal growth seen in neurons lacking Cdk5 and MBP with the extended culture period. Taken together these (Fang et al., 2011; Hahn et al., 2005). One potential mechanism for results suggest that lack of Cdk5 activity effects the maturation of the modulation of oligodendrocyte process development seen in oligodendrocytes and their ability to myelinate on schedule. the absence of Cdk5 may be phosphorylation of the paxillin by Cdk5 that influence paxillin's association with the focal adhesion kinase (FAK) in an oligodendrocyte cell line (Miyamoto et al., Discussion 2007). Whether similar signaling pathways are operative in primary oligodendrocytes is currently unclear. Finally, the reduc- The molecular mechanisms mediating the development of tion in myelination in Cdk5 null animals may also reflect the oligodendrocytes are highly complex with multiple growth factors reduction in the cellular polarity or motility of the myelinating and extracellular signals regulating the induction, proliferation, processes. Indeed, the mechanics of the initiation of myelination migration, differentiation of OPCs and maturation of oliogoden- are still not well understood, although it is clear they reflect the drocytes. How these signals are integrated within the cells to polarization of oligodendrocytes, the compartmentalization of the influence their biology is not well understood. Here we show that membrane domains and the motility of the myelinating process the intracellular signaling molecule Cdk5 plays an important role (Kippert et al., 2007). in regulating the appearance of oligodendrocytes at a specific While alterations in the morphology and myelination of Cdk5 stage in their development. In Cdk5 null animals during embryo- null oligodendrocytes are consistent with changes in cell motility, nic development the number and distribution of mature oligoden- the reduction in expression of MBP is less easily explained. It may drocytes is significantly reduced in the spinal cord. In aged- be, however that the reduction in MBP expression is also a direct matched WT animals MBP+ cells are localized in the ventral effect of modulation of intracellular motile activity in the absence midline of the spinal cord as well as more laterally while in of Cdk5. Earlier studies demonstrated that the mRNA for MBP is Cdk5 null animals they are restricted to midline regions. Further- not located in the cytoplasm adjacent to the oligodendrocyte more, while the number and distribution of Olig2+ progenitor nucleus, but rather is transported to the terminal of the processes cells is relatively unaffected, the number of MBP+ cells is sig- where it is translated into protein that is locally inserted into the nificantly reduced in Cdk5 null animals throughout development. membrane (Boccaccio et al., 1999; Boccaccio and Colman, 1995). The reduction in mature oligodendrocytes is not a reflection of The mechanisms of mRNA transport have been extensively studied reduced proliferation by OPCs since the proportion of BrdU+ cells and the data support a hypothesis in which clusters of mRNAs are was not significantly different between Cdk5−/− and wild type transported along microtubules through the activity of kinesins animals. Consistent with these data, conditional deletion of Cdk5 and dyneins (Carson et al., 1998; Carson et al., 1997; Shan et al., from olig1+ cells resulted in reduced expression of MBP and PLP in 2003). This activity is likely a consequence of cell polarization and postnatal animals although there was no obvious effect in the has recently been shown to be regulated both by adapter proteins number of PDGFαR+ cells or their proliferation in either white or such as heterogeneous nuclear ribonucleoprotein (hnRNP) (White gray matter. The reduction in myelin gene expression is reflected et al., 2012). This intracellular transport, similar to fast axonal in a reduction in the level of myelination in both the spinal cord transport, (Drake et al., 1985) also occurs in neurons where it and the optic nerve that remains pronounced until at least the targets dendritic RNAs to post synaptic sites. In neurons, retro- second postnatal week. Analysis of the expression and function of grade transport of proteins important in polarity and migration Cdk5 expression in cells of the oligodendrocyte lineage revealed such as LIS1 and NDEL is dependent on the activity of Cdk5 and its expression in both early and late stages of the oligodendrocyte this related intracellular transport is completely blocked in cul- lineage. Both kinase inhibition and RNAi-induced knockout down tured neurons treated with a dominant negative Cdk5 (Pandey and of Cdk5 impaired the biochemical and morphological maturation Smith, 2011). Based on these data we propose that the disruption of oligodendrocytes, while over expression of Cdk5 resulted in in the maturation of oligodendrocytes in the absence of Cdk5 is more robust oligodendrocyte maturation. Morphological analyses largely a reflection of perturbed transport and targeting of mRNAs of cultured oligodendrocytes demonstrated a reduction in the including MBP that are essential for directed oligodendrocyte number of primary processes and branching in the absence of precursor migration, process outgrowth and myelination. Cdk5 that correlated with reduced expression of MBP. Thus, in the Recent studies have begun to reveal the distinct roles for oligodendrocyte lineage the primary role of Cdk5 appears to be to different intracellular signaling pathways in the maturation of modulate maturational and morphological changes in oligoden- oligodendrocytes and myelination. For example, over expression of drocytes rather than the cell division kinetics of the OPC AKT1 in oligodendrocytes and their precursors via the PLP pro- population. moter results in exuberant production of myelin in the CNS that One common linkage between many of the biological phenom- ultimately results in axonal compression and death suggesting ena impacted by Cdk5 in neuronal cells is a connection to cell that signaling through the AKT pathway selectively regulates the polarity and motility (Morfini et al., 2004). Previous in vitro studies extent of myelination in oligodendrocytes rather than in cell death suggested that Cdk5 modulates OPC migration (Miyamoto et al., as it does in many other cell types (Narayanan et al., 2009; Tyler 2008), however the distribution of OPCs in the developing spinal et al., 2009). Consistent with a role for AKT signaling in modulating cord was not significantly different in transgenic animals lacking myelination inhibition of mTOR by rapamycin treatment results in Y. Yang et al. / Developmental Biology 378 (2013) 94–106 105 a blockade in oligodendrocyte development (Gomez et al., 2011; Gomez, O., Sanchez-Rodriguez, A., Le, M., Sanchez-Caro, C., Molina-Holgado, F., Guardiola-Diaz et al., 2012). By contrast, inhibition of the ERK Molina-Holgado, E., 2011. Cannabinoid receptor agonists modulate oligodendrocyte differentiation by activating PI3K/Akt and the mammalian target of pathway alters the timing of myelination in the CNS (Fyffe- rapamycin (mTOR) pathways. Br. J. Pharmacol. 163, 1520–1532. Maricich et al., 2011) and possibly more importantly reduces the Gray, N., Detivaud, L., Doerig, C., Meijer, L., 1999. ATP-site directed inhibitors of thickness of the myelin sheaths that develop around large dia- cyclin-dependent kinases. Curr. Med. Chem. 6, 859–875. Guardiola-Diaz, H.M., Ishii, A., Bansal, R., 2012. Erk1/2 MAPK and mTOR signaling meter axons. Like Cdk5 neither the AKT nor ERK pathways appear sequentially regulates progression through distinct stages of oligodendrocyte to be essential for the proliferation of OPCs. Rather OPC prolifera- differentiation. Glia 60, 476–486. tion may be mediated predominantly by p38 kinase (Chew et al., Hahn, C.M., Kleinholz, H., Koester, M.P., Grieser, S., Thelen, K., Pollerberg, G.E., 2005. 2010). Further studies are required to define the interactions Role of cyclin-dependent kinase 5 and its activator P35 in local axon and growth cone stabilization. Neuroscience 134, 449–465. between AKT and ERK signaling and Cdk5. Hawasli, A.H., Benavides, D.R., Nguyen, C., Kansy, J.W., Hayashi, K., Chambon, P., In conclusion our studies demonstrate an important role for Greengard, P., Powell, C.M., Cooper, D.C., Bibb, J.A., 2007. Cyclin-dependent Cdk5 in the development of oligodendrocytes and myelination. kinase 5 governs learning and synaptic plasticity via control of NMDAR degradation. Nat. Neurosci. 10, 880–886. Several intracellular pathways converge on Cdk5 and future He, X., Takahashi, S., Suzuki, H., Hashikawa, T., Kulkarni, A.B., Mikoshiba, K., studies will be directed at defining which of these pathways are Ohshima, T., 2011. Hypomyelination phenotype caused by impaired differentia- essential for the timely myelination and myelin repair after injury. tion of oligodendrocytes in Emx1-cre mediated Cdk5 conditional knockout mice. Neurochem. Res. 36, 1293–1303. Hellmich, M.R., Pant, H.C., Wada, E., Battey, J.F., 1992. Neuronal cdc2-like kinase: a cdc2-related protein kinase with predominantly neuronal expression. Proc. – Acknowledgments Natl. Acad. Sci. U. S. A. 89, 10867 10871. Jorda, R., Paruch, K., Krystof, V., 2012. Cyclin-dependent kinase inhibitors inspired by roscovitine: purine bioisosteres. Curr. Pharm. Des. 18, 2974–2980. This work was supported by the National Institutes of Kamei, H., Saito, T., Ozawa, M., Fujita, Y., Asada, A., Bibb, J.A., Saido, T.C., Sorimachi, Health (NIH) to R.H.M and Y.Y (RO1NS077942), NIH to Q.R. L H., Hisanaga, S., 2007. Suppression of calpain-dependent cleavage of the CDK5 activator p35 to p25 by site-specific phosphorylation. J. Biol. Chem. 282, (R01NS072427 and RO1NS075243), J.A.B. (MH083711, DA033485, 1687–1694. and NS073855); and Myelin Repair Foundation to R.H. M. Kerstetter, A.E., Padovani-Claudio, D.A., Bai, L., Miller, R.H., 2009. Inhibition of CXCR2 signaling promotes recovery in models of multiple sclerosis. Exp. Neurol. 220, 44–56. References Kippert, A., Trajkovic, K., Rajendran, L., Ries, J., Simons, M., 2007. Rho regulates membrane transport in the endocytic pathway to control plasma membrane specialization in oligodendroglial cells. J. Neurosci. 27, 3560–3570. Alexander, K., Yang, H.S., Hinds, P.W., 2004. Cellular senescence requires CDK5 Lew, J., Huang, Q.Q., Qi, Z., Winkfein, R.J., Aebersold, R., Hunt, T., Wang, J.H., 1994. A repression of Rac1 activity. Mol. Cell Biol. 24, 2808–2819. brain-specific activator of cyclin-dependent kinase 5. Nature 371, 423–426. Barres, B.A., Hart, I.K., Coles, H.S., Burne, J.F., Voyvodic, J.T., Richardson, W.D., Raff, M.C., Lu, Q.R., Sun, T., Zhu, Z., Ma, N., Garcia, M., Stiles, C.D., Rowitch, D.H., 2002. Common 1992. Cell death and control of cell survival in the oligodendrocyte lineage. Cell 70, developmental requirement for Olig function indicates a motor neuron/oligo- 31–46. dendrocyte connection. Cell 109, 75–86. Benavides, D.R., Quinn, J.J., Zhong, P., Hawasli, A.H., DiLeone, R.J., Kansy, J.W., Lu, Q.R., Yuk, D., Alberta, J.A., Zhu, Z., Pawlitzky, I., Chan, J., McMahon, A.P., Stiles, C. Olausson, P., Yan, Z., Taylor, J.R., Bibb, J.A., 2007. Cdk5 modulates cocaine reward, D., Rowitch, D.H., 2000. Sonic hedgehog–regulated oligodendrocyte lineage motivation, and striatal neuron excitability. J. Neurosci. 27, 12967–12976. genes encoding bHLH proteins in the mammalian central nervous system. Boccaccio, G.L., Carminatti, H., Colman, D.R., 1999. Subcellular fractionation and Neuron 25, 317–329. association with the cytoskeleton of messengers encoding myelin proteins. McKinnon, R.D., Matsui, T., Dubois-Dalcq, M., Aaronson, S.A., 1990. FGF modulates J. Neurosci. Res. 58, 480–491. the PDGF-driven pathway of oligodendrocyte development. Neuron 5, Boccaccio, G.L., Colman, D.R., 1995. Myelin basic protein mRNA localization and 603–614. polypeptide targeting. J. Neurosci. Res. 42, 277–286. McKinnon, R.D., Smith, C., Behar, T., Smith, T., Dubois-Dalcq, M., 1993. Distinct Calver, A.R., Hall, A.C., Yu, W.P., Walsh, F.S., Heath, J.K., Betsholtz, C., Richardson, W.D., effects of bFGF and PDGF on oligodendrocyte progenitor cells. Glia 7, 245–254. 1998. Oligodendrocyte population dynamics and the role of PDGF in vivo. Neuron Mi, S., Sandrock, A., Miller, R.H., 2008. LINGO-1 and its role in CNS repair. Int. 20, 869–882. J. Biochem. Cell Biol. 40, 1971–1978. Carson, J.H., Kwon, S., Barbarese, E., 1998. RNA trafficking in myelinating cells. Curr. Miller, R.H., 1996. Oligodendrocyte origins. Trends Neurosci. 19, 92–96. Opin. Neurobiol. 8, 607–612. Miller, R.H., Dinsio, K., Wang, R., Geertman, R., Maier, C.E., Hall, A.K., 2004. Carson, J.H., Worboys, K., Ainger, K., Barbarese, E., 1997. Translocation of myelin Patterning of spinal cord oligodendrocyte development by dorsally derived basic protein mRNA in oligodendrocytes requires microtubules and kinesin. Cell BMP4. J. Neurosci. Res. 76, 9–19. Motil. Cytoskeleton 38, 318–328. Miyamoto, Y., Yamauchi, J., Chan, J.R., Okada, A., Tomooka, Y., Hisanaga, S., Tanoue, Chew, L.J., Coley, W., Cheng, Y., Gallo, V., 2010. Mechanisms of regulation of A., 2007. Cdk5 regulates differentiation of oligodendrocyte precursor cells oligodendrocyte development by p38 mitogen-activated protein kinase. through the direct phosphorylation of paxillin. J. Cell Sci. 120, 4355–4366. J. Neurosci. 30, 11011–11027. Miyamoto, Y., Yamauchi, J., Tanoue, A., 2008. Cdk5 phosphorylation of WAVE2 Cicero, S., Herrup, K., 2005. Cyclin-dependent kinase 5 is essential for neuronal cell regulates oligodendrocyte precursor cell migration through nonreceptor tyrosine cycle arrest and differentiation. J. Neurosci. 25, 9658–9668. kinase Fyn. J. Neurosci. 28, 8326–8337. Drake, P.F., Oblinger, M.M., Lasek, R.J., 1985. Synthesis and fast axonal transport of Morfini, G., Szebenyi, G., Brown, H., Pant, H.C., Pigino, G., DeBoer, S., Beffert, U., proteins in the isolated Aplysia nervous system. Brain Res. 332, 47–57. Brady, S.T., 2004. A novel CDK5-dependent pathway for regulating GSK3 Emery, B., Agalliu, D., Cahoy, J.D., Watkins, T.A., Dugas, J.C., Mulinyawe, S.B., Ibrahim, activity and kinesin-driven motility in neurons. EMBO J. 23, 2235–2245. A., Ligon, K.L., Rowitch, D.H., Barres, B.A., 2009. Myelin gene regulatory factor is Narayanan, S.P., Flores, A.I., Wang, F., Macklin, W.B., 2009. Akt signals through a critical transcriptional regulator required for CNS myelination. Cell 138, the mammalian target of rapamycin pathway to regulate CNS myelination. 172–185. J. Neurosci. 29, 6860–6870. Fang, W.Q., Ip, J.P., Li, R., Ng, Y.P., Lin, S.C., Chen, Y., Fu, A.K., Ip, N.Y., 2011. Cdk5- Nishiyama, A., Chang, A., Trapp, B.D., 1999. NG2+ glial cells: a novel glial cell mediated phosphorylation of Axin directs axon formation during cerebral population in the adult brain. J. Neuropathol. Exp. Neurol. 58, 1113–1124. cortex development. J. Neurosci. 31, 13613–13624. Ohshima, T., Ward, J.M., Huh, C.G., Longenecker, G., Veeranna, Pant, H.C., Brady, R.O., Fok-Seang, J., Miller, R.H., 1994. Distribution and differentiation of A2B5+ glial Martin, L.J., Kulkarni, A.B., 1996. Targeted disruption of the cyclin-dependent precursors in the developing rat spinal cord. J. Neurosci. Res. 37, 219–235. kinase 5 gene results in abnormal corticogenesis, neuronal pathology and Fuller, M.L., DeChant, A.K., Rothstein, B., Caprariello, A., Wang, R., Hall, A.K., Miller, perinatal death. Proc. Natl. Acad. Sci. U. S. A. 93, 11173–11178. R.H., 2007. Bone morphogenetic proteins promote gliosis in demyelinating Pandey, J.P., Smith, D.S., 2011. A Cdk5-dependent switch regulates Lis1/Ndel1/ spinal cord lesions. Ann. Neurol. 62, 288–300. dynein-driven organelle transport in adult axons. J. Neurosci. 31, 17207–17219. Fyffe-Maricich, S.L., Karlo, J.C., Landreth, G.E., Miller, R.H., 2011. The ERK2 mitogen- Patrick, G.N., Zhou, P., Kwon, Y.T., Howley, P.M., Tsai, L.H., 1998. p35, the neuronal- activated protein kinase regulates the timing of oligodendrocyte differentiation. specific activator of cyclin-dependent kinase 5 (Cdk5) is degraded by the J. Neurosci. 31, 843–850. ubiquitin-proteasome pathway. J. Biol. Chem. 273, 24057–24064. Gao, L., Macklin, W., Gerson, J., Miller, R.H., 2006. Intrinsic and extrinsic inhibition Pringle, N.P., Richardson, W.D., 1993. A singularity of PDGF alpha-receptor expres- of oligodendrocyte development by rat retina. Dev. Biol. 290, 277–286. sion in the dorsoventral axis of the neural tube may define the origin of the Gilmore, E.C., Herrup, K., 2001. Neocortical cell migration: GABAergic neurons and oligodendrocyte lineage. Development 117, 525–533. cells in layers I and VI move in a cyclin-dependent kinase 5-independent Raff, M.C., Abney, E.R., Miller, R.H., 1984. Two glial cell lineages diverge prenatally in manner. J. Neurosci. 21, 9690–9700. rat optic nerve. Dev. Biol. 106, 53–60. Gilmore, E.C., Ohshima, T., Goffinet, A.M., Kulkarni, A.B., Herrup, K., 1998. Cyclin- Raff, M.C., Lillien, L.E., Richardson, W.D., Burne, J.F., Noble, M.D., 1988. Platelet-derived dependent kinase 5-deficient mice demonstrate novel developmental arrest in growth factor from astrocytes drives the clock that times oligodendrocyte cerebral cortex. J. Neurosci. 18, 6370–6377. development in culture. Nature 333, 562–565. 106 Y. Yang et al. / Developmental Biology 378 (2013) 94–106

Richardson, W.D., Pringle, N., Mosley, M.J., Westermark, B., Dubois-Dalcq, M., 1988. Vermeulen, K., Van Bockstaele, D.R., Berneman, Z.N., 2003. The cell cycle: a review A role for platelet-derived growth factor in normal gliogenesis in the central of regulation, deregulation and therapeutic targets in cancer. Cell Prolif. 36, nervous system. Cell 53, 309–319. 131–149. Robinson, S., Miller, R., 1996. Environmental enhancement of growth factor- Wang, S., Sdrulla, A.D., diSibio, G., Bush, G., Nofziger, D., Hicks, C., Weinmaster, G., mediated oligodendrocyte precursor proliferation. Mol. Cell Neurosci. 8, 38–52. Barres, B.A., 1998. Notch receptor activation inhibits oligodendrocyte differ- fi Rowitch, D.H., 2004. Glial speci cation in the vertebrate neural tube. Nat. Rev. entiation. Neuron 21, 63–75. – Neurosci. 5, 409 419. White, R., Gonsior, C., Bauer, N.M., Kramer-Albers, E.M., Luhmann, H.J., Trotter, J., Sasaki, Y., Cheng, C., Uchida, Y., Nakajima, O., Ohshima, T., Yagi, T., Taniguchi, M., 2012. Heterogeneous nuclear ribonucleoprotein (hnRNP) F is a novel compo- Nakayama, T., Kishida, R., Kudo, Y., Ohno, S., Nakamura, F., Goshima, Y., 2002. nent of oligodendroglial RNA transport granules contributing to regulation of Fyn and Cdk5 mediate semaphorin-3 A signaling, which is involved in regula- myelin basic protein (MBP) synthesis. J. Biol. Chem. 287, 1742–1754. tion of dendrite orientation in cerebral cortex. Neuron 35, 907–920. Xin, M., Yue, T., Ma, Z., Wu, F.F., Gow, A., Lu, Q.R., 2005. Myelinogenesis and axonal Shan, J., Munro, T.P., Barbarese, E., Carson, J.H., Smith, R., 2003. A molecular mechanism recognition by oligodendrocytes in brain are uncoupled in Olig1-null mice. for mRNA traffickinginneuronaldendrites.J.Neurosci.23,8859–8866. J. Neurosci. 25, 1354–1365. Sherman, D.L., Brophy, P.J., 2005. Mechanisms of axon ensheathment and myelin Yip, Y.P., Capriotti, C., Drill, E., Tsai, L.H., Yip, J.W., 2007. Cdk5 selectively affects the growth. Nat. Rev. Neurosci. 6, 683–690. Shimizu, T., Kagawa, T., Wada, T., Muroyama, Y., Takada, S., Ikenaka, K., 2005. Wnt migration of different populations of neurons in the developing spinal cord. – signaling controls the timing of oligodendrocyte development in the spinal J. Comp. Neurol. 503, 297 307. cord. Dev. Biol. 282, 397–410. Zhang, H., Miller, R.H., 1996. Density-dependent feedback inhibition of oligoden- – Tsai, H.H., Macklin, W.B., Miller, R.H., 2006. Netrin-1 is required for the normal drocyte precursor expansion. J. Neurosci. 16, 6886 6895. development of spinal cord oligodendrocytes. J. Neurosci. 26, 1913–1922. Zukerberg, L.R., Patrick, G.N., Nikolic, M., Humbert, S., Wu, C.L., Lanier, L.M., Gertler, F.B., Tsai, L.H., Delalle, I., Caviness Jr., V.S., Chae, T., Harlow, E., 1994. p35 is a neural- Vidal, M., Van Etten, R.A., Tsai, L.H., 2000. Cables links Cdk5 and c-Abl and specific regulatory subunit of cyclin-dependent kinase 5. Nature 371, 419–423. facilitates Cdk5 tyrosine phosphorylation, kinase upregulation, and neurite out- Tyler, W.A., Gangoli, N., Gokina, P., Kim, H.A., Covey, M., Levison, S.W., Wood, T.L., growth. Neuron 26, 633–646. 2009. Activation of the mammalian target of rapamycin (mTOR) is essential for oligodendrocyte differentiation. J. Neurosci. 29, 6367–6378.