Increase of proliferating oligodendroglial progenitors in the adult mouse brain upon Sonic hedgehog delivery in the lateral ventricle

Authors

  • Karine Loulier,

    1. CNRS, Signal Transduction and Developmental Neuropharmacology, UPR9040 Laboratoire de Neurobiologie Cellulaire et Moléculaire, Institut de Neurobiologie Alfred Fessard, IFR 2118, Gif sur Yvette, France
    Search for more papers by this author
  • Martial Ruat,

    1. CNRS, Signal Transduction and Developmental Neuropharmacology, UPR9040 Laboratoire de Neurobiologie Cellulaire et Moléculaire, Institut de Neurobiologie Alfred Fessard, IFR 2118, Gif sur Yvette, France
    Search for more papers by this author
  • Elisabeth Traiffort

    1. CNRS, Signal Transduction and Developmental Neuropharmacology, UPR9040 Laboratoire de Neurobiologie Cellulaire et Moléculaire, Institut de Neurobiologie Alfred Fessard, IFR 2118, Gif sur Yvette, France
    Search for more papers by this author

Address correspondence and reprint requests to M. Ruat or E. Traiffort, CNRS, Signal Transduction and Developmental Neuropharmacology, UPR9040 Laboratoire de Neurobiologie Cellulaire et Moléculaire, Institut de Neurobiologie Alfred Fessard, IFR 2118, 1 avenue de la terrasse, 91198 Gif sur Yvette, France.
E-mail (M. Ruat): ruat@nbcm.cnrs-gif.fr
E-mail (E. Traiffort): Elisabeth.Traiffort@nbcm.cnrs-gif.fr

Abstract

Sonic hedgehog signaling is required for the maintenance of stem cell niches in the postnatal subventricular zone and the proliferation of neural progenitors in the mature hippocampus. We show here that delivery of Sonic hedgehog protein into the lateral ventricle of adult mice increases cell proliferation in the corpus callosum and cerebral cortex. In this latter area, the number of neural progenitors expressing the proteoglycan NG2 is enhanced 2 days after the injection. In both areas, mRNA up-regulation of the transcriptional target gene Patched was observed in cells expressing the oligodendroglial transcription factor Olig1. Twenty-six days following the adenovirus-mediated delivery of Sonic hedgehog into the lateral ventricle, newly generated cells in the cerebral cortex and in the corpus callosum are influenced towards the initial steps of oligodendrogenesis, as indicated by a 50% increase in the number of cells expressing the oligodendroglial marker DM20. Our experiments demonstrate that the number of oligodendrocyte precursor cells in the cerebral cortex and corpus callosum can be increased upon delivery of Sonic hedgehog proteins and highlight the potential capacity of the adult brain to mobilize a pool of premyelinating cells.

Abbreviations used
BDNF

brain-derived neurotrophic factor

BrdU

bromodeoxyuridine

CNPase

2′,3′-cyclic nucleotide 3′-phosphohydrolase

EGF

epidermal growth factor

eGFP

enhanced green fluorescent protein

FGF

fibroblast growth factor

GFAP

glial fibrillary acidic protein

LV

lateral ventricle

MBP

myelin basic protein

OL

oligodendrocyte

OLP

oligodendroglial precursor

PSA-NCAM

polysialylated neural cell adhesion molecule

Ptc

Patched

Shh

Sonic hedgehog

ShhN

aminoterminal domain of Sonic hedgehog

Smo

Smoothened

SVZ

subventricular zone

Sonic hedgehog (Shh) is a member of the hedgehog family of secreted proteins involved in embryonic development. Synthesized as a large precursor protein, Shh undergoes unique post-translational processing including autoproteolysis and lipid modifications of its biologically active aminoterminal (ShhN) domain. ShhN binding to its receptor Patched (Ptc) relieves the inhibition that Ptc exerts on the seven-transmembrane domain protein Smoothened (Smo). This leads to the activation of downstream Shh target genes including the transcription factors of the Gli family and Ptc itself (Briscoe and Therond 2005). During the early embryonic stages, Shh is responsible for the ventralization of the neural tube and specifies the fate of ventral cells including motoneurons and interneurons of the spinal cord, dopaminergic neurons of the midbrain and serotoninergic cells of the ventral forebrain (McMahon et al. 2003).

Gain and loss-of-function experiments in mouse have further indicated that Shh is required for the specification of a first wave of oligodendroglial progenitors (OLP) originating in the ventral regions of the embryonic spinal cord and forebrain through induction of genes encoding the basic helix-loop-helix (bHLH) proteins, Olig1 and Olig2 (McMahon et al. 2003; Rowitch 2004; Briscoe and Therond 2005). In the dorsal spinal cord, a later production of OLP was reported to be Shh-independent (Cai et al. 2005; Fogarty et al. 2005; Vallstedt et al. 2005), whereas in the neocortex, Shh also potently induces precursors to adopt an oligodendroglial fate (Murray et al. 2002).

In the postnatal and adult nervous system, the cellular Shh cascade is still present (Traiffort et al. 1999; Coulombe et al. 2004). ShhN modulates electrophysiological properties of mature neurons (Pascual et al. 2005) and the expression of its receptor Hedgehog interacting protein (Hip) in neurons expressing the neuronal nitric oxide synthase suggests a functional link between nitric oxide and Shh signaling (Loulier et al. 2005). During postnatal development, Shh regulates the proliferation of cells in the dorsal brain, namely the neocortex, tectum and cerebellum (Ruiz i Altaba et al. 2002). It controls the number of neocortical cells with stem cell properties and is required for the maintenance of stem cell niches in the subventricular zone (SVZ) of the lateral ventricles (LV), a major neurogenesis area in the adult (Machold et al. 2003; Palma and Ruiz i Altaba 2004). In adulthood, the adenovirus-associated mediated transfer of Shh in the hippocampus regulates the proliferation of adult neural progenitors in the subgranular zone (SGZ) of the dentate gyrus (Lai et al. 2003). In addition, the in vivo administration of ShhN in rodent striatum (Charytoniuk et al. 2002) or of a Smo agonist orally in mouse (Machold et al. 2003) activates the Shh signaling pathway in the adult SVZ. However, in vivo analysis of Smo and Shh null mice and in vitro data obtained with neurospheres derived from adult neural stem cells were not conclusive as to whether stimulation of adult SVZ cells in vivo by Shh induces cell proliferation (Machold et al. 2003; Palma et al. 2005).

Those observations prompted us to investigate the effects induced by the stereotaxic delivery of Shh into the LV on the neural progenitors of adult mice in vivo. Here we show that the activation of Shh pathway occurs in several brain areas, including the SVZ, the corpus callosum and the cerebral cortex. We observed a potent increase in the number of proliferating cells, namely those expressing the oligodendroglial markers NG2 and DM20 in the cerebral cortex or DM20 in the corpus callosum.

Materials and methods

Adenovirus production

Adenoviral vectors were generated by the Genethon Viral Production Network (Genopôle, Evry, France): pAd/CMV:huShh:IRES:eGFP contains human Shh cDNA (from Dr C. Tabin) under the control of the constitutive cytomegalovirus promoter, placed in tandem sequence with enhanced green fluorescent protein (eGFP) under the control of an internal ribosomal entry site; pAd/CMV:IRES:eGFP excludes Shh cDNA and was used as a control. Replication-defective recombinant type 5 adenoviruses (AdhuShh-eGFP and Ad-eGFP) were obtained.

Recombinant protein and bromodeoxyuridine injection

Male adult OF1 mice (25–30 g) were anesthetized with pentobarbital and stereotaxically injected into the right lateral ventricle (LV; anteroposterior +0.2 mm, lateral +0.8 mm, dorsoventral −2.5 mm) (Shimazaki et al. 2001), with 5 µL (3 µg) of human ShhN (Biogen Idec, Boston, MA, USA) or the vehicle solution as described (Charytoniuk et al. 2002). One or two days later, mice (n = 5 for each group) received two intraperitoneal (i.p.) injections of bromodeoxyuridine (BrdU) as described (Charytoniuk et al. 2002) and were killed 2 h following the last injection (‘day-1 mice’ and ‘day-2 mice’, respectively). Other animals (n = 5) received four injections of BrdU on day 2 and were killed 15 days after the stereotaxic injection (‘day-15 mice’). All procedures were carried out in accordance with the European Communities Council Directive (86/806/EEC) for the care and use of laboratory animals.

Adenoviral vector injection

A total of 75 × 106 plaque forming units (pfu) of either Ad-eGFP or AdhuShh-eGFP were injected into the right LV (3 µL) of mice as described above. At day 4, mice either were killed (‘day-4 mice’, n = 2 for each group) or were injected once daily between day 4 and day 18 before being killed at day 26 (‘day-26 mice’, n = 6 for each group).

Tissue processing and immunostaining

Frontal brain sections (16 µm) were processed and immunostained with primary antibodies against BrdU (1 : 200), 2′,3′-cyclic nucleotide 3′-phosphohydrolase (CNPase, 1 : 200, Sigma, Saint Quentin Fallavier, France), ShhN (1 : 500), glial fibrillary acidic protein (GFAP, 1 : 500; Dako, Cergy Saint-Christophe, France), rabbit NG2 (1 : 500), mouse NeuN (1 : 1000), mouse RIP (1 : 100000) (Chemicon, Paris, France), rabbit Olig1 (1 : 700; a generous gift from Drs C. D. Stiles and J. A. Alberta, Boston, MA, USA) as described (Traiffort et al. 2001; Charytoniuk et al. 2002). The secondary antibodies used were: biotinylated goat anti-rat (1 : 200, Vector Laboratories, Paris, France), FITC- or TRITC-labelled goat anti-rat (1 : 400, Jackson ImmunoResearch Laboratories Inc., Marseille, France), FITC-labelled goat anti-rabbit (1 : 400, Santa Cruz Biotechnologies, Le Perray en Yvelines, France), highly cross-adsorbed goat anti-mouse and anti-rabbit Alexa Fluor 546 (1 : 1000, Molecular Probes, Cergy Pontoise, France). Immunofluorescent signals were detected by a conventional epifluorescence microscope (Leica DMRXA2; Leica Microsystems, Rueil Malmaison, France) or an upright laser scanning confocal microscope (Leica TCS SP2). Photomicrographs presented are from the cerebral hemisphere contralateral to the injection site unless mentioned. In situ hybridization experiments using Ptc, Gli1, DM20 or myelin basic protein (MBP) specific probes were performed as described (Charytoniuk et al. 2002). After alkaline phosphatase revelation, sections were either mounted with Aquaperm FluidTM medium (CML, Nemours, France) or washed in phosphate-buffered saline before revelation of anti-BrdU immunohistochemistry as described (Charytoniuk et al. 2002). For double Ptc in situ hybridization and Olig1 immunohistofluorescence, TSA™-Plus Fluorescein System (PerkinElmer Life Sciences, Courtaboeuf, France) was used for Ptc riboprobe detection according to the protocol provided by the supplier.

Cell counting

All cell counts were performed by two independent experimenters according to a double-blind protocol. BrdU+ nuclei were counted under a light microscope within the entire area of the cortex, corpus callosum, striatum and SVZ observed in 3–5 slices per animal (n = 5 for each group of ‘day 1’, ‘day 2’ and ‘day 15 mice’, n = 6 for each group of ‘day-26 mice’). The lateralities of the slices analyzed were comparable for each mouse and encompassed bregma levels −1.22 and +1.42. BrdU+ cells coexpressing another marker were quantified in the contralateral cortex except when stated in three to six consecutive slices from the sampled lateralities described above. Two counting areas (620 µm × 815 µm, represented as red boxes in Fig. 2c, see Results section) were defined with the 10 × objective. The total number of BrdU+ nuclei was evaluated in each counting area. Results were expressed as a percentage of this total number. Statistical analysis was performed using consecutively the unpaired bilateral Student's t-test and one-way anova. Statistical significance was considered for p-value < 0.05 and high significance for p-value < 0.01.

Figure 2.

 Characterization of proliferating cells in brain after an acute amino terminal domain of Sonic hedgehog (ShhN) treatment. (a, b) Bromodeoxyuridine (BrdU) immunohistochemistry on frontal brain sections from ShhN-injected ‘D2 mice’. Distribution of BrdU+ nuclei in the cortex [Cx, arrows in (a)], subventricular zone [SVZ, arrowhead in (b)] and corpus callosum [cc, (b)] of the contralateral injected side. The dotted line indicates the border line between the cortex and the corpus callosum. (c) Scheme indicating the distribution of BrdU+ cells in the cortex (Cx), corpus callosum (cc), lateral septum (ls), striatum (st) and subventricular zone (SVZ). Red boxes indicate representative areas used for cell counting. (d, e) Double BrdU immunohistochemistry (brown) and Patched (Ptc) in situ hybridization (blue) performed on frontal sections from ShhN-injected mice at the level of the Cx (d) or cc (e). Double-labelled cells are indicated by arrows. Scale bars (µm): 20 (a, b), 25 (d, e).

Western blot analysis

Expression of ShhN was investigated on homogenates and culture media from CHO-K1 cells (1.5 × 106/3 mL) harvested 24 and 48 h after infection with increasing concentrations of AdhuShh-eGFP, and on homogenates of 30 cryostat-cut consecutive sections evenly harvested between bregma levels −1.22 and +1.42 for each adenovirus-infected brain analyzed. Western blot analysis was performed using the ShhN antiserum (Traiffort et al. 2001) on either 12 µL of culture medium or 30 µg of proteins from brain homogenates.

Results

Proliferation of cells in response to Sonic hedgehog

To investigate whether stimulation of adult SVZ cells in vivo by Shh induces cell proliferation, we have stereotaxically injected 3 µg of ShhN directly into the right LV of adult mice (Fig. 1a). The animals were then killed 2 days later. We first demonstrated activation of the pathway by a robust up-regulation of Ptc mRNA, a downstream transcriptional target of Shh signaling. The increase of Ptc transcription was observed in chains of cells in the SVZ and rostral migratory stream, in scattered cells in the striatum, lateral septum, corpus callosum and cerebral cortex from ShhN- (n = 5) but not from vehicle-treated animals (n = 5; Figs 1b–f). Such activation was observed in both ipsi- and contralateral sides, indicating that the active ShhN protein had diffused throughout the cerebrospinal fluid. Further evidence for pathway activation in these brain areas was demonstrated by the increased mRNA of the transcription factor Gli1 from ShhN-treated (Figs 1h and i) but not from vehicle-treated (Fig. 1g) animals.

Figure 1.

 Influence of amino terminal domain of Sonic hedgehog (ShhN) injection in the lateral ventricle (LV) on the expression of Patched (Ptc) mRNA in the brain of adult mice. (a) Time course of the delivery of ShhN or vehicle and bromodeoxyuridine (BrdU). Mice were injected into the right LV on day 0 (D0), received BrdU on day 2 (D2) and were killed 2 h later. (b–i) In situ hybridization experiments performed on frontal (b–d, g–i) or sagittal (e, f) brain sections indicate that Ptc (b–f) and Gli1 (g–i) transcripts are induced in the subventricular zone (SVZ) (arrowheads), the cortex (Cx), the corpus callosum (cc) and the striatum (b, c, e, h, i) of ShhN-injected but not vehicle-injected (d, g) mice. Ptc transcription is also detected in the rostral migratory stream (RMS) of ShhN-treated (e) but not vehicle-treated (f) animals. Arrows in (b, d, g, h) indicate the injection site. Scale bars (µm): 200 (b, d–h), 40 (c).

To study cell proliferation, mice received two i.p. injections of BrdU at 2-h intervals and were killed 2 h after the last injection. Analysis of frontal brain sections from ShhN- and vehicle-treated animals showed the presence of BrdU+ cells in the SVZ, the corpus callosum, the cerebral cortex and also to a lesser extent in other brain areas such as the lateral septum and the striatum (Figs 2a–c). Furthermore, we observed the presence of double-labelled Ptc+/BrdU+ cells in the cortex, the corpus callosum and the SVZ of ShhN treated mice but not of control mice (Figs 2d and e and data not shown), indicating the presence of proliferating Shh-responsive cells in these regions. ShhN treatment was not followed by a significant increase of cell proliferation in the SVZ, nor in the striatum compared to controls in both contralateral and ipsilateral sides. However, in the contralateral side, BrdU+ cell number was significantly enhanced by two to three-fold in the corpus callosum (ShhN, 34 ± 2 cells per section, total counted cells n = 1370; control 14 ± 1, n = 416; p < 0.01) and more than three-fold in the cortex (ShhN, 85 ± 5, n = 3381; control 25 ± 5, n = 749; p < 0.01) upon ShhN treatment (Fig. 3). Similar results were obtained from the injected side (data not shown).

Figure 3.

 Quantification of bromodeoxyuridine-positive (BrdU+) cells in brain regions after amino terminal domain of Sonic hedgehog (ShhN) injection. A significant increase in BrdU+ cell number is observed in the contralateral cortex (Cx) and corpus callosum (cc), but not in the striatum (st) and subventricular zone (SVZ) from ShhN-treated compared to vehicle-treated animals. Values are means ± SEM from three to five animals. **p < 0.01.

Sonic hedgehog signaling induces NG2+ progenitor proliferation in the mature cerebral cortex

To characterize the phenotype of the proliferating cells in both the corpus callosum and cerebral cortex, frontal sections were stained for BrdU and NG2, a chondroitin sulfate proteoglycan expressed by cells representing the major pool of adult brain proliferative progenitors distributed throughout the adult CNS and described mainly as OLP (Dawson et al. 2003). Immunohistochemical analysis revealed expression of NG2 reactivity on both the body and processes of cells that were often bipolar and distributed both in the white and gray matter of treated and control animals. Among BrdU+ cells analyzed by conventional and confocal microscopy, the percentage of BrdU+/NG2+ cells was significantly higher (p < 0.05) in the cortex of ShhN-treated animals (Figs 4a and b; Table 1), but not in the corpus callosum (data not shown). Analysis of BrdU+/GFAP+ staining (Figs 4c and d) indicated that the percentage of BrdU+/GFAP+ cell number was decreased but did not reach significance in ShhN-treated mice (Figs 4c and d; Table 1).

Figure 4.

 Phenotype of bromodeoxyuridine-positive (BrdU+) cells upon amino terminal domain of Sonic hedgehog (ShhN) injection in D2 mice. (a, b) Frontal sections of ShhN-treated animals were analyzed by immunohistochemistry using anti-NG2 [red in (a), green in (b)] and anti-BrdU [green in (a), red in (b)] antibodies. Double-labelled cells are observed by conventional [arrows in (a)] and confocal (b) microscopy. (c, d) A subpopulation of cortical GFAP+ cells [yellow in (c), red in (d)] displays BrdU+ nuclei [blue in (c), green in (d)] observed by conventional (c) and confocal (d) microscopy in ShhN-treated mice. Inset in (c) is a higher magnification of the boxed area. Scale bars (µm): 50 (a, c). GFAP, glial fibrillary acidic protein.

Table 1.   Bromodeoxyuridine-labelled cortical cell types recruited 2 days after amino terminal domain of Sonic hedgehog (ShhN) injection
 VehicleShhN
  • Data represent the percentages of cells double-labelled for bromodeoxyuridine (BrdU) and a cell-specific marker over the total population of BrdU+ cells counted in the contralateral cortex. The total number of double-labelled cells counted from three to five animals appears in brackets. Values are means ± SEM.

  • a

    p < 0.05.

NG2+22 ± 3 (47)31 ± 2a (243)
GFAP+39 ± 9 (59)28 ± 4 (173)

To determine when the first BrdU+/NG2+ cells appear in the cortex, we have carried out an experiment where mice (n = 5) were killed 24 and 2 h after ShhN and BrdU injections, respectively (Fig. 5a). Ptc transcription was already detected in chains of cells in the SVZ and occasionally in cells of the corpus callosum and cerebral cortex (Fig. 5b and data not shown), indicating that up-regulation of Shh signaling occurred more rapidly in areas closely related to the LV, such as the SVZ, than in remote areas such as the cerebral cortex. Comparison of BrdU+ cell number in brain areas of 24 or 48 h ShhN-treated animals showed that BrdU+ cells were at least seven times less numerous in the cerebral cortex of 24 h ShhN-treated animals (Fig. 5c and Table 2). Among BrdU+ cells, 13% (Fig. 5d) and 31% (Table 1) were also positive for NG2 in the cerebral cortex of 24 h or 48 h ShhN-treated animals, respectively. These data indicate that the robust increase of cell proliferation and of BrdU+/NG2+ cells observed in the cerebral cortex of 48 h ShhN-treated animals coincides with Ptc transcription, supporting the idea that Shh signaling mediates these effects.

Figure 5.

 Patched (Ptc) transcription and bromodeoxyuridine (BrdU) incorporation in brain areas 24 h after amino terminal domain of Sonic hedgehog (ShhN) injection. (a) Time course of the delivery of ShhN or vehicle and BrdU. Mice were injected into the right lateral ventricle (LV) on day 0 (D0), received BrdU on day 1 (D1) and were killed 2 h later. (b) In situ hybridization experiment performed on a frontal brain section indicates that Ptc transcripts are induced mainly in the subventricular zone (SVZ) (arrowhead) of ShhN-injected mice. (c) Immunohistochemistry on a consecutive slice identifies BrdU+ nuclei mainly in the SVZ. (d) A frontal section of ShhN-treated animal was analyzed by immunohistochemistry using anti-NG2 (red) and anti-BrdU (green) antibodies at the level of the cerebral cortex. A double-labelled cell (boxed area) is observed by conventional microscopy. cc, corpus callosum. Scale bars (µm): 100 (b, c), 50 (d).

Table 2.   Quantification of bromodeoxyuridine-positive (BrdU+) cells in the adult mouse brain after amino terminal domain of Sonic hedgehog (ShhN) injection
 24 h48 h
  1. Data represent the number of BrdU+ cells counted in each brain section in the contralateral hemisphere. Each value is the mean ± SEM from three to five animals.

Cortex12 ± 185 ± 5
Corpus callosum15 ± 134 ± 2
Subventricular zone151 ± 9114 ± 10
Striatum6 ± 110 ± 2

To further demonstrate that cells of the oligodendroglial lineage respond to ShhN in both the corpus callosum and the cerebral cortex, frontal sections from 48 h ShhN-treated animals were stained for Ptc and the basic helix-loop-helix (bHLH) Olig1 transcription factor expressed by developing and adult oligodendrocytes (Lu et al. 2000; Zhou et al. 2000; Arnett et al. 2004). Immunohistofluorescent analysis demonstrated the presence of Olig1 cytoplasmic reactivity in cells located both in the corpus callosum and in the cerebral cortex (Figs 6b and f). Interestingly, we observed the presence of double-labelled Ptc+/Olig1+ cells in both regions (Fig. 6a–h) but not in the SVZ (data not shown). These data demonstrate that Olig1-expressing cells respond to ShhN in both the corpus callosum and the cerebral cortex of mice that received ShhN in the LV.

Figure 6.

 Patched (Ptc) transcript up-regulation occurs in Olig1+ oligodendroglial cells after amino terminal domain of Sonic hedgehog (ShhN) injection. Experiments using Patched (Ptc) in situ hybridization (a, e) and Olig1 immunohistofluorescence (b, f) were performed on frontal brain sections from ShhN-treated mice and analyzed at the level of the cerebral cortex (a–d) and corpus callosum (e–h). Cells coexpressing Ptc (green) and Olig1 (red) in the cortex (c, d) and corpus callosum (g, h), respectively, are indicated by arrowheads. Cell nuclei are labelled with DAPI (blue, d, h). Scale bars: 50 µm.

Analysis of the origin of newly generated cells

Whereas no effect on BrdU index has been found in the SVZ at 2 days after ShhN injection (see above), we have further investigated the source of the cells that can contribute to ShhN-induced proliferation. SVZ proliferating cells migrate along the rostral migratory stream to reach the olfactory bulb where they differentiate into GABAergic interneurons (Alvarez-Buylla and Lim 2004). As migrating neuroblasts expressing markers such as polysialylated neural cell adhesion molecule (PSA-NCAM) or doublecortin may migrate not only rostrally, but also caudally in specific experimental conditions to the corpus callosum and cortex (Magavi et al. 2000; Picard-Riera et al. 2002), we looked for the presence of cell chains labelled with these markers that would reflect such migration within the short time period after ShhN injection. In both the corpus callosum and cortex from ShhN-treated animals, we detected neither migrating PSA-NCAM nor doublecortin positive cells, nor chains of BrdU+ nuclei, which both would have reflected cell migration from the SVZ or from adjacent tissues (data not shown). In contrast, we observed a higher percentage of BrdU+ nuclei doublets (inset Fig. 2a) in the cortex of ShhN-treated mice (ShhN, 11 ± 2%, total counted cells n = 89; control, 6 ± 3%, n = 13; p < 0.05), strongly suggesting that the newly generated cells have their origin within the cortex itself.

Characterization of proliferating cell phenotype

In order to permit newly generated cells to reach a mature phenotype, we allowed another group of treated mice to survive 13 days after the last BrdU injection, which was carried out on day 2 (Fig. 7a). We then analyzed the phenotype of BrdU+ cells in the cortex on frontal sections from these animals. In ShhN-treated ‘day-15 mice’, Ptc mRNA expression was reduced as evidenced in the SVZ, the corpus callosum and the cortex compared to ‘day-2 mice’ (Figs 7b and 1b, c and e). However, the number of BrdU+ cells remains significantly higher in ShhN-treated mice compared to control (Fig. 7c; ShhN, 155 ± 13 cells per section, total counted cells n = 4330; control, 111 ± 11, n = 2337; p < 0.05). Double immunohistofluorescence experiments were carried out using antibodies against BrdU and several specific cell markers for OLP (NG2), astrocytes (GFAP) or differentiated neurons (NeuN). An increase in the number of BrdU+/NG2+ cells was still observed (ShhN, 32 ± 4%, total double-labelled cells counted n = 776; control, 21 ± 4%, n = 369; p < 0.05). The percentages of BrdU+/GFAP+ cells (ShhN, 29 ± 2, n = 121; control, 28 ± 4, n = 50) or BrdU+/NeuN+ cells (ShhN, 8 ± 3, n = 28; control, 8 ± 4, n = 16) were not different between ShhN- and vehicle-treated mice. Interestingly, a significantly higher number of newly generated BrdU+ cells was found juxtaposed to mature neurons (Fig. 7d) in the ShhN-treated group (ShhN, 45 ± 2%, n = 164; control, 34 ± 4%, n = 70; p < 0.05) reminiscent of the perineuronal satellite cells previously reported to belong to the astro- or oligodendroglial lineage (Kuhn et al. 1997). These data suggest that an acute injection of ShhN into the LV does not drive further differentiation of newly generated cells in the cortex into mature neurons or astrocytes.

Figure 7.

 Long-term effect of acute amino terminal domain of Sonic hedgehog (ShhN) injection in the lateral ventricle (LV) on Patched (Ptc) transcription and cell proliferation. (a) Experimental protocol. (b) In situ hybridization on a frontal section from a ShhN-treated animal indicating Ptc expression remains induced in the subventricular zone (SVZ) (short arrows) as well as in the other brain areas surrounding the injected (long arrow) and-non-injected side of the LV. (c) Bromodeoxyuridine-positive (BrdU+) cells were counted in the contralateral cortex from vehicle- or ShhN-treated mice. Cell numbers are means ± SEM, from five animals. *p < 0.05. (d) Double immunohistofluorescence using BrdU (green) and NeuN (red) on frontal sections from a ShhN-treated animal observed with conventional microscope. BrdU+ cells are juxtaposed to NeuN+ cells. DAPI (blue) labels the cell nuclei. cc, corpus callosum; Cx, cerebral cortex; ms, medial septum; st, striatum. Scale bars (µm): 200 (b), 50 (d).

Characterization of the adenovirus-mediated transfer of aminoterminal domain of Sonic hedgehog into the lateral ventricle of adult mouse

To obtain a sustained endogenous expression of ShhN in the adult brain, we used an adenoviral vector bearing the gene encoding the full-length human Shh protein and eGFP (AdhuShh-eGFP). This vector was developed in order to express the Shh precursor protein, which should be able to be fully processed to produce the active aminoterminal Shh product (Porter et al. 1995; Traiffort et al. 2004). CHO-K1 cells were first infected with AdhuShh-eGFP and the presence of a 22 kDa ShhN peptide produced in a time and concentration-dependent manner was identified in the culture medium by western blot analysis (Fig. 8a). Adult mice were then injected (75 × 106 pfu) with AdhuShh-eGFP or a control adenoviral vector (Ad-eGFP) into the right LV and were allowed to survive for 4 days (Fig. 8b) or for 26 days (Fig. 9a). eGFP fluorescence was detected at the level of the ependymal cell layer both in the ipsi- and contralateral sides on frontal sections from animals injected with both vectors (data not shown). These data confirmed that the adenoviral particles are able to move throughout the cerebrospinal fluid, as previously reported by others (Benraiss et al. 2001). At day 4, a robust Shh immunoreactivity was observed in infected cells bordering both the lateral and the third ventricle walls and was detected over several cell diameters inside the parenchyma in AdhuShh-eGFP-, but not in Ad-eGFP-treated animals (Figs 8c–e). This immunoreactivity corresponds to ShhN peptide identified by western blot from tissue sections analyzed at this level in AdhuShh-eGFP injected animals, suggesting the local production of ShhN peptide in the vicinity of SVZ cells (Fig. 8f). Further analysis of Shh signaling components indicated a robust increase of both Gli1 (Fig. 8g) and Ptc (data not shown) transcription in the same brain regions that were previously identified after recombinant ShhN injection, including the SVZ, the corpus callosum and the cortex (see Figs 1b–f), indicating activation of the pathway. Such transcription was not observed in tissue sections from Ad-eGFP animals (Fig. 8h and data not shown).

Figure 8.

 Characterization of Sonic hedgehog (Shh) signaling by adenovirus-mediated transfer of Shh in the lateral ventricle (LV). (a) Western blot on culture media harvested from CHO cells infected 24 or 48 h earlier with increasing concentrations of AdhuShh-eGFP indicated as plaque forming units (pfu) per cell. Amino terminal domain of Sonic hedgehog (ShhN) protein is detected as a 22 kDa signal. (b) Experimental protocol. (c–e) A strong Shh immunoreactivity is associated with the ependymal and subependymal cell layers (black arrowheads) of the injected (black arrows) and-non-injected side of the LV and with the ependymal layer of the third ventricle (3v, black and white arrow) of AdhuShh-eGFP (d, e), but not Ad-eGFP control mice (c).  (e) is an enlargement of the area boxed in (d).  (f) Western blot on brain tissue from Ad-eGFP (lane 2) or AdhuShh-eGFP (lanes 3, 4) mice. In AdhuShh-eGFP mice, ShhN and Shh are identified as major 22 and 47 kDa signals, respectively. Shh-infected CHO cells were used as positive control (lane 1). (g, h) Frontal sections from AdhuShh-eGFP (g) or Ad-eGFP (h) mice were hybridized with Gli1 riboprobe. Gli1+ cells are detected in the cortex (Cx), corpus callosum (cc) and SVZ of AdhuShh-eGFP mice only. ls, lateral septum; st, striatum. Scale bars (µm): 200 (c, d), 25 (e), 50 (g, h).

Figure 9.

 Adenovirus-mediated Sonic hedgehog (Shh) signaling orients newly generated cortical cells towards the oligodendrocyte (OL) lineage. (a) Experimental protocol. (b, c) Gli1 mRNA is observed throughout the cortex (Cx) from AdhuShh-eGFP (b), but not Ad-eGFP (c) mice. The inset in (b) is an enlargement of the boxed area. (d) DM20 in situ hybridization (blue) followed by bromodeoxyuridine (BrdU) immunohistochemistry (brown) shows that newly generated cells express (inset 1) or not (inset 2) DM20 transcripts in the cortex of AdhuShh-eGFP mice. A BrdU/DM20+ cell is shown (inset 3). Scale bars (µm): 50 (b, c), 12.5 (d). cc, corpus callosum; Cx, cerebral cortex.

Adenovirus-mediated production of aminoterminal domain of Sonic hedgehog orients newly generated cells to the oligodendroglial lineage

Animals that were allowed to survive until day 26, received a daily injection of BrdU between day 4 and day 18 to label proliferating cells (Fig. 9a). At day 26, the Shh pathway was still active throughout the cortex from AdhuShh-eGFP animals, as indicated by Gli1 transcription compared to control mice (Figs 9b and c). We then examined the percentage of cells double-labelled for BrdU and the oligodendroglial marker NG2 in the cerebral cortex on sections from these animals. We observed that the percentage of BrdU+/NG2+ cells was significantly higher in animals that received AdhuShh-eGFP compared to control animals (AdhuShh-eGFP, 39 ± 3%, total counted cells n = 436; control 32 ± 2%, n = 214; p < 0.05) (Fig. 10). To further investigate the phenotype of proliferating cells in the cerebral cortex and corpus callosum, first we used markers for premature (DM20 and MBP transcripts) or more mature (RIP, CNPase antigens) oligodendroglial cells. In the cortex of ShhN-treated mice, we observed a 50% increase in the percentage of BrdU+/DM20+ cells (AdhuShh-eGFP, 27 ± 1%, total counted cells n = 356; control, 18 ± 1%, n = 187; p < 0.01) (Figs 9d and 10). Interestingly, a similar increase in the percentage of BrdU+/DM20+ cells was found in the corpus callosum (AdhuShh-eGFP, 27 ± 1%, total counted cells n = 786; control, 19 ± 2%, n = 409; p < 0.01) (data not shown). Thus, these data indicate that newly generated cells in the cortex and corpus callosum express the transcript of the major splice variant of the proteolipid protein PLP. In contrast, in these regions, we did not observe a significant difference between the percentages of BrdU+ cells expressing the other oligodendroglial markers (MBP, RIP, CNPase) or markers of astrocytes (GFAP) and mature neurons (NeuN) (Fig. 10 and data not shown). Altogether these experiments indicate that the activation of Shh signaling in vivo in the adult brain is associated with an increase in cell proliferation in the cerebral cortex and corpus callosum and the regulation of brain precursor fate towards the first maturation steps of the oligodendroglial lineage.

Figure 10.

 Phenotypical characterization of amino terminal domain of Sonic hedgehog (ShhN)-induced proliferating cells in the cerebral cortex. AdhuShh-eGFP injection increases the number of bromodeoxyuridine-positive (BrdU+) cells expressing the NG2 proteoglycan or DM20 transcripts, but not myelin basic protein (MBP) transcripts, 2′,3′-cyclic nucleotide 3′-phosphohydrolase (CNPase), RIP, glial fibrillary acidic protein (GFAP) or NeuN markers. Values derive from counting of double-labelled cells observed in immunohistofluorescence experiments and are given as percentage of the total number of BrdU+ cells. They are means ± SEM from six mice for each group. The percentages obtained for double-labelled BrdU+/DM20+ cells correspond to 356 and 187 counted cells for Shh and control animals, respectively. **p < 0.01; *p < 0.05.

Discussion

We have delivered active forms of Shh in the LV using various strategies in order to determine its potential role in the regulation of neural progenitors in the mature brain. Our in vivo experiments demonstrate that the number of proliferating progenitors in the cerebral cortex and the corpus callosum can be increased upon brain administration of Shh proteins. The up-regulation of the transcripts of the target gene Ptc in both areas and their expression in Olig1+ cells indicate that cells of the oligodendroglial lineage are responsive to Shh proteins. They also indicate that Shh signaling is further involved in the expansion of the proliferating cell populations expressing the NG2 proteoglycan in the cerebral cortex and the DM20 oligodendroglial marker in cortical and callosal regions.

The decrease of BrdU+ cells in the SVZ of animals treated for 48 h compared to animals treated for 24 h (Table 2) may indicate a mobilization of cells from the SVZ to the corpus callosum and the cerebral cortex. However, several arguments are in supportive of an in situ proliferation of cells in both the corpus callosum and cerebral cortex. First, we did not detect any increase in the number of BrdU+ cells in the SVZ, as already reported after activation of the Shh pathway in this region consecutive to an intrastriatal injection of ShhN (Charytoniuk et al. 2002). Second, we did not observe chains of BrdU+ cells or migrating neuroblasts in the corpus callosum or the cerebral cortex, which would have reflected the mobilization of cells displaying a backward and radial migration from the SVZ as previously described (Magavi et al. 2000; Picard-Riera et al. 2002). Third, the significant increase in the number of doublets of BrdU+ cells in the cerebral cortex is in accordance with an in situ proliferation. Finally, the presence of double-labelled Ptc+/BrdU+ cells in the corpus callosum and cerebral cortex of ShhN-treated animals clearly indicates the ability of a population of proliferating cells in those regions to respond to Shh signaling. It is worthwhile to notice that the increase of transcription of the target gene Ptc in the corpus callosum and cerebral cortex parallels the increase of proliferation of cells in these areas. However, the cortex constitutes an area remote from the site of delivery of Shh protein, which leads to question about the mechanism through which Shh signaling pathway can be activated. We have previously demonstrated that ShhN is axonally transported in an anterograde manner from the hamster retina to the superior colliculus (Traiffort et al. 2001) and lesion of the fornix interrupts basal forebrain anterograde transport of ShhN to the hippocampus (Lai et al. 2003). Therefore, after Shh protein delivery, ShhN present in the LV might reach its target cells by axonal transport after being endocytosed by neurons close to one of the brain ventricles. This might be the case for cells of the basal amygdaloid nucleus bordering the LV and projecting to the frontal cortex (Kita and Kitai 1990). It is noteworthy that recombinant ShhN used in the present study increases proliferation of cerebellar granule cell precursors, differentiation of the cell line C3H10T1/2 and controls electrophysiological properties of differentiated neurons at a subnanomolar concentration (Pascual et al. 2005). In our present study, we observed a strong activation of Ptc and Gli1 transcription in several brain areas, including the corpus callosum and the cerebral cortex, suggesting that the concentration of ShhN was sufficient to activate the pathway. Nevertheless, we failed to detect ShhN by immunoreactivity in these regions, presumably reflecting the limit of detection of our immunohistochemistry protocol. However, the present data remind the report made by others of newly generated BrdU+ cells in several parenchymal areas, including the cerebral cortex, following the intracerebroventricular injection of brain-derived neurotrophic factor (BDNF). Those cells were distributed as a decreasing gradient from the closest to more remote regions around the lateral and third ventricles and were suggested, in part, to be associated with cell division occurring in situ in the parenchyma (Pencea et al. 2001).

Other growth factors, such as FGF (fibroblast growth factor), EGF (epidermal growth factor) or BDNF, have already been reported to induce cell proliferation in the adult cerebral cortex, but neither the extent of the process nor the precise fate of the newly generated cells have been definitively characterized (Kuhn et al. 1997; Pencea et al. 2001). Our data do not support a potential indirect effect of Shh via the induction of such endogenous trophic factors. Indeed, exogenous FGF and EGF infused in the LV (Kuhn et al. 1997) were both reported to increase the number of newborn cells in the SVZ, which is not observed upon administration of ShhN. In addition, BDNF infusion did not influence the number of BrdU+ cells in the frontal cortex (Benraiss et al. 2001) unless BrdU is infused directly into the LV for several days (Pencea et al. 2001).

In the adult brain, NG2-expressing cells have been reported to display features of stem cells distributed throughout the entire gray and white matter (Dawson et al. 2003). Recent data indicate that NG2+ cells in the SVZ can give rise to interneurons in the hippocampus and the olfactory bulbs (Aguirre and Gallo 2004; Aguirre et al. 2004) and that a small population of NG2+ cells residing within the cortex itself may be the source of new neurons (Dayer et al. 2005). These cells are, however, mainly characterized as OLP able to differentiate into oligodendrocyte (OL) in vitro and in vivo, and potentially increase the population of mature OL during adult life (Horner et al. 2002; Dawson et al. 2003). They can be mobilized from the SVZ in response to the demyelination process occurring in an animal model of experimental autoimmune encephalomyelitis (Picard-Riera et al. 2002). Therefore, NG2+ cells are considered as a promising therapeutic approach towards alleviating extensive demyelination processes.

Here, we have shown that Shh signaling promotes the proliferation of NG2+ cells in the cortex as evidenced by the marked increase in the number of these cells upon delivery of an active form of ShhN in the LV. The transcriptional induction of DM20 in BrdU+ cells in the cortex and corpus callosum of AdhuShh-eGFP-treated day-26 mice further supports the idea that Shh signaling can drive progenitor cell maturation towards the oligodendroglial lineage. During development, DM20 is expressed by premyelinating OL before PLP and is required for the maturation of these cells (Spassky et al. 1998; Wight and Dobretsova 2004). The 50% increase in the percentage of BrdU+/DM20+ cells in the cortex and corpus callosum of animals treated with the Shh adenoviral vector, indicates a strong effect on this cell population. Whether Shh responsive NG2+ and DM20+ cells in the cerebral cortex actually represent consecutive steps of the same lineage or two distinct types of progenitors remains to be further investigated. The phenotypic characterization of dividing cells in AdhuShh-eGFP-treated day-26 mice excludes that NG2+ cell increase is associated with the generation of new neurons in the cortex, which suggests that Shh does not influence adult neurogenesis in the cortex in contrast to the hippocampus (Lai et al. 2003). The absence of a significant increase in the number of NG2+ cells in the corpus callosum compared to the increase observed in the cortex might be accounted for by the heterogeneity described in the NG2+ cell population, mainly characterized at the level of their K+ and Na+ channel expression profiles and electrophysiological properties (Chittajallu et al. 2004).

The cerebral cortex and corpus callosum both contain progenitors sensitive to the proliferative activity of Shh, as previously found in the hippocampus (Lai et al. 2003), whereas the SVZ and striatum are insensitive to this effect, despite a potent activation of the Shh pathway. This result indicates that Shh effects are region-specific and is in agreement with the absence of a direct mitogenic effect of Shh in neurospheres from adult SVZ (Palma et al. 2005). However, these data disagree with the increase and the decrease in cell proliferation observed in the adult mouse SVZ after peripheral administration of a Smo agonist and antagonist, respectively (Machold et al. 2003; Palma et al. 2005). It is possible that Smo activation can be achieved independently of Shh signaling and that some of the Ptc effects can be mediated independently of Smo. Interestingly, the anti-apoptotic effect of Shh in the neural tube was suggested to be Ptc-related and Smo-independent (Thibert et al. 2003).

The mitogenic effect of Shh reported here on NG2+ cells in the cerebral cortex from adult mice reminds that Shh was previously described as a selective mitogen for OLP derived from E17 rat cortical plate (Murray et al. 2002), but differs from results reporting Shh is not required for OLP proliferation in chick embryo spinal cord (Orentas et al. 1999). Such discrepancies may be linked to species or tissue differences and to the competence of cell precursors previously reported to allow cell proliferation in response to activation of Shh pathway only at selected periods during embryogenesis (Rowitch et al. 1999).

In the developing cerebral cortex, the oligodendrogenesis process involves several steps including mainly the early Shh and FGF-dependent formation of OLP in the ventral forebrain (Tekki-Kessaris et al. 2001; Kessaris et al. 2004), and a second postnatal wave of precursors originating in the SVZ (Ivanova et al. 2003). In addition, a local production of OLP derived from Emx-1 expressing cells was suggested to occur in the cerebral cortex after E17 and to give rise to OL in the cortex, corpus callosum and fimbria (Gorski et al. 2002). Mice invalidated for the receptor Smo subsequently to E12.5 stage have confirmed the role of Shh signaling in supporting the generation of OL from the ventral forebrain and indicated Shh requirement for the generation of pallial OL later during development (Machold et al. 2003). Whether Shh signaling is involved in the generation of new OL in the adult brain, via a mechanism that may recapitulate the developmental processes described above requires further work using specific and selective pharmacological tools. In this respect, the up-regulation of the Nkx2.2 and Olig2 genes that are early markers of OL lineage, has been observed in a demyelination model in the adult brain (Fancy et al. 2004; Watanabe et al. 2004). Further work should delineate the specific combinatorial transcriptional code induced by ShhN in newly generated cells and thus determine why this combination is apparently not sufficient to confer MBP promoter activation. It was recently shown that progressive stages of OL lineage maturation are characterized by dynamic changes in the subcellular distribution and combinatorial profiles of several transcriptional regulators (Gokhan et al. 2005).

Taken together with the persistence of NG2+ cells observed in an experimental autoimmune encephalomyelitis animal model (Reynolds et al. 2002) or in demyelinating plaques of multiple sclerosis patients (Chang et al. 2000), the proliferative effect of Shh on NG2+ cells and its transcriptional effect on DM20 gene might be of potential therapeutic interest for demyelinating diseases and complementary to the survival effect recently reported for leukaemia inhibitory factor (LIF) (Butzkueven et al. 2002) and ciliary neurotrophic factor (CNTF) (Linker et al. 2002). In agreement with our results and the above mentioned hypothesis, exogenous Shh administration promotes an increase of OLP after adult rodent spinal cord demyelination (Bambakidis et al. 2003). Moreover, the induction of Shh and its receptor Ptc was recently suggested to be required for overcoming the maturation-inhibitory effects of the Notch-1 signaling pathway in a model of nonautoimmune demyelination (Mastronardi et al. 2004).

Acknowledgements

We thank A. Galdes and D. P. Baker (Biogen Idec, Cambridge, USA) for ShhN protein, A. Carleton and P. M. Lledo for technical advice and V. Zuliani and the Production Service Unit of Genethon for large scale production of adenoviral vectors within the Gene Vector Production Network (GVPN, http://www.genethon.fr/gvpn) program. KL is the recipient of a doctoral fellowship from the ‘Ministère de la Recherche’ and from La Ligue contre le Cancer and Association pour la Recherche contre le Cancer (ARC). This work was supported in part by a grant from Association pour la Recherche sur la Sclérose en Plaques (ARSEP), La Ligue contre le Cancer, Association pour la Recherche contre le Cancer (ARC) and Fondation pour la Recherche sur le Cerveau (FRC).

Ancillary