 |
||||
|
||||
|
||||
|
||||
Previous Article | Next Article
The Journal of Neuroscience, July 1, 2001, 21(13):4740-4751
Neuregulin Signaling Regulates Neural Precursor Growth and the Generation of Oligodendrocytes In Vitro
Viviane Calaora1, Bernard Rogister2, Keren Bismuth1, Kerren Murray1, Heidi Brandt2, Pierre Leprince2, Mark Marchionni3, and Monique Dubois-Dalcq1 1 Neurovirologie et Régénération du Système Nerveux, Institut Pasteur, 75724 Paris, France, 2 Centre de Recherches en Neurobiologie Cellulaire et Moléculaire, Université de Liège, Liège, 4020 Belgium, and 3 Cambridge NeuroScience, Inc., Norwood, Massachusetts 02139
ABSTRACT
TOP
ABSTRACT
INTRODUCTION
Neuregulin 1 (Nrg-1) isoforms have been shown to influence the emergence and growth of oligodendrocytes, the CNS myelin-forming cells. We have investigated how Nrg-1 signaling of ErbB receptors specifically controls the early stages of oligodendrocyte generation from multipotential neural precursors (NPs). We show here that embryonic striatal NPs express multiple Nrg-1 transcripts and proteins as well as their specific receptors, ErbB2 and ErbB4, but not ErbB3. The major isoform synthesized by striatal NPs is a transmembrane type III isoform called cysteine-rich domain Nrg-1. To examine the biological effect of Nrg-1, we added soluble ErbB3 (sErbB3) to growing neurospheres. This inhibitor of Nrg-1 bioactivity decreased mitosis of NPs and increased their apoptosis, resulting in a significant reduction in neurosphere size and number. When NPs were induced to migrate and differentiate by adhesion of neurospheres to the substratum, the level of type III isoforms detected by RT-PCR and Western blot decreased in parallel with a reduction in Nrg-1 fluorescence intensity in differentiating astrocytes, neurons, and oligodendrocytes. Pretreatment of growing neurospheres with sErbB3 induced a threefold increase in the proportion of oligodendrocytes generated from NPs migrating out of the neurosphere. This effect was not observed with an unrelated soluble receptor. Addition of sErbB3 during NP growth and differentiation enhanced oligodendrocyte maturation as shown by expression of galactocerebroside and myelin basic protein. We propose that both type III Nrg-1 signaling and soluble ErbB receptors modulate oligodendrocyte development from NPs.
Key words: neuregulin; neural stem cells; neurosphere growth; oligodendrocyte differentiation; ErbB receptors; embryonic mouse striata
INTRODUCTION
There is much interest in the molecular signals that regulate development of multipotential neural precursors (NPs) and their fate (Edlund and Jessell, 1999
). Among candidate regulators are neuregulin 1 (Nrg-1) growth factors and ErbB receptor tyrosine kinases (Carraway and Burden, 1995
How NRG-1 controls early stages of oligodendrocyte development is not clear. Oligodendrocyte progenitors (OPs) are triggered to divide by GGF2, and oligodendrocyte survival is enhanced by NDF and axon-derived Nrg-1 (Canoll et al., 1996
To investigate this question, the neurosphere system appears particularly appropriate (Reynolds and Weiss, 1996
MATERIALS AND METHODS
Preparation of neurosphere cultures. C57-Bl6 mice 14 d of age (IFFA-CREDO) were killed in compliance with animal regulations of the Pasteur Institute. The striata were removed from the embryos and mechanically dissociated before cells were plated, as described previously (Reynolds et al., 1992
Adhesion of neurospheres. After 6 d in culture, three neurospheres of the same size (80-100 µm in diameter) were plated onto 12 mm coated coverslips (poly-D-lysine/fibronectin; 10 µg/ml each; Sigma) and 2 ml of serum-free medium without EGF was then added to each well. After 2 d of adhesion, the majority of cells had migrated out of the sphere.
Neurosphere treatment. Recombinant Human GGF2 (Cambridge NeuroScience Inc., Norwood, MA) was used at 2-200 ng/ml. For experiments with soluble receptors (see description of production and purification below), neurospheres were grown for 6 d as above in the presence of different concentrations of sErbB3 (0.1-1 µg/ml) or soluble subunit
of the glycine receptor (sGlyr; 0.5 or 1 µg/ml). Every 2 d, EGF, sErbB3, and sGlyr were replenished. In some experiments, GGF2 (100 ng/ml), sErbB3 (0.5 µg/ml or 6.6 µM), and sGlyr (0.5 µg/ml or 7.6 µM) were added just after sphere adhesion for 2 d before immunocytochemical staining.
RT-PCR analysis. Total RNA was purified with an RNAble kit (Eurobio, Les Ulis, France) from E14 striata, adult spinal cord, immortalized Schwann cell cultures (MSC80; courtesy of A. Baron-Van Evercooren, la Salpêtrière, Paris, France), or neurospheres cultured in suspension for 6 d or after 6 d of adhesion. Seven micrograms of total RNA were used in RT reactions using Superscript II (Life technologies, Gaithersburg, MD). One microliter of a 1:7 dilution of the cDNA synthesized from the RT reaction was subjected to PCR in a 50 µl reaction with 0.2 mM dNTP, 1.5 mM MgCl2, 0.5 mM primers, and 1 µl (3.5 U) of Taq polymerase (Expand High Fidelity PCR system, Roche, Meylan, France). Primers used to analyze NRG-1 mRNAs encoding different isoforms are described in Table 1 and localized in Figure 1. Reactions were run on a Perkin-Elmer 2400 thermocycler (Roissy-Charles de Gaulle, France) (30 cycles; 94°C for 2 min, 94°C for 15 sec, 53°C for 30 sec, 68°C for 3 min, and a final extension time at 72°C for 7 min).
View this table: [in this window]
[in a new window]
Table 1. Primer localization and sequences Specific primers designed to amplify ErbB2, ErbB3, and ErbB4 are listed in Table 1. cDNA synthesized by RT was diluted at 1:2 before PCR amplification, then processed as described above with an annealing temperature of 60°C.
Western blot. Proteins were extracted as described by Raabe et al. (1996)
Immunocytochemistry of tissues and cells. The following antibodies were used at the indicated dilution: goat polyclonal antibody to the C terminus of the mature form of human Nrg-1 (HRG-C19; 1:200) and rabbit polyclonal antibody to the C terminus of the precursor form of human Nrg-1 (HRG-C20; 1:200) were obtained from Santa Cruz Biotechnology. The rabbit polyclonal antibodies against the cytoplasmic domain of the precursor form of human ErbB2 (Neu, C-18; 1:100) and ErbB4 (C-19; 1:100) were also from Santa Cruz Biotechnology. Control antibodies for Nrg-1 and ErbB staining were non-immune goat and rabbit IgG used at the same final concentration (Jackson ImmunoResearch Laboratories, West Grove, PA). TujI is a mouse IgG2a antibody against neuron-specific class III
Cryosections (10 µm)on Superfrost Plus glass slides (Menzel-Glaser, Braunschweig, Germany) were prepared from striata of E14 mice or from floating neurospheres that had been fixed in 4% paraformaldehyde in PBS overnight at 4°C. Adherent neurospheres were fixed the same way for 30 min at room temperature before staining. Primary antibodies were applied overnight at 4°C or 30 min at room temperature (O4 staining) in a humidified chamber. Slides were then incubated in appropriate secondary antibodies and with DAPI for 30 min at room temperature. For intracellular labeling with HRG-C19, HRG-C20, ErbB2, and ErbB4 antibodies, a permeabilization with PBS 0.1% Triton X-100 was performed for 20 min at room temperature before blocking with normal goat serum. Double and triple immunolabeling were performed as described previously (McKinnon et al., 1990
Preparations were photographed with a Leica microscope (DMRB model) equipped with a Coolsnap camera system (Princeton Instruments, Evry, France). For the confocal analysis, a Leica Optovar inverted microscope equipped with a Leica model TCSAD Confocal Instrument was used (courtesy of E. Perret, Department of Retrovirology, Pasteur Institute, Paris).
Purification of soluble receptors. The soluble form of ErbB3 receptor was constructed using the ectodomain of ErbB3 cloned in frame to an Fc portion (hinge, CH2, and CH3 domains) of human immunoglobulin G, called CDM7-IgB-3, courtesy of Y. Yarden (Chen et al., 1996
We extracted 40 × 106 cells in 40 ml of an ice-cold buffer containing 8 M urea, 0.1 M NaH2PO4, and 0.01 M Tris, pH 8.0. Cells were homogenized in a Potter tissue homogenizer and then centrifuged for 15 min at 15,000 × g at 4°C. The supernatant was loaded on an Ni-NTA column (Qiagen, Westburg, Leusden, The Netherlands) equilibrated with extraction buffer. The use of urea as a denaturing agent resulted in enhanced attachment of recombinant sErbB3 to Ni-NTA and therefore increased the yield of purification. After loading, the column was washed with a 0-100% gradient of refolding solution (0.02 M Tris, pH 7.4, 500 mM NaCl, and 20% glycerol) at 0.5 ml/min for 90 min. Two steps of elution were performed in refolding solution containing successively 10 and 250 mM imidazole. sErbB3 was present in the last elution step as demonstrated by Western blotting using anti-myc antibodies (see Fig. 4A). sGlyr was purified as described for sErbB3 on Ni-NTA gel except that the extraction buffer also contained 0.25% sodium deoxycholate. The refolding step was replaced by a wash in extraction buffer adjusted to pH 6.3, and the elution was performed with extraction buffer adjusted to pH 4.5 and containing 2% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid. sGlyr was selectively eluted as demonstrated in Figure 4A and appeared as a 75 kDa species. The higher than predicted molecular weight (MW) is probably attributable to the MW of the Tag (6.7 kDa) and glycosylation. Silver gel performed at the end of each soluble receptor purification did not reveal any contaminant. To test the biological activity of sErbB3, a tritiated thymidine assay was run on rat Schwann cells (Mahanthappa et al., 1996
Cell proliferation assay. Two different assays were used. First, the neurospheres were grown in suspension for different times and in the presence of 10 µM bromodeoxyuridine (BrdU; Sigma) for the last 24 hr. These spheres were plated on poly-D-lysine/fibronectin-coated coverslips for 30 min at 37°C to allow neurosphere adhesion but not differentiation. Neurospheres were then fixed for 30 min at room temperature. The spheres were then processed for BrdU staining (Ben-Hur et al., 1998
Second, the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT; Sigma) reduction assay (Mosmann, 1983
O.D. at 630 nm. Background absorbance of medium in the absence of cells was subtracted.
TUNEL analysis. TUNEL method was used to label apoptotic nuclei in floating neurospheres as described previously (Gavrieli et al., 1992
Statistics. For each experimental condition, cell counting was performed blind in three independent experiments. The total number of spheres and the size of 50 floating neurospheres in triplicate Petri dishes were analyzed with a Leica microscope (DMIL) equipped with a graduated eye-piece. To determine the percentages of BrdU- and TUNEL-positive cells in each experiment, all cells in 30 spheres per slide were counted with a Leica microscope (DMRB model). In experiments with adherent neurospheres, we counted the number of each type of differentiated cells and the total number of cells per sphere. In the latter case, we scanned the DAPI staining of the entire sphere outgrowth with the Coolsnap camera and counted the nuclei with NIH Image software (Version 1.62).
The mean fluorescence per cell of Nrg-1 expression was measured for NPs and differentiated cells with NIH image software on confocal images of neurospheres stained with either HRG-C20 or HRG-C19 antibodies. Data points from the MTT assay were averages from five independent experiments. All values are shown as mean ± SEM. The statistical significance of differences between control and experimental conditions was analyzed using the one-tailed Student's t test (Excel 98). A value of p < 0.05 is considered statistically significant (*p < 0.05; **p < 0.02; ***p < 0.01).
RESULTS
To study Nrg-1 signaling in multipotential NPs, we isolated the striatal periventricular zone of E14 mouse embryos and cultured the dissociated cells as neurospheres (Svendsen and Rosser, 1995
Neuregulin 1, ErbB2, and ErbB4 are expressed in E14 striatum and neurospheres
We first characterized by RT-PCR and Western blot which Nrg-1 isoforms were expressed in striatum and in neurospheres. We designed sets of primers that distinguish between the different amino termini of type I and/or II as well as type III isoforms by positioning the forward primer in regions specific for each amino terminus and positioning the reverse primer in the EGF-like domain present in all NRG-1 isoforms (see Materials and Methods) (Fig. 1A). The adult spinal cord was used as positive control because the three types of NRG-1 isoforms are expressed in this region (Shinoda et al., 1997
View larger version (35K):[in this window]
[in a new window]
Figure 1. RT-PCR and Western blot analysis of NRG-1 isoforms in E14 striatum and neurospheres. A, Schematic representation of the different splice variants of the NRG-1 gene modified from Adlkofer and Lai (2000)
In adult spinal cord and E14 striatum, Ig-EGF primers detected three bands from 480 to 370 bp that correspond to both type I and II isoforms either with or without a glycosylation domain (Fig. 1B). A weak signal was detected only at 480 bp in neurospheres (Fig. 1B). GGF2-EGF primers also detected three bands of type II isoforms (from 570 to 460 bp) in spinal cord and E14 striata, but there was no signal in neurospheres (Fig. 1B). In contrast, the type III isoform (CRD-EGF primers) gave an intense 500 bp band and a weaker band at 440 bp in spinal cord, striatum, and growing neurospheres with a decrease in intensity in differentiated spheres (Fig. 1B). Because the type III NRG-1 transcripts appear the most abundant, we designed other sets of primers that parse type III transcripts into secreted (Fig. 1A, s) and transmembrane isoforms with a long and short cytoplasmic tail (Fig. 1, A, C, a and c, respectively). Although the spinal cord and E14 striatum expressed these three type III isoforms, only the secreted form (s, 732 bp band) and the transmembrane form with the long C-terminal domain (a, 1852 bp band) were weakly detected in floating neurospheres (Fig. 1C). The 1852 bp product was cloned into a pCRII vector, and sequencing of seven clones revealed coding regions for type III transmembrane type (a) isoforms, corresponding to
,
We then analyzed ErbB2, -3, and -4 expression in E14 striata and growing and adherent neurospheres. Although we consistently detected both ErbB2 and ErbB4 by RT-PCR and Western analysis, an ErbB3 signal was not detected in these samples (Fig. 2A,B). Positive controls include Schwann cells, which showed a strong ErbB3 signal using both techniques but no ErbB4 (Fig. 2A,B) (Grinspan et al., 1996
View larger version (43K):[in this window]
[in a new window]
Figure 2. Expression of ErbB receptors in neurospheres before and after adhesion. A, RT-PCR analysis: RNA was extracted from four samples (denominated as in Fig. 1D) and from Schwann cells used as positive control. ErbB2 gave a signal in all samples. ErbB3 is absent in E14 striata and neurospheres but present in Schwann cells and adult spinal cord. ErbB4 is stronger than ErbB2 in all CNS samples and absent in Schwann cells. B, Western blot analysis was performed on extracts denominated as in Figure 1. Anti-ErbB2 antibodies detect a signal ~180 kDa in all samples. Anti-ErbB3 antibodies detect a band (160 kDa) only in cultured Schwann cell extracts but not in E14 striata or growing or adhering neurospheres. Anti-ErbB4 antibodies detect a 185 kDa signal in all samples except Schwann cells.
Together, these results indicate that in E14 striatum and NPs derived from this region, the predominant Nrg-1 products consist of type III 80 kDa transmembrane isoforms, which are downregulated in differentiating neurospheres. These Nrg-1 isoforms most likely signal through ErbB2-ErbB4 heterodimers or ErbB4 dimers expressed by NPs.
Nrg-1 and ErbB receptors are coexpressed in neural precursors but levels of expression vary during differentiation
To determine whether Nrg-1 isoforms and their receptors were expressed in the same or different NPs in vivo and in vitro, we used confocal fluorescence microscopy to examine cryosections of E14 striata and floating neurospheres double labeled with polyclonal antibodies to the mature form of Nrg-1 (HRG-C19 antibody) and to the ErbB2 or ErbB4 receptor. Merging of the two images taken in the fluorescein (ErbB receptor) and rhodamine (Nrg-1) mode showed that the majority of striatal cells (Fig. 3A,B) and NPs (Fig. 3C,D) coexpressed at least one of these two ErbB receptors and Nrg-1. ErbB2 and ErbB4 staining was detected both in the cell cytoplasm and at the membrane whereas Nrg-1 staining was mostly associated with the membrane (Fig. 3A'-D'). The orange color resulting from colocalization of Nrg-1 with ErbB2 or ErbB4 was mostly associated with the plasma membrane and appears more frequent in cells at the periphery of the neurosphere. These observations correlate well with the synthesis of a transmembrane Nrg-1 isoform by NPs and suggest that signaling through the ErbB2 and ErbB4 receptors occurs mostly in an autocrine or paracrine way, or both.
View larger version (73K):[in this window]
[in a new window]
Figure 3. Nrg-1, ErbB2, and ErbB4 expression in E14 striatum and neurospheres. A-D, Confocal imaging was done after double immunolabeling with goat antibody to human Nrg-1 (HRG-C19, rhodamine in A-D) and a rabbit antibody against human ErbB2 (fluorescein; A, C) or ErbB4 receptor (fluorescein; B, D). The merged images show the coexpression of each ErbB receptor with Nrg-1 in orange in striatum (A, B) and neurospheres (C, D). A'-D', The strong membrane signal of individual cells (arrows) is shown in the close-up corresponding to the frames indicated in A-D. V, Ventricule. E-J, Expression of Nrg-1 in differentiated cells in the outgrowth from neurospheres. Note that Nrg-1 is in red in E and F and in green in G-J and that nuclei are stained blue by DAPI. E, F, Six days after sphere adhesion, many cells in the outgrowth are preoligodendrocytes labeled with O4 antibody (green). Single (E) and double (F) label show scattered dots of Nrg-1 (red, HRG-C20 antibody) colocalized with O4 staining. White arrows in F indicate three double-labeled cells: one is strongly O4 positive (left arrow), whereas the two cells at right arrows are weakly stained with O4 antibodies. Nrg-1 is also detected in a double-labeled spindle-shaped OP in the left corner of F. G, H, Neuronal cells stained with TujI antibody (red) express Nrg-1 (green, HRG-C20 antibody) as indicated by the orange label in three cells indicated by white arrows in H. I, J, A cell with a weak GFAP staining (red) shows a strong signal for Nrg-1 (green, HRG-C20 antibody) resulting in orange labeling of this immature astrocyte (J, white arrow). In contrast, strongly GFAP-positive astrocytes (in red) are not stained by Nrg-1 antibody (compare I and J at the two black arrows). Scale bars, 10 µm. K, Quantitative analysis of confocal images of cells stained with two different antibodies to Nrg-1 proteins (HRG-C19 and HRG-C20). Fluorescence intensity of Nrg-1 labeling decreased in differentiated cells of adherent neurospheres after 6 d when compared with undifferentiated NPs within neurospheres.
We then examined whether expression of the Nrg-1 and ErbB receptors changes during differentiation of NPs into different CNS cell types after adhesion of neurospheres. Scattered, punctuate staining of Nrg-1, ErbB2, and ErbB4 was detected in both undifferentiated and differentiated cells of the sphere outgrowth (Figs. 3, 4). Quantitative analysis by confocal microscopy showed an important decrease of intensity of Nrg-1 expression when compared with NPs in neurospheres (Fig. 3K). The comparison between single labeling for Nrg-1 and double labeling with cell-specific markers showed that Nrg-1 was mostly expressed by O4-positive preoligodendrocytes (Fig. 3E,F, white arrowheads) as well as by multiprocessed neuronal cells stained for the isoform III of tubulin in their cell bodies and processes (Fig. 3G,H, white arrowheads). In contrast, some mature astrocytes were negative for Nrg-1 (Fig. 3I,J, black arrowheads), whereas immature astrocytes showed more Nrg-1 staining (Fig. 3I,J, white arrowheads). Double labeling illustrated in Figure 4 showed that ErbB2 was expressed by the majority of differentiated cells (Fig. 4A,C,E, white arrowheads), although some oligodendrocytes, neurons, and astrocytes were negative for ErbB4 (Fig. 4B,D,F, black arrowheads). Thus the majority of differentiated cells still express Nrg-1 and its two receptors.
View larger version (76K):[in this window]
[in a new window]
Figure 4. ErbB expression in differentiated cells. Double labeling with nuclei stained blue by DAPI. Note that ErbB receptors are in red in A and B and in green in C-F. A, B, ErbB2 (A) and ErbB4 (B) (red) are detected in both the cell body and processes of preoligodendrocytes producing an orange staining of the cytoplasm around the blue nucleus (white arrowheads). In B, note one oligodendrocyte strongly stained by O4 (green), which is ErbB4 negative (black arrowhead). C, D, Neurons labeled with TujI (red) and ErbB2 (green) show orange label in their cell bodies and processes (C, white arrowheads). In contrast, only a few TujI-positive cells express ErbB4 (orange cells in D at white arrowheads). Other bipolar neuronal cells are mostly red, showing weak or no labeling for ErbB4 (D, black arrowheads). E, F, All astrocytes coexpress GFAP (red) and ErbB2 (green), resulting in orange staining (white arrowheads), but some do not express ErbB4 (red-labeled astrocytes at black arrowheads in F). Scale bar, 10 µm.
Nrg-1 controls the growth and survival of neurospheres
To investigate the effects of Nrg-1 on NPs, we first tested whether addition of GGF2 (at 50 or 200 ng/ml) to the medium of growing neurospheres would increase their rate of formation or their size, but we observed no significant effect compared with untreated spheres after 6 d of growth (data not shown). We also investigated whether ErbB2 and ErbB4 receptors were phosphorylated after addition of GGF2, using immunoprecipitation followed by phosphotyrosine antibody labeling (Vartanian et al., 1997
Given the expression of Nrg-1 isoforms and ErbB receptors in neurospheres (Figs. 1-4), we reasoned that endogenous Nrg-1 had probably saturated the receptors at 6 d. We therefore used a different assay to test GGF2 effects in conditions of minimal endogenous Nrg-1. Freshly dissociated striatal NPs were seeded with different concentrations of GGF2, and their growth was measured after only 2 d by MTT assay. In this condition, GGF2 addition produced a small but significant increase in cell growth (1.5-fold with 2 ng/ml of GGF2, twofold with 20 ng/ml of GGF2, and 1.8-fold with 200 ng/ml of GGF2 in comparison with control condition). These results suggest a role of Nrg-1 in the early growth of NPs. Therefore, we engineered a cDNA encoding sErbB3, the amino-terminal domain of which specifically binds to Nrg-1 ligands, which we purified from transfected CHO cells. Although soluble ErbB4 has been used before as an Nrg-1 inhibitor (Pinkas-Kramarski et al., 1997
View larger version (40K):[in this window]
[in a new window]
Figure 5. Effect of soluble ErbB3 on neurosphere formation and growth. A, Purification of soluble ErbB3 (sErbB3) and
View this table: [in this window]
[in a new window]
Table 2. sErbB3 biological activity on rat Schwann cells To evaluate the effects of this Nrg-1 inhibitor on neurosphere growth, we added 0.1 µg/ml to 1 µg/ml of sErbB3 or 1 µg/ml of sGlyr at the time of seeding of the NPs. sErbB3 decreased the neurosphere size but did not inhibit neurosphere formation (Fig. 5B). After 6 d of growth in suspension, there was a significant and dose-dependent reduction in both the size and number of neurospheres growing in the presence of sErbB3, whereas addition of sGlyr had no effect compared with neurospheres grown in control medium (Fig. 5C,D). To determine whether sErbB3 inhibited cell proliferation or survival, or both, we first evaluated DNA synthesis by adding BrdU during the last 24 hr to neurospheres cultured for 2 or 6 d in the presence or absence of sErbB3 (0.5 µg/ml) and sGlyr (0.5 µg/ml) (Fig. 5E). The same experimental design and time points were used to assay apoptosis by the TUNEL method (Fig. 5F). sErbB3 significantly decreased cell proliferation and increased the percentage of apoptotic cells after 2 or 6 d in culture (Fig. 5E,F). However, these effects varied with time. At 2 d, sErbB3 decreased the proliferative index 4.5-fold (from 27 ± 1% in control condition to 6 ± 3%), whereas at 6 d the decrease was merely 13% (from 81 ± 2% to 68 ± 2%) (Fig. 5E), suggesting either a low efficiency of sErbB3 blocking after 6 d or that other mitogens were stimulating proliferation at this stage. In contrast, sErbB3 treatment increased NP apoptosis two to three times at both time points, whereas there was no significant change in the apoptosis index between 2 and 6 d in control and sGlyr-treated cultures (Fig. 5F). These experiments demonstrate that endogenous Nrg-1 enhances the growth and survival of NPs in neurospheres.
sErbB3 treatment of neurospheres enhances oligodendrocyte generation after adhesion
To investigate whether sErbB3 influences cell differentiation after adhesion of neurospheres, we analyzed by immunocytochemistry the different CNS cell types in the outgrowth of control or sErbB3-treated neurospheres (as described above). In one group, the inhibitor was washed out before neurosphere adhesion, whereas it was added for 2 d in the other group (Fig. 6). The total number of cells that had migrated out of each neurosphere was determined for each condition (Fig. 6A), and the respective ratios of oligodendrocytes, neurons, and astrocytes to this total cell number were determined (Fig. 6B). As predicted from the results of the experiments on neurospheres (Fig. 5), neurosphere treatment with sErbB3 before adhesion significantly decreased the total number of cells migrating out of neurospheres, whereas sErbB3 added after adhesion had no such effect (Fig. 6A). Most of the outgrowth cells were still undifferentiated NPs at 2 d. Among the differentiated cells, surprisingly, the ratio of O4-positive preoligodendrocytes to total cells was three times greater after adhesion of sErbB3-treated neurospheres (15 ± 3%) than of control (5 ± 1%) or sGlyr-treated neurospheres (6 ± 2%). This effect was specific for preoligodendrocytes because the ratio of GFAP-positive astrocytes or TujI-positive neuronal cells was not increased. Addition of sErbB3 after adhesion did not increase the percentage of preoligodendrocytes (7 ± 2%) (Fig. 6B). These results indicate that sErbB3 treatment of growing neurospheres can selectively prime the precursors for oligodendrocyte differentiation. The preferential recruitment of these cells from the undifferentiated NPs could also indicate an accelerated process of differentiation in the presence of the Nrg-1 inhibitor.
View larger version (40K):[in this window]
[in a new window]
Figure 6. Effect of soluble ErbB3 and GGF2 on neurosphere differentiation. The neurospheres were grown for 6 d in the presence of 0.5 µg/ml of sErbB3 or sGlyr or in the presence of 100 ng/ml of GGF2 and then adhered for 2 d to allow differentiation in the presence or absence of these soluble proteins. Cultures were either triple labeled with O4 (for preoligodendrocytes), TujI (for neuronal cells), or GFAP (for astrocytes) antibodies (B) or double labeled with O1 and MBP antibodies (for oligodendrocytes; C, D) and counterstained with DAPI (A-D). The ratio of each cell type to the total number of cells in the outgrowth was determined. A, The total number of DAPI-positive cells that migrated out of each adherent sphere decreased when the growing spheres were first cultured in the presence of sErbB3. After adhesion, sErbB3 shows no effect on the number of migrating cells. In contrast, addition of GGF2 increased the total number of cells. B, The ratio of O4-positive cells increased by approximately three times in the presence of sErbB3, whereas the ratio of neuronal cells and astrocytes stayed the same. The increase in the oligodendrocyte ratio was not seen if sErbB3 was added only after neurosphere adhesion (B, C). In the presence of GGF2, the percentage of oligodendrocytes decreased and astrocytes significantly increased, whereas the percentage of neuronal TuJI-positive cells was not significantly changed. C, The percentage of either O1- or MBP-positive cells in the outgrowth had also increased at 2 d after adhesion of neurospheres pretreated with sErbB3. This effect was not observed with neurospheres treated with sErbB3 only after adhesion. Each bar represents the mean ± SEM of three spheres from triplicate experiments done three times. Student t test: *p < 0.05; **p < 0.02; ***p < 0.01. D, After sErbB3 treatment, cells coexpressing MBP (green dots at arrowheads) in processes of galactocerebroside-positive oligodendrocytes (O1 antibody in red) showed an increased complexity of the processes (arrow) compared with oligodendrocytes in control and sGlyr-treated cultures. Scale bar, 10 µm.
The effect of the soluble receptor was compared with that of GGF2 before and after adhesion. GGF2 treatment increased the number of cells in the outgrowth and moderately decreased the percentage of preoligodendrocytes (from 5 ± 1 to 3 ± 0.3%) (Fig. 6A,B). In addition, GGF2 increased the percentage of GFAP-positive cells from 0.5 ± 0.2 to 4 ± 0.5% as described previously (Pinkas-Kramarski et al., 1994
Because soluble ErbB receptor emerged as a strong inducer of oligodendrocyte development, we also studied the effect of sErbB3 pretreatment of growing neurospheres on expression of GC and MBP, two markers of postmitotic oligodendrocytes. Such sErbB3 pretreatment resulted in a significant increase in the percentage of GC-positive cells after adhesion (from 6 ± 2% in control to 12 ± 2%) and a doubling of the percentage of MBP-positive cells (from 1 ± 0.3% in control to 2 ± 0.4%) (Fig. 6C). Interestingly, when sErbB3 treatment was maintained throughout neurosphere growth and differentiation, the percentage of GC-positive cells (Fig. 6C) further increased from 12 ± 2 to 16 ± 3%, and the complexity of oligodendrocyte processes was increased when compared with control (Fig. 6D). As observed with the O4 labeling, addition of sErbB3 exclusively after adhesion did not significantly modify the percentage of GC- and MBP-positive cells versus controls (Fig. 6C). Together these results demonstrate that sErbB3 favored oligodendrocyte development and may accelerate their differentiation.
DISCUSSION
Our study demonstrates that NPs within the embryonic striatum and grown in neurospheres synthesize multiple NRG-1 transcripts and proteins as well as the ErbB2 and ErbB4 receptors. Although both secreted and transmembrane type III NRG-1 isoforms were identified by RT-PCR in growing neurospheres, only one dominant Nrg-1, the transmembrane SMDF/CRD-NRG isoform, was demonstrated by Western analysis. ErbB receptors and Nrg-1 are coexpressed in NPs, and treatment with sErbB3 (which binds Nrg-1 ligands) decreases proliferation and increases apoptosis of these precursors, resulting in a significant reduction in neurosphere size and number. Thus Nrg-1 made by striatal neural precursors promotes their growth via autocrine or paracrine signaling, or both. Furthermore, the ability of GGF2 to enhance only the initial stage of neurosphere growth in low-density cultures suggests that the level of Nrg-1 bioactivity was rapidly saturating in this culture system. After adhesion, GGF2 treatment favored astrocyte rather than oligodendrocyte differentiation (Pinkas-Kramarski et al., 1994
Previous studies have shown that FGF2 and EGF enhance neural stem cell growth in vivo and in vitro (for review, see Brüstle and Dubois-Dalcq, 2001
Our study also provides evidence that type III transmembrane SMDF/CRD is the Nrg-1 isoform specifically involved in promoting NP growth in neurospheres. Although the RT-PCR signal detected the secreted and long transmembrane isoform, Western blot analysis revealed only the transmembrane type III SMDF/CRD in the striatum and in differentiated neurospheres. This does not exclude the possibility that the soluble protein was made but not detected in our experimental conditions. Although type I or II NRG-1 transcripts were detected in E14 striatum (Meyer et al., 1997
The strong and specific effect of sErbB3 treatment on oligodendrocyte generation from NPs is of particular interest. When this Nrg-1 inhibitor was added during sphere growth for 6 d, the relative oligodendrocyte population in the outgrowth had tripled compared with control, whereas the proportion of neuronal and astrocytic cells in the outgrowth remained unchanged. There are several possible interpretations of these findings (Fig. 7). The simplest one is that Nrg-1 acts not only as mitogen for NPs but also inhibits differentiation of early OPs (Canoll et al., 1996
View larger version (31K):[in this window]
[in a new window]
Figure 7. Diagram of the hypothetical mechanisms of action of Nrg-1 signaling in NPs and oligodendrocyte differentiation. NPs, growing in neurospheres, strongly expressed Nrg-1. After adhesion, NPs migrated out of neurospheres and progressively differentiated into astrocytes, neurons, and oligodendrocytes, all weakly expressing Nrg-1. Addition of sErB (in this study, sErbB3) during neurosphere formation increased NP apoptosis, decreased NP proliferation rate, and increased oligodendrocyte differentiation, suggesting an increase in OP selection or commitment, or both. The framed panels show two different interpretations of these data. In a, sErbB blocks endogenous Nrg-1, soluble or transmembrane, decreasing its mitotic and survival effect on NPs; in b, sErbB binds to transmembrane Nrg-1 to trigger an intracellular signal of an unknown nature.
What then is the physiological stimulus of differentiation of oligodendrocytes in vivo? Interestingly, two isoforms of the ErbB4 receptor have been described that differ in the juxtamembrane portion of the extracellular domain JMa [(23 amino acid (aa)], which is cleavable, and JMb (13 aa), which is noncleavable (Elenius et al., 1997
In conclusion, our study demonstrates how Nrg-1 promotes the growth of NPs and underlines the importance of Nrg-1 signaling in oligodendrocyte development. This process could occur in two steps. First, Nrg-1 signaling would enhance and maintain the growth of NPs and early oligodendrocyte precursors, and subsequently the regulated release of a soluble Nrg-1 receptor would inhibit mitosis and induce (or permit) oligodendrocyte differentiation. Future investigations in vivo and in vitro may help determine whether this hypothesis is correct.
FOOTNOTES Received Oct. 13, 2000; revised April 2, 2001; accepted April 17, 2001.
This work was supported by Grant 0480/I 16298 of the Multiple Sclerosis Society of Great Britain and Northern Ireland to M.D.-D. and V.C. B.R. and P.L. are Research Associates of the "Fonds National de la Recherche Scientifique" of Belgium (FNRS), which also supported this work. We also thank the Fondation Médicale Reine Elisabeth, Fonds Charcot, the Belgian League against Multiple Sclerosis, and the Concerted Action of the French Belgian Community. We thank Geneviève Rougon and Roberto Bruzzone for critical reading of this manuscript. Dr. Y. Yarden kindly provided the ErbB3 vector (CDM7-IgB-3).
Correspondence should be addressed to Dr. Monique Dubois-Dalcq, Unité de Neurovirologie et Régénération du Système Nerveux, Institut Pasteur, 25 rue du Dr. Roux, 75724 Paris, Cedex 15, France. E-mail: mdalcq{at}pasteur.fr.
REFERENCES
- Adlkofer K, Lai C (2000) Role of neuregulins in glial cell development. Glia 29:104-111[ISI][Medline].
- Ben-Hur T, Rogister B, Murray K, Rougon G, Dubois-Dalcq M (1998) Growth and fate of PSA-NCAM+ precursors of the postnatal brain. J Neurosci 18:5777-5788
[Abstract/Free Full Text] .- Bermingham-McDonogh O, McCabe KL, Reh TA (1996) Effects of GGF/neuregulins on neuronal survival and neurite outgrowth correlate with erbB2/neu expression in developing rat retina. Development 122:1427-1438[Abstract].
- Bermingham-McDonogh O, Xu YT, Marchionni MA, Scherer SS (1997) Neuregulin expression in PNS neurons: isoforms and regulation by target interactions. Mol Cell Neurosci 10:184-195.
- Betz H, Kuhse J, Schmieden V, Laube B, Kirsch J, Harvey RJ (1999) Structure and functions of inhibitory and excitatory glycine receptors. Ann NY Acad Sci 868:667-676
[Abstract/Free Full Text] .- Brüstle O, Dubois-Dalcq M (2001) CNS stem cells, their gliogenic potential. In: Glial cell development (Jessen K, Richardson WD, eds). New York: Oxford UP, in press.
- Burden S, Yarden Y (1997) Neuregulins and their receptors: a versatile signaling module in organogenesis and oncogenesis. Neuron 18:847-855[ISI][Medline].
- Calver AR, Hall AC, Yu WP, Walsh FS, Heath JK, Betsholtz C, Richardson WD (1998) Oligodendrocyte population dynamics and the role of PDGF in vivo. Neuron 20:869-882[ISI][Medline].
- Canoll PD, Musacchio JM, Hardy R, Reynolds R, Marchionni MA, Salzer JL (1996) GGF/neuregulin is a neuronal signal that promotes the proliferation and survival and inhibits the differentiation of oligodendrocyte progenitors. Neuron 17:229-243[ISI][Medline].
- Canoll PD, Kraemer R, Teng KK, Marchionni MA, Salzer JL (1999) GGF/neuregulin induces a phenotypic reversion of oligodendrocytes. Mol Cell Neurosci 13:79-94[ISI][Medline].
- Carraway KL, Burden SJ (1995) Neuregulins and their receptors. Curr Opin Neurobiol 5:606-612[ISI][Medline].
- Chen X, Levkowitz G, Tzahar E, Karunagaran D, Lavi S, Ben-Baruch N, Leitner O, Ratzkin BJ, Bacus SS, Yarden Y (1996) An immunological approach reveals biological differences between the two NDF/heregulin receptors, ErbB-3 and ErbB-4. J Biol Chem 271:7620-7629
[Abstract/Free Full Text] .- Ciccolini F, Svendsen CN (1998) Fibroblast growth factor 2 (FGF-2) promotes acquisition of epidermal growth factor (EGF) responsiveness in mouse striatal precursor cells: identification of neural precursors responding to both EGF and FGF-2. J Neurosci 18:7869-7880
[Abstract/Free Full Text] .- Corfas G, Rosen KM, Aratake H, Krauss R, Fischbach GD (1995) Differential expression of ARIA isoforms in the rat brain. Neuron 14:103-115[ISI][Medline].
- Eagleson KL, Daigneau L, Levitt P (1998) The role of ErbB receptor signaling in cell fate decisions by cortical progenitors: evidence for a biased, lineage-based responsiveness to different ligands. Mol Cell Neurosci 12:349-362[Medline].
- Edlund T, Jessell TM (1999) Progression from extrinsic to intrinsic signaling in cell fate specification: a view from the nervous system. Cell 96:211-224[ISI][Medline].
- Elenius K, Corfas G, Paul S, Choi CJ, Rio C, Plowman GD, Klagsbrun M (1997) A novel juxtamembrane domain isoform of HER4/ErbB4. Isoform-specific tissue distribution and differential processing in response to phorbol ester. J Biol Chem 272:26761-26768
[Abstract/Free Full Text] .- Fernandez PA, Tang DG, Cheng L, Prochiantz A, Mudge AW, Raff MC (2000) Evidence that axon-derived neuregulin promotes oligodendrocyte survival in the developing rat optic nerve. Neuron 28:81-90[ISI][Medline].
- Flores AI, Mallon BS, Matsui T, Ogawa W, Rosenzweig A, Okamoto T, Macklin WB (2000) Akt-mediated survival of oligodendrocytes induced by neuregulins. J Neurosci 20:7622-7630
[Abstract/Free Full Text] .- Francis A, Raabe TD, Wen D, DeVries GH (1999) Neuregulins and ErbB receptors in cultured neonatal astrocytes. J Neurosci Res 57:487-494[ISI][Medline].
- Gatti R, Belletti S, Orlandini G, Bussolati O, Dall'Asta V, Gazzola GC (1998) Comparison of annexin V and calcein-AM as early vital markers of apoptosis in adherent cells by confocal laser microscopy. J Histochem Cytochem 46:895-900
[Abstract/Free Full Text] .- Gavrieli Y, Sherman Y, Ben-Sasson SA (1992) Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 119:493-501
