Genetically-defined lineage tracing of Nkx2.2-expressing cells in chick spinal cord

Developmental Biology 349 (2011) 504–511

Contents lists available at ScienceDirect

Developmental Biology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / d e v e l o p m e n t a l b i o l o g y

Genomes & Developmental Control

Genetically-defined lineage tracing of Nkx2.2-expressing cells in chick spinal cord Hitoshi Gotoh a,c, Katsuhiko Ono a,b,c, Hirohide Takebayashi a,b,d, Hidekiyo Harada e,f, Harukazu Nakamura e, Kazuhiro Ikenaka a,b,⁎ a

Division of Neurobiology and Bioinformatics, National Institute for Physiological Sciences, Okazaki, Aichi 444-8787, Japan Department of Physiological Sciences, School of Life Science, The Graduate University of Advanced Studies (Sokendai), Miki-cho, Kanagawa 240-0193, Japan Department of Biology, Kyoto Prefectural University of Medicine, Kyoto 603-8334, Japan d Department of Morphological Neural Science, Graduate School of Medical Sciences, Kumamoto University, Kumamoto 860-8556, Japan e Department of Molecular Neurobiology, Graduate School of Life Sciences and Institute of Development, Aging and Cancer, Tohoku University, Miyagi 980-8575, Japan f Division of Cell & Developmental Biology, College of Life Sciences, University of Dundee, Dow Street,DD1 5EH, UK b c

a r t i c l e

i n f o

Article history: Received for publication 24 April 2010 Revised 1 September 2010 Accepted 6 October 2010 Available online 15 October 2010 Keywords: Oligodendrocytes Chick Nkx2.2 Neural tube Lineage tracing

a b s t r a c t In the spinal cord, generation of oligodendrocytes (OLs) is totally dependent on the presence of Olig2, a basic helix-loop-helix transcription factor. However, it also requires Nkx2.2 for its generation, whose expression follows the expression of Olig2. Although it is believed that oligodendrocytes originate from the pMN domain, Nkx2.2 is present in the p3 domain located ventral to the pMN domain. According to recent reports, it is possible that oligodendrocytes are directly derived from the p3 domain in addition to the pMN domain in the chick spinal cord. We examined this hypothesis in this paper. To analyze OL development in the spinal cord, chick embryos are widely used for genetic modification by electroporation or for transplantation experiments, because it is relatively easy to manipulate them compared with mouse embryos. However, genetic modification by electroporation is not appropriate for glial development analyses because glia proliferate vigorously before maturation. In order to overcome these problems, we established a novel method to permanently introduce exogenous gene into a specific cell type. We introduced the CAT1 gene, a murine retroviral receptor, by electroporation followed by injection of murine retrovirus. By using this method, we successfully transduced murine retrovirus into the chick neural tube. We analyzed cell lineage from the p3 domain by restricting CAT1 expression by Nkx2.2-enhancer and found that most of the labeled cells became OLs when the cells were labeled at cE4. Moreover, the labeled OLs were found throughout the white matter in the spinal cord including the most dorsal spinal cord. Thus p3 domain directly generates spinal cord OLs in the chick spinal cord. © 2010 Elsevier Inc. All rights reserved.

Introduction In the developing neural tube, neurons and glial cells are generated from neural progenitor cells that are located in the ventricular area. These progenitor cells are divided into several ‘domain structures’ depending on the expression of transcription factors (Helm and Johnson, 2003; Tanabe and Jessell, 1996). It is known that these domain structures define the subtypes of neurons, such as motoneurons and many classes of interneurons. Recently, it was reported that astrocytes also have positional identity depending on the domain structure, and are subdivided into several subtypes (Hochstim et al., 2008). These observations indicate that progenitor cells are from a heterogenous population through the neurogenic to the gliogenic phase. Oligodendrocytes (OLs) develop from ventral spinal cord as well as more dorsal spinal cord (Cai et al., 2005; Fogarty et al., 2005; ⁎ Corresponding author. Division of Neurobiology and Bioinformatics, National Institute for Physiological Sciences, 5-1 Higashiyama, Myodaiji, Okazaki, Aichi 4448787, Japan. Fax: +81 564 59 5247. E-mail address: [email protected] (K. Ikenaka). 0012-1606/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2010.10.007

Vallstedt et al., 2005). The majority of oligodendrocytes arise from ventral spinal cord and the Olig2 gene, which is present in the pMN domain, is indispensable for their generation (Lu et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002). Nkx2.2, a homeodomain transcription factor, defines the p3 domain that is located ventral to the pMN domain and is also required for the development of OLs in collaboration with Olig2 (Qi et al., 2001; Xu et al., 2000). Oligodendrocyte progenitor cells (OPCs) in the ventral spinal cord are thought to arise from the pMN domain, because of the lineage tracing analysis using Olig2-CreER mice (Masahira et al., 2006) or the absence of OLs in Olig2-deficient mice (Lu et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002). However, it has been reported that the OPC marker, O4 or PDGFRα, is present in the Nkx2.2-expressing p3 domain, especially in chick spinal cord (Soula et al., 2001). Finally, it has been proposed that OPCs migrate out from the ventricular zone as Olig2+/PDGFRα+/Nkx2.2+ cells from the p* domain and as Olig2−/PDGFRα−/Nkx2.2+ cells from the p3 domain in the chick spinal cord. The latter population is considered to be Olig2+ after arriving at the white matter (Fu et al., 2002). These observations suggest that the origin of OPCs in the ventral spinal cord might be

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

heterogeneous and OLs can be generated from both the pMN and p3 domains. However, there is no direct evidence showing that p3derived cells generate mature myelinating oligodendrocytes in vivo. The chick-quail chimera system (Le Douarin, 1973), in which donor cells are directly injected into host organs, is used for lineage tracing analysis of relatively ‘large areas’ such as the dorsal or ventral areas. However, lineage tracing analysis in more restricted domain structures cannot be achieved by this method. Recently it became possible to introduce exogenous genes into chick embryos by electroporation (Nakamura and Funahashi, 2001). The expression of exogenous genes is transient, because introduced plasmids are diluted during cell proliferation. Therefore, it is inappropriate to analyze the glial cell lineage because they proliferate vigorously until their final maturation. To overcome this problem, we designed a novel method by which murine retrovirus and their receptor were used in combination. The infectivity of murine retrovirus is dependent on the expression of cationic amino acid transporter 1 (mCAT1) at the plasma membrane of the cells (Albritton et al., 1989). Chick cells are thought to lack corresponding receptors, thus cannot be infected by murine retrovirus. If mCAT1 is exogenously expressed under the regulation of Nkx2.2-promoter or enhancer, murine retrovirus should transduce cells in the p3 domain that express mCAT1. We found that murine retrovirus transduced chick embryos depending on the exogenous expression of mCAT1. By using this method, we analyzed the cell lineage from Nkx2.2-expressing p3 domain at E4, a late stage of neurogenesis, and found that Nkx2.2-expressing cells in the p3 domain mainly differentiate into mature OLs in various regions of the spinal cord. Our observations presented in vivo evidence that cells that express Nkx2.2 are the source of mature myelinating OLs in chick spinal cord and that they distribute to all areas in the spinal cord.

505

paraformaldehyde in 0.1 M PB. After post fixation overnight with 4% paraformaldehyde/PB, the trunk or the spinal cord was immersed in 20% sucrose/PBS for 1 day. The trunk or the spinal cord was then frozen, and 20 μm sections were cut using a cryostat. In situ hybridization was performed as previously described (Ding et al., 2005). For immunocytochemical analysis, cells were seeded onto a glass coverslip. After transfection, coverslips were washed briefly with PBS, and incubated with 4% paraformaldehyde in 0.1 M PB for 20 min. The sections or the coverslips were incubated with phosphate buffered saline supplemented with 0.1% Triton X-100 (PBS-T), incubated with a blocking buffer composed of 1% BSA, PBS-T for 1 h and then incubated with primary antibody for 16 h at 4 °C. After washing the sections three times with PBS, the sections were processed with the ABC detection kit (Vector, USA) according to the manufacturer's protocol. The signals were visualized by peroxidase reaction with 3,3′-diaminobenzidine as the substrate. For fluorescent immunohistochemistry, brain sections were incubated with a primary antibody, followed with AlexaFluor 488-, AlexaFluor 543-, or AlexaFluor 633-labeled secondary antibody with DAPI, and observed using a fluorescence microscope (DP70 and DP71; Olympus, Japan) or confocal microscope (FV-1000; Olympus, Japan). The primary antibodies used in this study are as follows: anti-NeuN, antiphosphohistone H3 (Millipore, USA), anti-GFAP (DAKO, USA), antiGSTπ, anti-Myc (MBL, Japan), anti-MBP (Nichirei, Japan), rabbit polyclonal anti-GFP (Invitrogen, USA), rat monoclonal anti-GFP (Nacalai Tesque, Japan), anti-BrdU (BD Pharmingen, USA), and anticleaved caspase 3 (Cell Signaling Technology, USA) antibodies. Rabbit polyclonal antibody to GLAST was kindly provided by Dr. Masahiko Watanabe, monoclonal antibody to Nkx2.2 (clone 74.5A5) was obtained from Developmental Studies Hybridoma Bank (Univ. of Iowa, USA), and monoclonal anti-PLP antibody (clone AA3) was used as previously described (Yamada et al., 1999).

Materials and methods Vectors A cDNA encoding the mCAT1 was amplified using myc-tagged reverse primers, resulting in C-terminal Myc-tagged mCAT1 and cloned into pCAG vector or p3T vector (MoBiTech, Germany). The enhancer region of Nkx2.2 (Lei et al., 2006) was amplified and cloned into pGL3-basic vector (Promega, USA). The luciferase sequence was replaced with mCAT1-myc, and β-globin minimal promoter was inserted between Nkx2.2 enhancer and mCAT1. pGAP-GFP that expresses membrane-targeted EGFP was used as previously described (Ono et al., 2004). A retroviral vector pLNRGW was obtained from Dr. Teoan Kim (Kwon et al., 2004), and vector encoding cPDGFRα was kindly provided by Dr. J-L Thomas (Spassky et al., 1998). Animals and cell culture Fertilized white leghorn eggs were obtained from the Ghen Corporation (Gifu, Japan) and incubated at 38 °C. Embryonic stages of chicks were determined according to Hamburger and Hamilton (Hamburger and Hamilton, 1951). For BrdU injection, 25 mg/kg of BrdU was injected into the egg yolk. The DF-1 chicken fibroblast cell line and NIH3T3 mouse fibroblast cell line were maintained in DMEM containing 10% fetal bovine serum, streptomycin and penicillin. mCAT1-expressing plasmids were transfected into cultured cells using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer's instructions. In situ hybridization, immunohistochemistry, and immunocytochemistry The trunk region of early chick embryos was directly immersed into 4% paraformaldehyde in 0.1 M phosphate buffer (PB). Spinal cords of late chick embryos were removed after perfusion with 4%

Preparation of EGFP-expressing high-titer retrovirus Preparation of EGFP-expressing retrovirus producing cells or preparation of high-titer retrovirus was performed as previously described (Nanmoku et al., 2003). Briefly, the retrovirus producing cell line, ΨMP34, was transfected with pLNRGW and treated with G418 (Nacalai Tesque, Japan) for 1 week. Stable clones were selected by a limited dilution method and stable clones that produced high-titer retrovirus were selected. The selected clones were cultured in 250 ml of DMEM supplemented with 10% FBS at 30 °C. The supernatant of virus-producing cells was concentrated 1/10,000 by two-rounds of centrifugation at 6000 g for 16 h at 4 °C and dissolved in Hanks' balanced salt solution (HBSS). For virus titration, NIH3T3 cells were seeded in 24-well plates at 1 × 104 cells/well together with serial dilutions of virus solution in the presence of 8 μg/ml of polybrene. The EGFP-positive cells were counted 48 h after infection, and the titer was estimated according to the following formula: titer of retrovirus (cfu/ml) = number of EGFP-positive cells/virus volume (ml) × 4 (replication factor of NIH3T3). In ovo electroporation and virus injection Plasmids indicated in each Figure were dissolved in distilled water at a concentration of 1.0 μg/μl for pCAG-mCAT1 or 500 ng/μl for pNkx2.2-mCAT1 and mixed with a 1/10 volume of 0.5% fast green. Approximately 0.1 μl of the mixed solution was injected into the neural tube of HH stage 17 to 19 chicken embryos (3 days after starting the incubation). Needle type electrodes were placed near the embryo, and a 30 V, 30 ms pulse was applied three times using an electronic stimulator (SEN-3310; Nihon Kohden, Japan). For the retroviral injection, approximately 0.1 μl of virus solution was injected into neural tube 24 h after electroporation unless otherwise indicated.

506

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

Results Exogenous expression of mCAT1 into chick cells confers infectivity to murine retrovirus To transduce chick cells with murine retrovirus, we introduced its receptor, mCAT1, into chick fibroblast cell line, DF-1. Twenty-four hours after transfection, cells were incubated with EGFP retrovirus. As a positive control, mouse fibroblast cell line, NIH3T3, was incubated with retrovirus (Figs. 1 A–C). GFP-positive cells were observed in NIH3T3 (Fig. 1 C), whereas there are no GFP-positive cells in DF-1 cells without mCAT1 transfection (Fig. 1 F). When cells were transfected with mCAT1-

myc (Figs. 1 H, I), EGFP-positive cells were present in mCAT1-expressing (Myc-positive) cells, suggesting that the infectivity of murine retrovirus to chick cells is dependent on the presence of exogenous mCAT1. To check further whether mCAT1 introduction confers infectivity to mCAT1 in vivo, we introduced an mCAT1-expressing construct into the chick neural tube. Mock or pCAG-mCAT1-myc that express mCAT1 under the regulation of cytomegalovirus enhancer and chick β-actin promoter was electroporated into E3 chick embryos. Solution containing murine retrovirus was injected 24 h after electroporation. We analyzed embryos at E14, and found that EGFP-positive cells were observed in one side of the neural tube when mCAT1 was introduced by electroporation (Figs. 1 L, M), whereas EGFP signals were not observed

Fig. 1. The infectivity of murine retrovirus into chick cells is dependent on the exogenous expression of mCAT1 both in vitro and in vivo. NIH3T3, a mouse fibroblast cell line, was transduced by murine retrovirus and 48 h after transduction, expression of EGFP was observed by immunostaining with anti-GFP (C) antibody. DF-1, a chick fibroblast cell line, was transfected with pCAG (mock; D–F) or pCAG-mCAT1-myc (G–I). After 24 h, cells were transduced by retrovirus. Expression of EGFP (F and I) or mCAT1-Myc (E and H) was examined by immunostaining with anti-GFP or anti-Myc antibodies, respectively. Bright field images of cells are shown in A, D, and G (Scale bar: 100 μm). Mock (J and K) or mCAT1 (L and M) was introduced by electroporation at HH stages 17–19 (E3). Solution containing EGFP-expressing retrovirus was infused into the neural tube 24 h after electroporation. Embryos were grown to E14 or E9, and ectopic expression of EGFP was examined by direct fluorescent microscopy (K, M) (Scale bar: 2 mm). Sections of E9 chick neural tube were immunostained by anti-GFP antibody (N and O) (Scale bar: 500 μm).

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

in the control spinal cord (Figs. 1 J, K). We further confirmed that infection of EGFP retrovirus is dependent on mCAT1 in spinal cord sections. At E9, there were no EGFP-positive cells among the control (n = 6; Fig. 1 N). In the mCAT1-transfected embryos, EGFP-positive cells were observed only in one side of the neural tube (n = 8; Fig. 1 O), suggesting that murine retrovirus can also infect cells that express mCAT1 in vivo. Effects of mCAT1 introduction in ovo and effects of virus-titer on the labeling efficiency It was reported that deficiency of mCAT1 in mice leads to perinatal lethality (Perkins et al., 1997). Therefore, we next examined whether

507

exogenous expression of mCAT1 results in abnormal cell proliferation or apoptosis. To analyze whether mCAT1-expressing cells show altered BrdU incorporation, chick embryos were electroporated using pCAG–mCAT1-myc (Figs. 2 D–F) or pGAP-GFP as a control (Figs. 2 A–C). BrdU was injected into chick embryos 2 h before harvesting them. At any time point, the number of cells incorporating BrdU was not altered among the nonelectroporated side, GFPelectroporated side and the mCAT1-electroporated side. We calculated the ratio of GFP/BrdU double-positive cells to total GFP-positive cells, and calculated the ratio of mCAT1-Myc/BrdU double-positive cells to total mCAT1-Myc-positive cells (Fig. 2 G), and found that forced expression of mCAT1 does not significantly affect the ability to incorporate BrdU. Next, we analyzed mitotic cells by using anti-

Fig. 2. Effect of mCAT1 expression on proliferation rate or apoptosis. Chick embryos at HH stage 17–19 (E3) were electroporated with pGAP-GFP (A–C, H–J, O–P) or pCAG-mCAT1 (D–F, K–M, Q–R). BrdU was injected into the egg yolk 2 h before collecting embryos. Figure A–F, H–M, and O–R show representative images of sections 24 h after electroporation (Scale bar: 200 μm). BrdU was immunostained with anti-BrdU antibody (B, E) together with anti-GFP antibody (A) or anti-Myc antibody (D). The ratio of GFP/BrdU doublepositive cells per total GFP-positive cells, or the ratio of mCAT1-myc/BrdU double-positive cells per total mCAT1-myc-positive cells was calculated for each experimental time point (G). Data represent means ± SD of five chick embryos. Mitotic cells were examined by immunostaining for phosphohistone H3 (I, L). The ratio of phosphohistone H3-positive electroporated cells per total electroporated cells was calculated (N). Data represent means ± SD of four chick embryos. Cleaved caspase 3-positive apoptotic cells were examined in GFP-electroporated embryos (O, P) or mCAT1-electroporated embryos (Q, R).

508

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

phosphohistone H3 antibody (Figs. 2 H–M). Introduction of mCAT1 did not induce any ectopic phosphorylated histone H3 immunoreactivity as compared with introduction of GFP. Quantitative data show no significant difference in the mitotic index of cells between GFPelectroporated embryos and mCAT1-electroporated ones (Fig. 2 N). Because BrdU is incorporated into newly synthesized DNA in S phase cells and phosphohistone H3 is a marker for the M phase, our observation suggests that exogenous expression of mCAT1 does not affect the cell cycle in chick embryos. We further analyzed apoptosis of electroporated cells by immunostaining using anti-cleaved caspase 3 antibody or DAPI staining for pyknotic cells. There was no excess cleaved caspase 3 immunoreactivity (Figs. 2 O–R) or pyknotic cells in mCAT1-positive cells (data not shown), suggesting that introduction of mCAT1 does not show any effects on apoptosis. These results indicate that mCAT1-electroporated cells can be transduced with murine retrovirus without any prominent effects on neural progenitor cells. It is known that high-titer virus is required for transduction in vivo as compared with that in vitro. Thus, we determined the titer of retrovirus required for transduction of murine retrovirus in ovo. We electroporated 1.0 μg/μl of pCAG-mCAT1 followed by injection with approximately 0.1 μl murine retrovirus at the indicated titer. Fortyeight hours after injection with murine retrovirus, we prepared 20 μm cryosections, and counted the maximum number of GFP-positive cells per 20 μm sections (Figs. 3 A–D). As indicated in Figs. 3C and D, 1 × 106 cfu/ml of murine retrovirus can transduce chick embryo although the number of transduced cells was small. Although expression level of mCAT1 in each cells is dependent on the promoter or enhancer sequence, these data demonstrated that over 106 cfu/ml titer is required for our gene modification method. Transduced chick cells with murine retrovirus have ability to differentiate into any neural cell type To analyze whether infected chick cells with murine retrovirus could differentiate into major neural cell types, namely neurons, oligodendrocytes, and astrocytes, embryos were electroporated with pCAG–mCAT1 at E3 (HH stage 17–19) followed by injection with murine retrovirus at E4. Embryos were grown to stage E14 when the expression of major cell type markers can be observed. We performed immunohistochemical analysis using anti-NeuN (Figs. 4 A–C), anti-GSTπ (Figs. 4 D–F), or anti-GLAST

antibodies (Figs. 4 G–I), together with anti-GFP antibody. GFP/NeuN double-positive neurons (Fig. 4 C), or GFP/GSTπ double-positive mature oligodendrocytes (Fig. 4 F), or GFP/GLAST double-positive mature astrocytes (Fig. 4 I) could be observed in spinal cord sections at E14, suggesting that chick progenitor cells that are transduced by murine retrovirus differentiated into various neural lineage cells and that there is no preferential expression of GFP in a specific cell type. mCAT1 expression under the regulation of Nkx2.2 enhancer region enables lineage tracing analysis from the p3 domain We next asked whether restricted expression of mCAT1 under the regulation of the Nkx2.2-enhancer enables us to analyze cell lineage from the p3 domain. The enhancer region of Nkx2.2 specified mCAT1 expression in a restricted area in the neural tube (Fig. 5 A), whereas DsRed expressed under the ubiquitous promoter was widely expressed along the entire neural tube (Fig. 5 B). Among 10 embryos that we examined, mCAT1 was expressed only in Nkx2.2-expressing cells (Figs. 5 C, D), namely p3 domain cells in all embryos. Since mCAT1 is a membrane protein and mCAT1 signal detected by Myc antibody was observed at the plasma membrane, mCAT1 that is yet unincorporated into the plasma membrane could not be detected by immunohistochemistry. To further confirm that the enhancer region of Nkx2.2 specifies mCAT1 expression only in the p3 domain, we analyzed localization of mCAT1 mRNA in the spinal cord section by in situ hybridization. When we introduced mCAT1-myc at the concentration of 500 ng/μl, we did not detect nonspecific signals and detected mCAT1 mRNA only in the Nkx2.2-expressing p3 domain. However, when higher concentration of Nkx2.2-mCAT1 plasmid (1.0 μg/μl) was introduced, nonspecific expression of mCAT1 mRNA was observed outside the p3 domain (data not shown). These results suggest that the enhancer region of Nkx2.2 specified mCAT1 expression precisely to the p3 domain when constructs were electroporated at an appropriate concentration. Cells in the p3 domain at E4 mainly differentiate into oligodendroglial lineage We then used this system to analyze cell lineage of p3 origin. pNkx2.2-mCAT1 was electroporated at E3 (HH stage 17–19) followed by injection with murine retrovirus at E4, and embryos were grown to

Fig. 3. Effect of virus-titer on cell labeling efficiency. Chick embryos were electroporated with pCAG-CAT1 at E3. Twenty-four hours after electroporation, approximately 0.1 μl of 1 × 1010 cfu/ml (A), 1 × 108 cfu/ml (B; bar indicates 200 μm), or 1 × 106 cfu/ml virus solution (C) was injected into the neural tube. Forty-eight hours after injection, sections were examined by immunohistochemistry using anti-GFP antibody. The maximum number of EGFP-positive cells per 20 μm section was counted (D). Data represent means ± SD of eight chick embryos.

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

509

Fig. 4. Chick cells infected with murine retrovirus can differentiate into any neural cell lineage. pCAG-mCAT1 was introduced by electroporation at HH stages 17–19 (E3). Solution containing EGFP-expressing retrovirus was infused into the neural tube 24 h after electroporation. Embryos were grown to stage E14. Spinal cord sections at E14 were doubleimmunostained with anti-NeuN (B), GSTπ (E), or GLAST (H) together with anti-GFP (A, D, and G). GFP-labeled cells were merged with NeuN-positive neurons (C), GSTπ-positive oligodendrocytes (F; arrowheads), or GLAST-positive astrocytes (I; arrow), respectively.

Fig. 5. Nkx2.2-enhancer specifies mCAT1 expression in the p3 domain. pNkx2.2-mCAT1-myc was electroporated at E3 together with pCAG-DsRed. Expression of mCAT1-Myc was examined by immunohistochemical staining with anti-Myc antibody (A, C) 24 h after electroporation. The expression of the electroporation marker, DsRed (B), or endogenous expression of Nkx2.2 (D) was examined. The bars indicate 200 μm. (E) Expression of mCAT1 mRNA was examined by in situ hybridization (purple) followed by immunohistochemistry using anti-Nkx2.2 (brown).

stage E7 (Fig. 6 A), E9 (Fig. 6 B) or E15 (Figs. 6 C, D). At E7, we observed GFP-positive cell reached the white matter. We characterized this cell and found that it expressed PDGFRα (Fig. 6 A), suggesting that Nkx2.2lineage cells differentiated into oligodendrocyte progenitor cells. At this age, no GFP-positive cells were observed in the dorsal spinal cord. At E9, we observed neuron-like cells in the gray matter or glia-like cells in the white matter. As compared with embryos at E7, GFP-positive cells were present in lateral to dorsal area of the spinal cord. We tried to characterize the GFP-positive cell type, and found that at E9, NeuNpositive neurons were generated (Figs. 6 E–G) whereas cells in the white matter did not express any mature cell type marker, such as MBP or GFAP (data not shown). We further examined the cell type of GFPpositive cells at a later developmental stage (E15). At this stage, there were GFAP-positive mature astrocytes (Figs. 6 H–J). It is known that the majority of spinal cord oligodendrocytes form a myelin sheath along the longitudinal axis, thus we immunostained longitudinal sections with mature oligodendrocytic markers. We observed many MBP- (Figs. 6 K–M) or PLP-positive (Figs. 6 N–P) mature myelinating oligodendrocytes. Morphology of MBP-positive oligodendrocytes was different from PLPpositive oligodendrocytes. Mature oligodendrocytes have many processes as compared with immature oligodendrocytes, thus it is possible that PLP-positive oligodendrocytes are relatively immature. Next, we examined the distribution of each cell type by double immunostaining using anti-NeuN, anti-GFAP, or anti-GSTπ antibodies, another mature oligodendrocytic marker. Fig. 6 Q shows schematic diagrams of the distribution and cell type of GFP-positive cells of four embryos examined. Neurons and astrocytes were observed only in the ventral spinal cord, whereas mature oligodendrocytes could be observed not only in the ventral spinal cord, but also in the dorsal spinal cord. Approximately 50% of GFP-positive cells were negative for these markers (Fig. 6 Q). These cells might be immature glial cells because

510

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

Fig. 6. Characterization of Nkx2.2-lineage cells and their distribution in the spinal cord. Embryos were electroporated with pNkx2.2-mCAT1-myc at E3, followed by injection with EGFP-expressing retrovirus at E4. Embryos were grown to E7 (A), E9 (B) or E15 (C, D), and spinal cord sections were stained with anti-GFP antibody. At E7, expression of PDGFRa was examined by in situ hybridization (purple) followed by immunostaining with anti-GFP (brown). Merged cell was indicated by an arrow head. At E9, uncharacterized cells in the lateral to dorsal white matter are shown in the boxed area. Fig. 5 C shows a representative image of GFP-positive cells in the ventral spinal cord. GFP-positive cells were also observed in the white matter of the dorsal spinal cord (D). Bar indicates 500 μm. Embryos were electroporated with pNkx2.2-mCAT1-myc at E3, followed by injection with EGFP-expressing retrovirus at E4. Coronal sections at E9 (E–G) or E15 (H–J) were double-immunostained using anti-NeuN (F) or anti-GFAP (I) antibodies together with anti-GFP antibody (E, H). G and J show merged images of E and F, or H and I, respectively. A GFP/GFAP double-positive cell is indicated by an arrow head (J). Longitudinal sections at E15 were doubleimmunostained using anti-MBP (L) or anti-PLP (O) antibodies together with anti-GFP antibodies (K, N). M and P show merged images of K and L, or N and O, respectively. MBP- or PLP-positive myelinated axons are indicated by arrows (M and P). Bar indicates 40 μm. In Fig. Q, adjacent sections were double-immunostained using anti-NeuN, anti-GSTπ, or antiGFAP antibodies together with anti-GFP antibodies. GFP-positive cells of one chick embryo were superimposed and marked with dots on a schematic diagram. Four independent experiments were performed. NeuN/GFP double-positive neurons are shown in white, GSTπ/GFP double-positive oligodendrocytes are shown in blue, and GFAP/GFP double-positive astrocytes are shown in red. Green dots show cells that were negative for all three differentiated cell markers. Among them, oligodendrocyte-like cells were determined by morphology and shown by blue dots (Bar: 200 μm). The Table indicates the proportion of each cell type observed in each embryo (n = 4).

they are present in the white matter, and negative for the neuronal marker, NeuN. We quantified immature oligodendrocytes according to their morphological distinction as shown by gray dots in Fig. 6 Q. Our observations demonstrate that we can analyze the lineage of p3 domain cells in the chick spinal cord, and that cells in the p3 domain at E4 differentiate mainly into oligodendrocytes that spread widely throughout the dorsal-to-ventral spinal cord. Discussion Chick embryonic spinal cords are widely used for developmental studies. As compared with mice, transgenic chickens are difficult to produce. Recently, in ovo electroporation has been widely used for overexpression or downregulation of a certain gene, which enables us to examine the in vivo function of a certain gene in a relatively short time. By using chick embryos, it is relatively easy to analyze the mechanisms of development, especially in the neural tube, compared with relatively difficult analyses in mouse. However, as neural progenitor cells or glial

cells are proliferative after electroporation, introduced plasmids are often diluted as they divide, which makes it impractical to perform the lineage tracing analysis or to observe the long-term effect of overexpression of a certain gene. Both in mice and in chick, retroviral vectors are widely used for the lineage tracing analysis because they allow permanent expression of an exogenous gene. To restrict the retroviral transduction to a subset of cells in mice, the combination of transgenically expressed TVA (receptor for chick retrovirus) and injection with chick retrovirus has been used (Fisher et al., 1999). We established a novel gene modification method in chick embryos that consists of exogenous expression of murine retroviral receptor followed by injection with murine retrovirus. Our observations indicate that murine retrovirus can transduce chick cells in vivo dependent on the expression of exogenous mCAT1 (Fig. 1). Moreover, mCAT1 expression did not affect proliferation, cell death, or ability to differentiate into various neural lineages (Figs. 2, 3). These findings suggest that our method can be used for lineage tracing analysis in ovo. As compared with transgenic mice strategies, our method has advantages including the relatively short time required for experiments in the neural

H. Gotoh et al. / Developmental Biology 349 (2011) 504–511

tube. We next applied mCAT1 expression under the regulation of Nkx2.2enhancer and analyzed the cell lineage from the Nkx2.2-positive p3 domain. We confirmed that Nkx2.2-enhancer precisely specifies mCAT1 expression to the p3 domain, although nonspecific expression of mCAT1 was detected when high concentration of mCAT1-expressing construct was introduced, suggesting that careful examination of cell-specific promoter or enhancer are required for application of our geneticallydefined lineage tracing method. We labeled cells at E4 when the border between Olig2-expressing pMN and Nkx2.2-expressing p3 domains is quite clear with at most a single cell layer of double-positive cells (Fu et al., 2002; data not shown). Labeled cells at E4 mainly differentiated into oligodendrocytes, but relatively rarely to ventral neurons and astrocytes. Moreover, mature oligodendrocytes that developed from the p3 domain in chick embryo distributed in the white matter widely along the dorsoventral axis. Because it has been reported that OPC developed from the ventral progenitor area to show extensive migration through the ventral to dorsal area (Ono et al., 1995), our analysis of p3-lineage cells is consistent with this observation. At E15, MBP-positive oligodendrocytes had many more branched processes than PLP-positive oligodendrocytes, although some PLP-positive oligodendrocytes showed similar morphology to MBP-positive oligodendrocytes. This difference may be due to the reactivity of anti-PLP antibody (clone AA3). It was reported that AA3 antibody reacts with both dm20 and PLP, and that AA3 shows immunoreactivity from E6 chick embryo (Perez Villegas et al., 1999). These observations suggest that relatively immature oligodendrocytes as well as more mature oligodendrocytes become AA3-positive. Lineage analysis at a single cell resolution could solve whether p3 domain cells share neuronal, astrocytic, or oligodendrocytic fate. However, we could not analyze at a single cell resolution, but observed small and limited number of cells transduced with murine retrovirus (data not shown). For example, we observed only six GFP-positive cells, three of which expressed PDGFRα, in one embryo shown in Fig. 6 A. These observations support that most of the p3 domain cells at E4 differentiate into oligodendrocyte lineage. It has been demonstrated using Olig2CreER mouse that the Olig2-expressing pMN domain (including the Olig2+/Nkx2.2+ P* domain) generates myelinating oligodendrocytes in all areas of the spinal cord, and our results indicate that the p3 domain also contributes to the generation of myelinating oligodendrocytes. This may be in contradiction to the finding that Olig2 expression is indispensable for generation of oligodendrocyte precursor cells. Although this observation in the chick spinal cord might be due to species difference between mice and chick, it has been shown that Nkx2.2expressing cells had once expressed Olig2 and altered their expression from Olig2 to Nkx2.2 in mice (Dessaud et al., 2007). Thus this early Olig2 expression may be needed for the Nkx2.2+/Olig2−cells to become oligodendrocytes. In conclusion, we established a novel geneticallydefined lineage tracing method that is useful to analyze cell lineage from a specific cell domain in the chick spinal cord. Moreover, we showed in vivo evidence that Nkx2.2-expressing cells in the p3 domain differentiate into mature oligodendrocytes and distribute widely in the white matter of the chick spinal cord. Acknowledgment We thank Dr. Masahiko Watanabe (Hokkaido University, Japan) for providing the GLAST antibody, Dr. Teon Kim (Catholic University of Daegu, Korea) for providing the retroviral vector, and Dr. J-L Thomas (INSERM, France) for providing chick PDGFRα vector. This work was supported by a Grant-in-Aid for Scientific Research (A) by JSPS. References Albritton, L.M., Tseng, L., Scadden, D., Cunningham, J.M., 1989. A putative murine ecotropic retrovirus receptor gene encodes a multiple membrane-spanning protein and confers susceptibility to virus infection. Cell 57, 659–666.

511

Cai, J., Qi, Y., Hu, X., Tan, M., Liu, Z., Zhang, J., Li, Q., Sander, M., Qiu, M., 2005. Generation of oligodendrocyte precursor cells from mouse dorsal spinal cord independent of Nkx6 regulation and Shh signaling. Neuron 45, 41–53. Ding, L., Takebayashi, H., Watanabe, K., Ohtsuki, T., Tanaka, K.F., Nabeshima, Y., Chisaka, O., Ikenaka, K., Ono, K., 2005. Short-term lineage analysis of dorsally derived Olig3 cells in the developing spinal cord. Dev. Dyn. 234, 622–632. Dessaud, E., Yang, L.L., Hill, K., Cox, B., Ulloa, F., Ribeiro, A., Mynett, A., Novitch, B.G., Briscoe, J., 2007. Interpretation of the sonic hedgehog morphogen gradient by a temporal adaptation mechanism. Nature 450, 717–720. Fisher, G.H., Orsulic, S., Holland, E., Hively, W.P., Li, Y., Lewis, B.C., Williams, B.O., Varmus, H.E., 1999. Development of a flexible and specific gene delivery system for production of murine tumor models. Oncogene 18, 5253–5260. Fogarty, M., Richardson, W.D., Kessaris, N., 2005. A subset of oligodendrocytes generated from radial glia in the dorsal spinal cord. Development 132, 1951–1959. Fu, H., Qi, Y., Tan, M., Cai, J., Takebayashi, H., Nakafuku, M., Richardson, W., Qiu, M., 2002. Dual origin of spinal oligodendrocyte progenitors and evidence for the cooperative role of Olig2 and Nkx2.2 in the control of oligodendrocyte differentiation. Development 129, 681–693. Hamburger, V., Hamilton, H.L., 1951. A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92. Helm, A.W., Johnson, J.E., 2003. Specification of dorsal spinal cord interneurons. Curr. Opinion. Neurobiol. 13, 42–49. Hochstim, C., Deneen, B., Lukaszewicz, A., Zhou, Q., Anderson, D.J., 2008. Identification of positionally distinct astrocyte subtypes whose identities are specified by a homeodomain code. Cell 133, 510–522. Kwon, M.S., Koo, B.C., Choi, B.R., Lee, H.T., Kim, Y.H., Ryu, W.S., Shim, H., Kim, J.H., Kim, N.H., Kim, T., 2004. Development of transgenic chickens expressing enhanced green fluorescent protein. Biochem. Biophys. Res. Commun. 320, 442–448. Le Douarin, N.M., 1973. A biological cell labelling technique and its use in experimental embryology. Dev. Biol. 30, 217–222. Lei, Q., Jeong, Y., Misra, K., Li, S., Zelman, A.K., Epstein, D.J., Matise, M.P., 2006. Wnt signaling inhibitors regulate the transcriptional response to morphogenetic Shh-Gli signaling in the neural tube. Dev. Cell 11, 325–337. Lu, Q.R., Sun, T., Zhu, Z., Ma, N., Garcia, M., Stiles, C.D., Rowitch, D.H., 2002. Common developmental requirement for Olig function indicates a motor neuron/oligodendrocyte connection. Cell 109, 75–86. Masahira, N., Takebayashi, H., Ono, K., Watanabe, K., Ding, L., Furusho, M., Ogawa, Y., Nabeshima, Y., Alvarez-Buylla, A., Shimizu, K., Ikenaka, K., 2006. Olig2-positive progenitors in the embryonic spinal cord give rise not only to motoneurons and oligodendrocytes, but also to a subset of astrocytes and ependymal cells. Dev. Biol. 293, 358–369. Nakamura, H., Funahashi, J., 2001. Introduction of DNA into chick embryos by in ovo electroporation. Methods 24, 43–48. Nanmoku, K., Kawano, M., Iwasaki, Y., Ikenaka, K., 2003. Highly efficient gene transduction into the brain using high-titer retroviral vectors. Dev. Neurosci. 25, 152–161. Ono, K., Bansal, R., Payne, J., Rutishauser, U., Miller, R.H., 1995. Early development and dispersal of oligodendrocyte precursors in the embryonic chick spinal cord. Development 121, 1743–1754. Ono, K., Yasui, Y., Ikenaka, K., 2004. Lower rhombic lip-derived cells undergo transmedian tangential migration followed by radial migration in the chick embryo brainstem. Eur. J. Neurosci. 20, 914–922. Perez Villegas, E.M., Oliver, C., Spassky, N., Poncet, C., Cochard, P., Zalc, B., Thomas, J.-L., Martinez, S., 1999. Early specification of oligodendrocytes in the chick embryonic brain. Dev. Biol. 216, 98–113. Perkins, C.P., Mar, V., Shutter, J.R., del Castillo, J., Danilenko, D.M., Medlock, E.S., Ponting, I.L., Graham, M., Stark, K.L., Zuo, Y., Cunningham, J.M., Bosselman, R.A., 1997. Anemia and perinatal death result from loss of the murine ecotropic retrovirus receptor mCAT-1. Genes Dev. 11, 914–925. Qi, Y., Cai, J., Wu, Y., Wu, R., Lee, J., Fu, H., Rao, M., Sussel, L., Rubenstein, J., Qiu, M., 2001. Control of oligodendrocyte differentiation by the Nkx2.2 homeodomain transcription factor. Development 128, 2723–2733. Soula, C., Danesin, C., Kan, P., Grob, M., Poncet, C., Cochard, P., 2001. Distinct sites of origin of oligodendrocytes and somatic motoneurons in the chick spinal cord: oligodendrocytes arise from Nkx2.2-expressing progenitors by a Shh-dependent mechanism. Development 128, 1369–1379. Spassky, N., Goujet-Zalc, C., Parmantier, E., Olivier, C., Martinez, S., Ivanova, A., Ikenaka, K., Macklin, W., Cerruti, I., Zalc, B., Thomas, J.-L., 1998. Multiple restricted origin of oligodendrocytes. J. Neurosci. 18, 8331–8343. Takebayashi, H., Nabeshima, Y., Yoshida, S., Chisaka, O., Ikenaka, K., Nabeshima, Y., 2002. The basic helix-loop-helix factor Olig2 is essential for the development of motoneuron and oligodendrocyte lineages. Curr. Biol. 12, 1157–1163. Tanabe, T., Jessell, T.M., 1996. Diversity and pattern in the developing spinal cord. Science 274, 1115–1123. Vallstedt, A., Klos, J.M., Ericson, J., 2005. Multiple dorsoventral origins of oligodendrocyte generation in the spinal cord and hindbrain. Neuron 45, 55–67. Yamada, M., Ivanova, A., Yamaguchi, Y., Lees, M.B., Ikenaka, K., 1999. Proteolipid protein gene product can be secreted and exhibit biological activity during early development. J. Neurosci. 19, 2143–2151. Xu, X., Cai, J., Fu, H., Wu, R., Qi, Y., Modderman, G., Liu, R., Qiu, M., 2000. Selective expression of Nkx-2.2 transcription factor in chicken oligodendrocyte progenitors and implications for the embryonic origin of oligodendrocytes. Mol. Cell. Neurosci. 6, 740–753. Zhou, Q., Anderson, D.J., 2002. The bHLH transcription factors OLIG2 and OLIG1 couple neuronal and glial subtype specification. Cell 109, 61–73.