SEARCH

SEARCH BY CITATION

Keywords:

  • basic helix-loop-helix (bHLH);
  • commissural interneuron;
  • cell migration;
  • dorsal midline cell;
  • in situ hybridization;
  • electroporation

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We examined the migration and differentiation of cells expressing Olig3, a basic helix-loop-helix transcriptional factor, in the developing spinal cord. Distribution of Olig3 lineage cells was demonstrated with in situ hybridization and X-gal staining in an Olig3-lacZ knock-in mouse. Olig3-positive cells first appeared in the dorsal spinal cord, except for the roof plate. Some of the dorsal Olig3 lineage cells co-expressed Islet1/2, Math1, or Brn3a, markers for dorsal interneuron. LacZ-positive cells were observed in the ventral-most part of the E10.5 spinal cord, suggesting that some dorsal Olig3 lineage cells migrate into the ventral-most part by E10.5. Ventral-ward migration of dorsal cells and contribution to commissural interneurons were substantiated by electroporation of EGFP expression plasmid in the dorsal spinal cord of chick embryo. Dorsal midline cells were also LacZ-positive during development. These findings suggest that dorsal Olig3 cells contribute to dorsal midline cells and commissural interneurons at intermediate and ventral levels. Developmental Dynamics 234:622–632, 2005. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The central nervous system (CNS) is composed of various types of cells including macroglial and neuronal cells. Such cells are specified for positions along the anteroposterior and dorsoventral axes, which are regulated initially by soluble factors with morphogen activity, such as sonic hedgehog, bone morphogenetic proteins, fibroblast growth factors, and Wnts, followed by expression of homeobox and basic helix-loop-helix (bHLH) transcription factors (Tanabe and Jessell,1996). Following cell type specification, cells begin to migrate toward their final position where they function. Migration and maturation of progenitor cells thus establish the functional architecture of the CNS (Hatten and Heintz,1995; Marín and Rubenstein,2003).

The three Olig genes, Olig1, Olig2, and Olig3, comprise a subfamily of bHLH transcription factors (Takebayashi et al.2000). In the developing spinal cord, Olig1 and Olig2 are initially expressed in the ventral ventricular zone and demarcate the pMN domain, and are subsequently distributed throughout the spinal cord (Lu et al.,2000; Takebayashi et al.,2000; Zhou et al.,2000). While Olig1 knockout mice exhibit delayed development of oligodendrocytes and hypomyelination (Lu et al.,2002), mice lacking the Olig2 gene are devoid of motor neurons and oligodendrocytes in the spinal cord (Lu et al.,2002; Takebayashi et al.,2002b; Zhou and Anderson,2002). The Olig1 and Olig2 genes thus play crucial roles in induction of motoneurons and oligodendrocyte precursor cells as well as differentiation of oligodendrocyte progenitor cells (Mizuguchi et al.,2001; Novitch et al.,2001; Zhou et al.,2001; Lu et al.,2002; Zhou and Anderson2002).

In contrast, the Olig3 gene is initially expressed in the dorsal-most portion of the spinal cord, except for the roof plate. Subsequently, Olig3 expression appears in the ventral half of the spinal cord; in fetal mouse ventral spinal cord, Olig3 is expressed in the border regions between the ventricular zone and mantle layer at the level of the p3, p2, and p0 domains, but not pMN domain (Takebayashi et al.,2002a). This pattern of expression in early spinal cord development strongly suggests that the functions of Olig3 differ from those of Olig1 and Olig2. However, unlike Olig1 and Olig2, the functions of Olig3 in cell type specification have just begun to be elucidated (Filippi et al.,2005; Müller et al.,2005). Furthermore, the cell lineage expressing the Olig3 gene remains elusive, especially in the later development of the CNS.

In the present study, we first examined Olig3 expression in detail in the developing spinal cord, including later stages of development, by Olig3 in situ hybridization. In addition, Olig3-lacZ knock-in (Olig3lacZ) mice were generated by homologous recombination and used for short-term lineage tracing experiments, since the half-life of the LacZ protein appears to be longer (Miano,2003) than that of Olig3. Unexpectedly, LacZ-positive cells were observed in the ventral-most part of the E10.5 spinal cord on sensitive X-gal staining, suggesting that some dorsal Olig3 lineage cells migrate into the ventral-most part of spinal cord by E10.5. The ventral-ward migration of dorsal cells and contribution to commissural interneurons at both intermediate and ventral-most levels were further substantiated by electroporation experiments with the EGFP gene in chick embryonic spinal cord. Dorsal midline cells were also LacZ-positive during development. These findings suggest that dorsal Olig3 lineage cells contribute to interneurons at the intermediate and ventral-most levels and to dorsal midline cells in the developing spinal cord.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Temporal and Spatial Pattern of Olig3 Expression in Spinal Cord Demonstrated by In Situ Hybridization

The expression of Olig3-mRNA in fetal mouse spinal cord was examined by in situ hybridization. Olig3 expression appeared in the neural tube by E9.5 (Takebayashi et al.,2002a). The pattern of expression in spinal cord was examined in detail in coronal sections. The dorsal-most portion expressed Olig3, with the exception of the roof plate (Fig. 1A). No expression was detected in the remaining part of the neural tube at this stage. E10.5 spinal cord also contained Olig3-expressing cells exclusively in its dorsal-most region (Fig. 1B). The dorsal Olig3 domain was reduced in size by E11.5 (Fig. 1C). In addition to the dorsal-most region, Olig3-expressing cells appeared in the ventral half of the spinal cord by this stage, as reported previously (Takebayashi et al.,2002a). Olig3-mRNA-positive cells in the ventral half were grouped into three distinct cell clusters (arrows in Fig. 1C; Takebayashi et al.,2002a). The Olig3-mRNA-positive cells in the dorsal-most region were observed both in the ventricular zone and in the thin mantle layer close to the pial surface, while those in the ventral half were positioned in the border regions between the ventricular zone and mantle layer (Fig. 1C).

thumbnail image

Figure 1. Distribution of Olig3-mRNA in fetal mouse spinal cord. A,B: E9.5 and E10.5, respectively. Olig3 expression is restricted to the dorsal-most portion of the neural tube. C: E11.5. Olig3-expressing domain in the dorsal spinal cord is reduced, while new Olig3 domains appear as three distinct spots in the ventral spinal cord (arrows). D: E12.5. Olig3 expression in the dorsal spinal cord remains in a region lateral to the roof plate and in the vertical cell strand (arrows), both of which are magnified in E. Three distinct ventral Olig3 spots are also observed (arrowheads). E: The discontinuous vertical cell strand is magnified. These cells course within the mantle layer. F: E13.5. Olig3-expressing cells are aligned along the outer edge of the ventricular zone. Dorsal-most Olig3 expression is restricted to only few cells. G: E14.5. Olig3 expression is observed in the intermediate and ventromedial parts of spinal cord, and also in the posterior median septum. The dorsal horn (DH) is devoid of Olig3-expressing cells, except for its basal portion (arrowheads). H: Higher magnification of the posterior median septum (arrows). Arrowhead indicates an Olig3-mRNA-containing cell. I: E15.5. Olig3-expressing cells are mostly restricted to the ventromedial part of spinal cord, and only a few cells at the intermediate level express Olig3 (arrowheads). Scales in A–D, F, G, I = 200 μm, and in E, H = 20 μm.

Download figure to PowerPoint

By E12.5, the Olig3-expressing region had been further reduced in size in the dorsal spinal cord. Cells lateral to the roof plate exhibited Olig3 expression. In addition, Olig3 expression was also observed in a few cells arranged dorsoventrally in the medial part of the mantle layer (Fig. 1D,E, arrows). The ventral Olig3-positive cells formed three distinct clusters (Fig. 1D, arrowheads), as observed at E11.5. The dorsal Olig3 domain was greatly reduced in size in E13.5 spinal cord compared with E12.5, and only a few cells exhibited Olig3 expression in the dorsal-most portion (Fig. 1F). These cells were aligned lateral to the dorsal ventricular zone, in a discontinuous cell strand. Although Olig3-expressing cells were roughly divided into dorsal and ventral groups in the mantle layer, the three distinct clusters in the ventral group were no longer recognized in the ventricular zone (Fig. 1F).

By E14.5, the white matter (or marginal zone) had developed in the peripheral part of the spinal cord and the contour of the gray matter (or mantle layer) had become H-shaped (Fig. 1G), somewhat resembling mature spinal cord. A small number of Olig3-expressing cells were positioned in the base of the dorsal horn (arrowhead in Fig. 1G). At this stage, the posterior median septum developed to separate the dorsal funiculus on both sides. While Olig3-expressing cells disappeared in the dorsal-most portion of the spinal cord, a limited number of such cells were observed in the posterior median septum (Fig. 1H). In addition, the base of the dorsal horn and the ventromedial part of the ventral horn contained Olig3-positive cells. This pattern of expression of Olig3 continued to at least E15.5 in the spinal cord (Fig. 1I), although the number of Olig3-expressing cells decreased.

Olig3lacZ expression in developing spinal cord

We generated lacZ knock-in mice as shown in Figure 2A–C, and carried out short-term lineage tracing. The reporter gene, lacZ, was knocked-in to the Olig3 locus (Fig. 2A). To confirm correct expression of lacZ in the Olig3 lineage cells, we compared localization of LacZ in the neural tube with that of Olig3-mRNA. Figure 2D shows E9.25 mouse with whole-mount X-gal staining. Olig3 signals were detected in the paramedian dorsal midline, similar to the pattern of localization of Olig3-mRNA (Takebayashi et al.,2002a). Expression of LacZ was further examined in coronal sections of E10.5 spinal cord. Almost all Olig3-positive cells (Fig. 3B) were immunoreactive to β-galactosidase (Fig. 3A,C), indicating correct expression of the reporter gene.

thumbnail image

Figure 2. Construction of the Olig3lacZ knock-in targeting vector. A: Targeted mutation of the Olig3 locus. Strategy for targeted replacement of Olig3 with lacZ. The open box indicates the Olig3 exon and the filled box the bHLH domain-encoding region. The lacZ and floxed PGK-neo cassette were inserted using an EcoRI site and XbaI site in the exon. The 5′ probes used for Southern-blot analysis are shown below the wild-type genomic map. B: Southern-blot analyses of genomic DNA isolated from mouse tails. BamHI-digested tail DNA was hybridized with 5′ probe. The 5′ probe detects 17.5-kb wild-type and 12.5-kb mutant bands. C: PCR analysis of tail DNA extracted from wild-type or Olig3+/- mice, using primers specific for wild-type Olig3 allele and Olig3lacZ allele. The positions of primers are shown in A. D: Whole-mount X-gal staining of E9.25 mouse. Paramedian dorsal portion (arrowheads) in the hindbrain (Hb) and spinal cord/neural tube (SpC/NT) show X-gal staining. Scale bar = 500 μm.

Download figure to PowerPoint

thumbnail image

Figure 3. Presence of LacZ in Olig3-immunoreactive cells. E11.5 Olig3lacZ mouse spinal cord was double-labeled with anti-β-galactosidase (A) and anti-Olig3 antibodies (B) and observed under a confocal laser scanning microscope. C: The images have been digitally merged. Olig3-immunoreactive cells (B, red) are simultaneously positive for β-galactosidase (A, green). Note that the roof plate is immunoreactive for β-galactosidase but not for Olig3 (arrows). Scale bar = 100 μm.

Download figure to PowerPoint

We next performed a short-term lineage tracing experiment using X-gal staining. As mentioned above, the dorsal-most portion of the spinal cord expressed LacZ in early, e.g., E9.5, neural tube (Fig. 4A). In addition to the X-gal-positive cells in the Olig3-expressing domain, the labeled cells were scattered in the subpial portion of the dorsolateral and lateral parts of the neural tube (arrows in Fig. 4A). Since we could not detect these cells by in situ hybridization (Fig. 1A), they had probably ceased Olig3 expression and had, therefore, become undetectable by in situ hybridization, though they were still detectable by sensitive X-gal staining. These X-gal-positive cells appeared to be derived from the dorsal Olig3-expressing domain and to be migrating ventrally.

thumbnail image

Figure 4. X-gal staining of developing Olig3lacZ mouse spinal cord. Coronal sections were subjected to X-gal histochemistry and counterstained with Nuclear Fast Red. A: E9.5. The dorsal-most portion of the neural tube exhibits intense X-gal reaction. In addition, a few X-gal-positive cells were sporadically observed in the subpial portion in the dorsolateral and lateral parts of spinal cord (arrows). B: E10.5. X-gal-positive cells are densely packed in the dorsal-most spinal cord. The dorsal-to-ventral cell strand (DVS) is formed from the dorsal Olig3 domain to the ventral-most level. Dorsally in the strand, the cells are positioned in the subpial portion, while in the ventral part of the strand X-gal-positive cells course medially to the ventral horn. Note that X-gal-positive cells reach the ventral-most part of the spinal cord (arrows). C: E11.5. The pattern of expression of lacZ is essentially the same as that at E10.5. In addition, weak X-gal-positive cells appear in the ventral ventricular zone, which are magnified in D. D: A higher magnification of the ventral ventricular zone. Weakly Olig3-positive cell clusters are indicated by arrows, and probably correspond to ventral Olig3-mRNA-expressing cells. The most ventrally located X-gal-positive cell cluster may include both ventrally- and dorsally-derived Olig3 lineage cells. E: E12.5. X-gal-positive cells are aligned along the outer edge of the ventricular zone and have accumulated in the intermediate level of spinal cord, which is magnified in F. Note that nearly all the roof plate cells are X-gal-positive. F: Higher magnification of the intermediate level. The mediolateral X-gal-positive cell cluster (curved arrow) is continuous with the DVS (smaller arrows). G: E14.5. X-gal cells are roughly grouped into three clusters, i.e., the posterior median septum (arrow), the mediolateral cell cluster, and the ventromedial cell group. H: E18.5. Three cell clusters are still observed in the dorsal midline (arrow), intermediate level, and the ventromedial part, although the number of labeled cells is decreased. I–K: E10.5. Double-immunofluorescence staining of β-galactosidase (I) and Islet1/2 (J) observed under a confocal laser scanning microscope. The images have been digitally merged in K. Note that double-labeled cells (arrowheads) are observed in the DVS. L,M: E10.5. X-gal staining followed by in situ hybridization of Math1 (L) and Brn3a (M). Light blue staining is X-gal reaction product while dark blue represents mRNA. Arrowheads indicate double-labeled cells in the dorsal spinal cord. Scale bars = 100 μm in A–C, E, G, H; = 50 μm in D, F, I–M.

Download figure to PowerPoint

In E10.5 spinal cord, in addition to the dorsal-most cells, a “dorsal-to-ventral cell strand” (DVS) of X-gal-positive cells was formed from the dorsal-most portion to ventral spinal cord (Fig. 4B). Unexpectedly, X-gal-positive cells were observed in the ventral-most part of the spinal cord at this stage (arrows in Fig. 4B). Since ventral Olig3-mRNA was not yet detected in E10.5 spinal cord (see Fig. 1B), these X-gal-positive cells were probably immigrating Olig3-expressing cells in the dorsal-most spinal cord. The DVS was aligned along the pial surface in the dorsal spinal cord and then coursed between the ventricular zone and the ventral horn in the ventral spinal cord.

The E11.5 spinal cord of the Olig3lacZ mice also exhibited the DVS (Fig. 4C). The cell strand was thicker and the number of labeled cells higher at this stage than at E10.5. X-gal-positive cells in the most ventral level were positioned just dorsal to the floor plate and lateral to the ventricular zone. At this stage, Olig3-mRNA expression had begun in the ventral ventricular zone (Fig. 1C), and weakly X-gal-labeled cells in the ventricular zone (arrows in Fig. 4D) probably corresponded to the ventral Olig3-expressing groups.

In E12.5 spinal cord, the DVS coursed between the ventricular zone and the mantle layer in the dorsal spinal cord as the mantle layer developed lateral to the strand (Fig. 4E). At this stage, X-gal-labeled cells began to accumulate in the intermediate level of the spinal cord, forming the mediolateral cell cluster (Fig. 4E,F). This cell cluster was continuous with the DVS (arrows in Fig. 4F). In the ventral spinal cord, X-gal-positive cells were distributed mainly in the ventromedial portion. Although the roof plate cells did not express Olig3-mRNA (Fig. 1), LacZ was labeled immunohistochemically or with X-gal histochemistry in these cells (Figs. 3A,C and 4C,E).

E14.5 Olig3lacZ mouse spinal cord contained more X-gal-positive cells in the intermediate and ventral levels. In addition, almost all of the cells in the posterior median septum were X-gal-positive (Fig. 4G). In contrast, few or no labeled cells were detected in the dorsal horn, except in its base. During the next three days, the number of X-gal-positive cells decreased greatly, and in the E18.5 Olig3lacZ mouse such cells were distributed in the intermediate region, probably including the base of the dorsal horn and the anterior horn (Fig. 4H). The posterior median septum also contained a small number of X-gal-positive cells (arrow in Fig. 4H).

Recently, eight types of dorsal interneurons have been distinguished based on combinatorial expression of various transcription factors (Bermingham et al.,2001; Gross et al.,2002; Helms and Johnson2003). To characterize dorsal Olig3 lineage cells in the early spinal cord, we used three transcription factors as dorsal interneuron markers: Islet1/2 (homeobox transcription factor for dI3), Math1 (bHLH transcription factor for dI1), and Brn3a (POU domain transcription factor for dI1, dI2, and dI3) (Bermingham et al.,2001; Gross et al.,2002; Helms and Johnson2003). E10.5 spinal cord was double-labeled with anti-β-galactosidase antibody and with anti-Islet1/2 antibody, and observed under confocal laser scanning microscopy.

Several LacZ-positive cells were also immunoreactive to Islet1/2 (Fig. 4I–K), and nearly half of Olig3 lineage cells were double-labeled with Islet1/2 in the DVS (44.4 ± 7.47%). In addition, Olig3 lineage cells appeared to express Math1 (Fig. 4L) or Brn3a (Fig. 4M) in the dorsal part of the spinal cord, though the sensitivity of in situ hybridization was decreased after X-gal staining. These patterns of expression indicate that Olig3 lineage cells differentiate into dI1, dI2, and dI3 dorsal interneurons, consistent with recent findings by Müller et al. (2005).

Ventral Migration of Cells Derived From Dorsal Spinal Cord

To verify ventral migration and differentiation of the cells derived from the dorsal-most region, including the Olig3 domain, in the spinal cord, we carried out electroporation of the EGFP gene to the dorsal-most portion of spinal cord and examined the migration and morphology of EGFP-expressing cells.

Prior to the electroporation experiment, we examined Olig3 expression in chick embryonic spinal cord with an in situ hybridization method. The E2 neural tube, the youngest stage examined, expressed Olig3-mRNA exclusively in its dorsal-most portion (Fig. 5A). No signals were detected in other parts of the neural tube. In E3 spinal cord, the dorsal-most portion exhibited intense expression of Olig3. In addition, three distinct spots of Olig3-mRNA were located in the ventral half of the spinal cord (Fig. 5B). This pattern of expression lasted until E4 in the spinal cord (not shown). Dorsal expression weakened with embryonic development (Fig. 5C). In E6 spinal cord, Olig3-expressing cells in the dorsal-portion were observed lateral to the roof plate. A short vertical cell strand also exhibited Olig3-mRNA expression (arrowheads in Fig. 5C). By contrast, at this stage the ventral spinal cord contained more Olig3-expressing cells in the ventral part than in the previous stages, as a result of which the ventral three Olig3 spots had become continuous. The spatio-temporal pattern of expression of Olig3 was thus similar in fetal mouse and early chick embryo, especially in dorsal spinal cord. The pattern of expression of Olig3 in E2 chick embryo neural tube appeared to correspond to that in E10.5 mouse spinal cord, while those in E3 and E4 chick spinal cord were similar to that in E11.5 mouse spinal cord.

thumbnail image

Figure 5. Dorsal-to-ventral migration of the cells derived from dorsal-most spinal cord in chick embryo. A–C: Olig3-mRNA distribution in chick embryo neural tube. A: E2. Olig3 expression is restricted to the dorsal-most portion of the neural tube. B: E3. The dorsal-most portion of the spinal cord exhibits intense expression of Olig3-mRNA, while weak expression is found in the ventral spinal cord as three distinct clusters (arrowheads). C: E6. Dorsal Olig3 expression is weak while ventral Olig3-expressing cells are increased in number. Arrowheads indicate the vertical Olig3 cell strand in the dorsal spinal cord. D-I: Electroporation of EGFP gene-containing plasmid to the dorsal-most spinal cord and subsequent distribution of EGFP-expressing cells. Brown color represents EGFP expression and dark blue in H and I represents Nkx2.2 immunoreactivity. D,F,H: Lower-magnification pictures showing sites of electroporation (larger arrowheads). Areas indicated by larger arrows in D, F, and H are magnified in E, G, and I, respectively. Smaller arrows indicate commissural fibers in the ventral commissure. Note that the EGFP-expressing cells have accumulated at the intermediate level. Smaller arrowheads in H indicate radial migration of cells which are crossing the DVS. E,G, I: Higher magnifications of D, F, and H, respectively. E and G are of the intermediate level, while I shows the ventral-most level. Arrowheads in G indicate EGFP-expressing cells in the intermediate level, and the arrow in G indicates a cell with fusiform shape, which appears to be migrating ventrally. Arrows in I indicate EGFP-expressing cells with commissural fibers located in the ventral-most level of the spinal cord, adjacent to Nkx2.2-positive p3 domain. Scale bars = 100 μm in A; = 20 μm in B–D, F, H; = 50 μm E, G, I.

Download figure to PowerPoint

We then performed electroporation to transfect the EGFP gene into the dorsal neural tube at E3 or E4, and the distribution of EGFP-expressing cells was observed three days after electroporation. Portions of spinal cord with electroporation into restricted regions were selected and analyzed. Of over 100 cases of electroporation examined, in at least 5 cases the EGFP gene was successfully transfected to a restricted dorsal region (larger arrowheads in Fig. 5D,F,H), which included the dorsal Olig3 domain. In these cases, EGFP-expressing cells can be considered as progeny of the dorsal spinal cord cells. In such cases of restricted electroporation, EGFP-expressing cell bodies were mostly located in the intermediate level of the mantle layer (larger arrows in Fig. 5E,G). The labeled cells extended fibers crossing the ventral commissure, which finally reached the contralateral ventral funiculus (smaller arrows in Fig. 5D,F,H). Thus, some dorsally derived cells differentiated into commissural interneurons at the intermediate level. Some EGFP-expressing cells were fusiform in shape (Fig. 5G, arrow), and appeared still to be migrating ventrally. Furthermore, EGFP-expressing cell bodies were also observed in the ventral-most level of the spinal cord, adjacent to the Nkx2.2 domain (Fig. 5H). These cells also extended commissural fibers (Fig. 5H,I). These findings indicate that dorsal cells contribute to commissural interneurons at the intermediate and ventral-most levels.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The present study examined the spatio-temporal pattern of expression of the Olig3 gene in fetal mouse spinal cord with an in situ hybridization method, and also with X-gal histochemistry of the Olig3lacZ mouse. Olig3 is expressed in both the dorsal and ventral spinal cord. Since the dorsal Olig3 lineage is only partly understood, we therefore focused primarily on the dorsal Olig3 cells, and in particular on their migration.

Descendants of the Olig3 Lineage Cells

We generated Olig3lacZ knock-in mice in this study. Since the floxed PGK-neo cassette was removed in these mice (see Experimental Procedures section), LacZ expression can be expected to be appropriately controlled by elements regulating Olig3 expression. Indeed, nearly all Olig3-positive cells were simultaneously positive for β-galactosidase at E10.5 (Fig. 3). LacZ expression in the Olig3lacZ knock-in mouse thus appeared to accurately mimic endogenous Olig3 expression. When the pattern of localization of Olig3-mRNA was compared with X-gal cell distribution in the Olig3lacZ mouse, X-gal-positive cells were found to be distributed much more widely, suggesting that the Olig3lacZ mouse can be used for short-term tracing experiments with Olig3 lineage cells.

We found that some LacZ-positive cells in the dorsal spinal cord co-expressed Islet1/2, Math1, or Brn3a, and that the dorsal Olig3 domain produces dI1, dI2, and dI3 (Class A) dorsal interneurons (Fig. 4I–M). A recent study by another group also showed that Olig3 lineage cells express other marker molecules for Class A dorsal interneurons such as Lhx2/9 and Foxd3, but not Lbx1, a marker for dI4, dI5, and dI6 (Class B) interneurons (Müller et al.,2005). Two studies have reported loss-of-function experiments with Olig3, in which loss-of-function of Olig3 affected these lineages. Filippi et al. (2005) found that the Olig3 gene regulates D2 interneuron differentiation and also functions to establish the neural crest-lateral neural plate boundary in zebrafish embryo. Müller et al. (2005), in an analysis of Olig3 knockout mice, demonstrated that Olig3 plays roles in specification of dI2 and dI3 interneurons. Although Olig3-mRNA was not detected in dorsal midline at the roof plate stage, LacZ was detected by X-gal histochemistry or β-galactosidase immunohistochemistry in this region of the Olig3lacZ mouse. These findings were also obtained in another Olig3 knock-in mouse strain (Müller et al.,2005). Since dorsal midline cells in the roof plate appeared to produce dorsal interneurons (dI2, dI3) migrating ventrally in a lineage tracing experiment using Gdf7-Cre mice (Lee et al.,2000), there might be turnover of the midline cells in the roof plate, and dorsal Olig3 lineage cells might replace cells on the roof plate. In later development of the spinal cord, dorsal midline glial cells in the dorsal median septum express Olig3-mRNA (Fig. 1G,H). Furthermore, nearly all cells in the septum of the Olig3lacZ mouse were X-gal-positive at E14.5, and thus some of the dorsal Olig3 lineage cells differentiate into the dorsal midline glia. Although the dorsal midline glia are considered as a subtype of astroglia in the mammalian spinal cord (Böhme1988; Parkinson and Del Bigio,1996), further studies are needed to elucidate their characteristics and lineage. Of note, inhibition of Olig3 function in zebrafish embryos resulted in the absence of gfap-positive cells, as well as D2 interneurons, in the dorsal spinal cord (Filippi et al.,2005).

Since few or no cells in the marginal zone (or future white matter) contain Olig3-mRNA or LacZ, Olig3 lineage cells of the rodent spinal cord may make little contribution to the oligodendrocyte lineage.

Dorsal-to-Ventral Migration of Olig3 Lineage Cells

Dorsal-to-ventral migration of spinal neurons had been suggested by Golgi studies (Ramon y Cajal,19091911; Wentworth,1984a b), and was substantiated by lacZ-retrovirus infection of the chick embryo (Leber and Sanes,1995). In more recent studies, dI1 and dI2 interneurons have been reported to migrate ventrally and to differentiate into commissural interneurons in the base of the dorsal horn (Bermingham et al.,2001; Helms and Johnson,2003). Olig3-mRNA and Olig3 protein were confined to the restricted dorsal-most portion (dorsal Olig3 domain) until E11. However, X-gal-positive cells were also distributed outside the dorsal Olig3 domain in the spinal cord of Olig3lacZ mice. These X-gal-positive cells were also found in the subpial portion of the E9.5 spinal cord, suggesting that they are migrating from the dorsal Olig3 domain and that the onset of tangential migration is as early as E9.5. The tangential migration of Olig3 lineage cells became more prominent at E10.5, and X-gal-positive cells formed the DVS. Since ventral Olig3 expression in the progeny of the ventral ventricular zone starts at E11.5 and is not yet detected at E10.5, the ventromedial X-gal-positive cells at this stage are probably derived from the dorsal Olig3 cells. Olig3-LacZ-expressing cells form the mediolateral cell cluster at the intermediate level of the spinal cord at E12.5, and the intra-mantle DVS is continuous with the mediolateral X-gal cell cluster, suggesting settlement of the dorsal Olig3 lineage cells in this region (Fig. 4E,F). To verify ventral migration and differentiation of the cells derived from the dorsal-most region, we performed an electroporation experiment with EGFP-expressing plasmid in the dorsal spinal cord of chick embryo. In this experiment, the EGFP-expressing cells, which appeared to be derived from the dorsal spinal cord, migrated ventrally, and accumulated at the ventral, as well as the intermediate, level of spinal cord, and extended their processes contralaterally through the ventral commissure. Although the origin of these EGFP-expressing cells could not be clearly shown to be the dorsal Olig3 domain in these experiments, the electroporated region contained the Olig3 domain. These observations, taken together with those for the DVS of X-gal-positive cells in Olig3lacZ spinal cord, suggest the possibility that the dorsally derived Olig3 cells migrate towards not only the intermediate level but also the ventral-most level of the spinal cord, and that at least some of them become commissural interneurons. Although the present study focused on Olig3 expression in the spinal cord, Olig3 is also expressed in the lower rhombic lip (LRL), the dorsolateral margin of the caudal hindbrain (Fig. 2D; Takebayashi et al.,2000a). The cells derived from the LRL also migrate ventrally to differentiate into precerebellar neurons in the brainstem (Ono et al.,2004). Olig3 lineage cells in spinal cord and hindbrain may share a common mechanism of ventral migration.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Animals

Pregnant mice of ICR strain were purchased from Shizuoka Laboratory Animal Center (Shizuoka, Japan). Noon of the day when vaginal plugs were observed was regarded as embryonic day 0.5 (E0.5). Fetuses at E8.5–E18.5 were used for experiments. Fertilized chicken eggs, purchased locally, were incubated at 38°C. The stages of development of the chick embryos were determined according to Hamburger and Hamilton (1951). All procedures were approved by the Animal Research Committee of the National Institutes of Natural Sciences.

Construction of Targeting Vector

The Olig3 genomic clones were obtained from a 129SvJ mouse genomic library (in λ DASH II: Stratagene, La Jolla, CA) by hybridization screening using mouse Olig3 cDNA probe. We subcloned 5-kb XbaI-EcoRI and 3-kb XbaI-BamHI fragments into a lacZ-loxP-neo-loxP gene cassette vector. All of the Olig3 ORF was deleted and replaced with lacZ using the EcoRI site overlapping the first methionine. The targeted allele was designated Olig3lacZ.

Electroporation and Generation of Germ Line-Transmitting Mice

TT2 ES cells (1 × 107) (Yagi et al.,1993) were electroporated with 50 μg of linearized targeting vector DNA in 0.5 ml of phosphate-buffered saline (PBS) using a Bio-Rad (Richmond, CA) Gene Pulser. The electroporated cells were then cultured on feeder cells, which had been prepared from G418-resistant primary embryonic fibroblasts (Invitrogen, La Jolla, CA) and selected with G418 (300 μg/ml) for 6 to 8 days. Selected clones were cultured in 96-well plates; the clones were then frozen and their genomic DNA was isolated for Southern-blot analysis.

Homologous recombinants, identified by Southern-blot analysis using 5′ external probe, were expanded for injection. Three recombinants were identified. Chimeric mice were generated by injection of the Olig3-mutant ES cells into 8-cell stage ICR embryos and implantation into pseudopregnant foster mothers at the Laboratory Animal Resource Center of Tsukuba University. The resulting chimeric males were bred with Balb/c females. The heterozygotic mice were backcrossed with C57BL/6NCrj inbred mice. The floxed PGK-neo cassette was removed by crossing with the Pc36 line, which expresses Cre recombinase activity under control of a ubiquitous CAG promoter (Niwa et al.,1991).

Genotype Analysis

For PCR genotyping, the wild-type or targeted allele was detected using a common sense primer (5′-CCATAAGGCCGCATCTCTTG-3′) and antisense primers specific for wild-type Olig3 (5′-GCTCCGACAGCTGCTTCTTG-3′) and Olig3lacZ (5′-GCCAGGGTTTTCCCAGTCAC-3′). These primer pairs amplify a 265-bp fragment from wild-type Olig3 and a 150-bp fragment from Olig3lacZ. PCR conditions were 95°C for 9 min followed by 95°C for 20 sec, 57°C for 30 sec, and 72°C for 30 sec, for 35 cycles using Ampli Taq Gold (Applied Biosystems, Foster City, CA).

Tissue Preparation

Pregnant mice were deeply anesthetized with pentobarbital and their offspring were removed from the uterus. Fetal mouse spinal cord was fixed by immersion in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) overnight (E15.5 or younger), or by perfusion through the heart with 4% PFA and then immersed overnight in the same fixative (E16.5 or older). Chick embryos were fixed by immersion in the fixative overnight. These tissues were subsequently immersed in 20% sucrose PBS overnight. The cervical and thoracic levels of the spinal cord were mainly used for morphological analysis, and were embedded in OCT compound and frozen in liquid nitrogen. The samples were stored at −80°C until cutting. Sections of spinal cord were cut sagittally or coronally with a cryostat and thaw-mounted into MAS-coated glass slides (Matsunami Co., Tokyo, Japan). The sections were used for in situ hybridization, X-gal staining, and immunohistochemistry.

X-Gal Staining

Sections of Olig3lacZ mice were immersed into staining solution composed of 1 mg/ml X-gal, 16 mM potassium ferrocyanide, 16 mM potassium ferricyanide, 2 mM magnesium chloride, 0.02% NP-40, and 0.01% sodium deoxycholate in PBS, and kept overnight at room temperature. After washing in PBS, they were lightly counterstained with Nuclear Fast Red. Some of the X-gal-stained sections were subsequently processed for immunohistochemistry or for in situ hybridization (see below).

Cell Migration Analysis

In ovo electroporation to chick embryo was performed as previously described (Ono et al.,2001,2004). Briefly, pGAP-GFP (Ono et al.,2001) or pEF1α-GFP (Tabata and Nakajima,2001) plasmids were used. Eggs were windowed at E3 or E4 (HH stage 18–24). A plasmid solution was injected into the central canal via a micropipette, and electrodes were dorsally placed on the epidermis covering the roof plate (anode) and the opposite side of the embryo (cathode). A 50-ms rectangular pulse of 35 V was delivered twice by a pulse generator (SEN-7203, Nihon Kohden, Tokyo, Japan). The window was sealed with adhesive tape, and the eggs were returned for further incubation. These embryos were fixed three days after electroporation.

In Situ Hybridization

mRNA for Olig3, Brn3a, and Math1 was detected by in situ hybridization using digoxygenin (DIG)-labeled antisense riboprobes (Akazawa et al.,1995; Eng et al.,2004; Takebayashi et al.,2002a; Zhou et al.,2001). The sections were treated with proteinase K (1 μg/ml, 5–30 min, room temperature), and hybridized with 0.3 μg/ml riboprobe in a hybridization buffer (50% formamide, 20 mM Tris-HCl at pH 7.5, 600 mM NaCl, 1 mM EDTA, 10% dextran sulfate, 200 μg/ml yeast tRNA, 1× Denhardt's solution, 0.25% SDS) at 65°C overnight. The sections were washed three times with 1× SSC containing 50% formamide at 65°C, followed by maleic buffer (0.1 M, pH 7.5) containing 0.1% Tween 20 and 0.15 M NaCl. DIG-labeled probe was visualized by overnight incubation of the sections with anti-DIG antibody conjugated to alkaline phosphatase (1:2,000; Roche) and NBT/BCIP reaction.

Immunohistochemistry

Immunohistochemical staining of sections were carried out as described previously (Ono et al.,2004). The sections were incubated with a primary antibody overnight at room temperature. They were then processed with the ABC method (ABC Elite kit, Vector Labs, Burlingame, CA) or with the indirect immunofluorescence method using species-specific secondary antibodies conjugated to Alexa Fluor 488 or 594 (Molecular Probes, Eugene, OR). Some of the samples double-labeled by the immunofluorescence method were observed under a confocal laser scanning microscope (LMS-510, Carl Zeiss, Oberkochen, Germany). For the ABC method, color was developed by incubating sections with PBS containing diaminobenzidine HCl (DAB) and hydrogen peroxide. The primary antibodies used were anti-Olig3 rat polyclonal antibody (1:2,000; Takabayashi et al.,2002), rabbit anti-β-galactosidase (1:1,000; ICN/Cappel, Aurora, OH), anti-EGFP rabbit IgG antibody (1:2,000; Molecular Probes), mouse anti-Islet 1/2 (1:5; Developmental Study Hybridoma Bank; DSHB), and mouse anti-Nkx2.2 (1:10; DSHB). Some of the EGFP-electroporated sections were double-stained with anti-EGFP and with anti-Nkx2.2. Sections were first treated with anti-EGFP followed by color development with the ABC method using DAB as a chromogen. Subsequently, they were microwave-irradiated in citrate buffer (0.01M, pH 6.0) at over 95°C for 5 min to inactivate EGFP-HRP and simultaneously to retrieve Nkx2.2 antigenicity. They were then incubated with anti-Nkx2.2 antibody and processed with the ABC method. Dark purple color was developed with DAB-Nickel as a chromogen (Adams1981).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The authors thank David Anderson (Caltech) for chick Olig3 cDNA, Eric Turner (UCSD) for mouse Brn3a cDNA, Ryoichiro Kageyama (Kyoto Univ.) for Math1 cDNA, Yoko Nabeshima for instructions on blastocyst injection, and Thomas Müller and Carmen Birchmeier for communication of unpublished results. We are also grateful to Daisuke Nakamura, Rie Taguchi, Minako Ito, and Chieko Ueda for technical assistance, and Seiji Hitoshi, Tetsuichiro Saito, and Mengsheng Qiu for helpful discussion. The monoclonal antibodies 74.5A5, 39.4D5, and 67.4E12 were obtained from the Developmental Studies Hybridoma Bank maintained by the University of Iowa. This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas-Advanced Brain Science Project and Research for the Future from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES