ABSTRACT
Permanent Schwann cells lines have been established in culture after stable transfection of newborn rat Schwann cells with the pJC-SVLTtsA vector, expressing a thermosensitive oncogene driven by the early promoter-enhancer region of the gliotropic GS/B variant of the papovavirus JC. The proliferation and differentiation of two clonal cell lines have been studied. The cells of these lines display the morphology of primary Schwann cells and express Schwann cell differentiation markers such as the S-100 protein, laminin, the low-affinity receptor to nerve growth factor and the glial fibrillary acidic protein. One of the lines is able to differentiate further. Indeed, in the presence of dorsal root ganglion neurones, the cells synthesize the myelin Po protein and are capable of some myelination, although to a lesser extent than secondary Schwann cells.
Introduction
During the development of the peripheral nervous system, Schwann cells (SCs) migrate along growing axons, ensheath them and form myelin. Schwann cell terminal differentiation has been shown to depend on axon signals (Salzer and Bunge, 1980; Brunden et al. 1990).
However, most of the molecular mechanisms involved in the control of the proliferation and differentiation of SCs remain to be elucidated. During the study of neural precursor cell lines, which were derived from the central nervous system, we identified NPDC-1, a gene encoding a protein that regulates the proliferation of neural precursor cells (Galiana et al. 1995); we assume that immortalized SC lines, constituting a homogeneous and unlimited cell material, should be useful for studying the mechanisms implicated in terminal differentiation of SCs. Indeed, primary SCs usually consist of a heterogeneous mixture of different developmental stages, and the low division potential of SCs makes it difficult to obtain a large number of cells. In addition, only very few immortalized or transformed SC lines have been obtained, even though a few groups of researchers have reported the generation of SC lines. Some lines have been established after transfection of the viral oncogenes v-Ha-ras (Ridley et al. 1988) and the Simian virus 40 large T antigen (SVLT) gene (Tennekoon et al. 1987), or have been derived from transgenic mice carrying the SVLT gene (Messing et al. 1994). Other authors have reported the establishment of long-term cell cultures by continuous stimulation of cellular growth using glial growth factor and forskolin (Porter et al. 1987; Feltri et al. 1992); Haynes et al. (1994) obtained an immortal cell line by intermittent exposure of SCs to cholera toxin. Spontaneous immortalization of SCs has also been described (Watanabe et al. 1995; Li et al. 1996). Recently, by using the promoter-enhancer region of a gliotropic strain of JC papovavirus, we have targeted expression of the tsA58 allelic form of the SVLT gene, encoding a thermosensitive large T antigen to glial precursor cells (Bernard et al. 1994).
Here, we report that stable introduction of this construct into rat SCs leads to the generation of immortalized SC lines that are conditional for growth. Two such derived clonal cell lines have been examined and shown to display a morphology and to express differentiation markers similar to those of primary SCs. One of these lines is able to differentiate further: when the cells are placed in the presence of dorsal root ganglion (DRG) neurones, the synthesis of the Po protein is induced and some myelination occurs.
Materials and methods
Schwann cell cultures
Primary SC cultures
These cultures were prepared according to the method of Brockes et al. (1979). Sciatic nerves were dissected from newborn rat, chopped and incubated in Dulbecco modified Eagle’s medium (DMEM, Seromed) containing 0.1 % collagenase (Class 3, Worthington) for 25 min at 37 °C. At the end of this incubation, trypsin (0.08 %; Seromed) was added for 10 min, after which the digestion was stopped by the addition of 10 % foetal calf serum (FCS, Gibco). Nerve fragments were dissociated by gentle triturations through a 21 gauge needle and then through a 23 gauge needle. The suspension was filtered through nylon mesh (60 μm mesh size: Do Thi et al. 1993) and centrifuged at 500 g for 10 min. The cells were plated in DMEM-F12 supplemented with 10 % FCS and incubated at 37 °C in a humidified 5 % CO2 incubator.
Secondary SC cultures
In order to eliminate the fibroblasts, primary SCs were removed 1 week later by mild trypsinization (0.25 % trypsin, 1 mmol l−1 EDTA) and treated with anti-Thy 1-1 antibody (dilution 1/4; Cedarlane) and rabbit complement (Cedarlane) before replating in plastic dishes (100 mm diameter; 106 cells per dish).
Transfection of secondary Schwann cells
Four days before transfection, growth of secondary SCs was stimulated by forskolin (5 μmol l−1) and insulin (1 μmol l−1) (Schumacher et al. 1993). These growth factors were maintained in the medium for a total of 16 days. Schwann cells (106 cells) were cotransfected with 25 μg of pJC-SVLTtsA vector (Fig. 1; details of the construct were reported by Bernard et al. 1994) and 5 mg of pSV2TKNeoβ-globin using the calcium phosphate precipitation procedure (Graham and Van der Eb, 1973). The next day, the medium was replaced with fresh DMEM, and the cultures were incubated at 34 °C for two additional days. To select the stably transfected cells, G418 (150 μg ml−1; Gibco) was added, and the selection medium was renewed twice a week. The colonies emerged 4–5 weeks later, when they were individually picked using cloning cylinders and expanded to mass cultures.
Cells from the CR1b4 clone were subsequently cotransfected with 8 μg of pRSVLacZ (pRSVLacZ was a gift from P. Herbomel) and 1 μg of pY3 vector conferring hygromycin resistance (Blochlinger and Diggelmann, 1984) by the calcium phosphate precipitation procedure. The stably transfected cells were then selected in the presence of 150 μg ml−1 hygromycin. The resistant clones emerged 3–4 weeks later, were picked and expanded. A clone expressing LacZ in 20 % of the cells (CR1b4LacZ) was used in transplantation experiments. In order to ensure that the cell population was homogeneous, exponentially growing cells expressing the LacZ gene were sorted through a cell sorter (Nolan et al. 1988). However, within a few days after replating, the percentage of cells expressing the LacZ gene had decreased to 20 %, suggesting that this proportion reflects the differential expression of the gene during the cell cycle or that a negative selection of β-galactosidase-expressing cells could occur.
Transplantation experiments
The left sciatic nerve of an 8-week-old rat was transected at mid-thigh level. Proximal and distal stumps were introduced into silicone rings (1 mm long, inner diameter 1.2 mm, outer diameter 2.5 mm). These rings were inserted into the openings of each end of a silicone tube (10 mm long, inner diameter 2.5 mm, outer diameter 3.5 mm; Knoops et al. 1990; Feltri et al. 1992), and the nerve was anchored to the tube with a single nylon suture (0.3 Ethilon-Ethnor), leaving a gap of 0.8 mm between the stumps. Before this operation, the tubes were filled with 106 cells (CR1b4LacZ) in 20 μl of DMEM.
The nerves were then allowed to regenerate through the tube for 6 weeks. In order to avoid rejection of the grafted cells, Cyclosporin A (SandimmunR, provided by Sandoz Laboratory) was administered every day to the rats subcutaneously at a dose of 20 mg kg−1, starting 1 day before the transplantation.
X-Gal staining of Schwann cells in regenerated nerves
After regeneration, nerves were removed and fixed in phosphate-buffered 2.5 % glutaraldehyde for 2 h and then placed in X-gal solution overnight (Evrard et al. 1988). The nerves were rinsed in phosphate buffer and prepared according to the method of Phend et al. (1995) using an osmium-free method of Epon embedding. Thick sections (1 μm) were counterstained with Cresol Violet to visualize myelinated fibres.
In vitro myelination
Neurone cultures from dorsal root ganglia were prepared according to the method of Eldridge et al. (1987). Dorsal root ganglia were dissected from 16-day-old rat embryos. Ten explanted ganglia were seeded per Petri dishes (35 mm diameter) previously coated with rat tail collagen (Eldridge et al. 1987) and incubated in DMEM/F12 medium (v/v) supplemented with 10 % FCS, 25 ng ml−1 NGF 7S (Sigma) at 37 °C. During this period, the cultures were subjected to three pulses of 10 μmol l−1 cytosine arabinoside (AraC, Sigma), each pulse lasting 2 days.
The medium was then replaced by serum-free defined medium (Bottenstein and Sato, 1979) supplemented with 7S NGF (25 ng ml−1). One day later, immortalized Schwann cells (5×105 cells) were added to the dishes and the cultures were incubated at 34 °C; in these cultures, the immortalized SCs could be identified easily by staining with monoclonal antibody against the SVLT antigen (Harlow et al. 1981). After incubation for 1 week, these co-cultures were fed with defined medium containing 20 % FCS, 25 ng ml−1 7S NGF and 50 ng ml−1 ascorbic acid. The cultures were maintained for a further 2 weeks in the myelinating medium.
Immunofluorescence
Monoclonal antibodies against Thy1-1 and neurofilament 160 (NF) were obtained from Sigma; monoclonal antibody against low-affinity NGF receptor (LNGFR) was a gift from J. de Vellis (UCLA); rabbit polyclonal anti-laminin and anti-Po were given to us by M. Vigny (Université Paris-7) and J. Bonnet (Université Bordeaux-2), respectively. Rabbit anti-S-100 was from Sigma, and rabbit anti-GFAP from Dakopatts. They were used at a dilution of 1:100.
Cells were seeded on glass coverslips and were processed for indirect immunofluorescence. Cells to be stained for LNGF-R and laminin were incubated for 45 min at room temperature with the appropriate antibodies diluted in phosphate-buffered saline plus 5 % foetal calf serum (PBS-FCS). They were then washed three times in PBS-FCS, incubated with the appropriate secondary fluochrome-conjugated antibody (1:100) for 45 min and rinsed five times in PBS-FCS. The cells were fixed in 4 % paraformaldehyde for 10 min, rinsed in PBS and mounted in Dabco-Gelvatol medium (Langanger et al. 1983). For Thy1-1 detection, the cells were lightly fixed in 2 % paraformaldehyde at 4 °C for 10 min before staining with the antibodies. Cells to be stained for SVLT, S-100, NF and GFAP were fixed in 4 % paraformaldehyde for 10 min, permeabilized with methanol (80 %) at −20 °C for 10 min and processed as above.
Staining of myelin
For Sudan Black staining, the cells were fixed for 10 min in 4 % paraformaldehyde in 0.1 mol l−1 sodium phosphate buffer, pH 7.4, post-fixed for 1 h in 1 % OsO4 in the same buffer at 4 °C, stained for 4 h with 0.5 % Sudan Black (Sigma) in 70 % ethanol and rinsed with 70 % ethanol before being mounted in glycerol gelatin (Sigma).
Myelin quantitation
The areas of maximal myelination were chosen systematically. The total length of the myelinated segments was measured in ten independent microscopic fields (area 1 mm2, Biocom). For each experimental condition, three dishes were analysed. The quantity of myelin was expressed as the total length of myelinated segments per mm2.
Electron microscopy
The cells were fixed overnight at 4 °C in 2.5 % glutaraldehyde diluted in 0.05 mol l–1 sodium phosphate buffer (pH 7.0) containing 0.1 mol l−1 sucrose. After several washes in this buffer, the cells were placed for 1 h in 2 % osmium tetroxide (Sigma) at 4 °C, rinsed, dehydrated and embedded in Epon. Ultrathin sections, obtained using an LKB ultramicrotome, were stained with uranyl acetate and lead citrate (Tougard and Picart, 1986) and examined in a Philips 300 electron microscope.
Southern hybridizations
DNA was isolated from SC lines as described by Maniatis et al. (1989). DNA (10 μg) was digested with restriction enzymes, subjected to electrophoresis on a 0.7 % agarose gel, transferred onto a 0.45 μm nitrocellulose membrane (Schleicher and Schuell) and hybridized with a 32P-labelled Bam HI–Xho I SVLT fragment.
Detection of the Po protein by immunoblotting
Expression of Po protein in Schwann cell lines was assayed by western blot analysis. Cells from dorsal root ganglion explants were rinsed in PBS and lysed with 50 mmol l−1 Tris–HCl (pH 8), 150 mmol l−1 NaCl, 1 % non-ionic detergent P40, 0.5 % deoxycholate, 0.1 % sodium dodecyl sulphate for 20 min at 4 °C. After centrifugation for 10 min at 10 000 g, the supernatant was collected and stored at −80 °C until used. The protein concentration was determined using the Bradford (1978) procedure. Protein (40 μg) was electrophoresed on a 10 % SDS–polyacrylamide gel (Laemmli, 1970). The fractionated proteins were transferred onto a nitrocellulose membrane (45 μm pore size) and probed with the same Po antibody as above (dilution 1:5000) and peroxidase-conjugated secondary antibody (Amersham, 1:5000). The antigen–antibody complexes were visualized by ECL detection (Amersham).
Results
Immortalized Schwann cell clones
Secondary SCs were transfected with pJC-SVLTtsA plasmid (Bernard et al. 1994), which carries the promoter region of the gliotropic GS/B strain of JC papovavirus (Loeber and Dörries, 1988) upstream from the tsA large T sequence of SV40. For convenient selection of the transfected cells, the NeoR gene, inserted in pSV2TKNeo (Nicolas and Berg, 1983), was co-introduced, conferring resistance to G418 to the cells (Colbere-Garapin et al. 1981). Individual colonies were isolated after 4–5 weeks, then expanded in Petri dishes (35, 60 and 100 mm diameter). Two clones, named CR3a1 (Fig. 2A) and CR1b4 (Fig. 2B), out of 25 isolates have been subcloned and characterized as described below.
The cells of both clones were bipolar and spindle-shaped with a morphology similar to that of primary cultured SCs. The CR1b4 cells formed aligned arrangements, whereas the CR3a1 displayed a slightly flattened morphology (Fig. 2A,B).
Growth properties
The growth of these clones was temperature-dependent. At 34 °C, the permissive temperature, the doubling time of CR3a1 and CR1b4 cells was 48 h and 76 h, respectively (Fig. 3). When shifted to 39 °C, CR3a1 cells divided very slowly and displayed a generation time of 5 days (Fig. 3). However, these cells could resume growth when shifted back to 34 °C, even after 5 days at the non-permissive temperature. In contrast, CR1b4 cells stopped dividing completely at 39 °C (Fig. 3). This growth-arrest was still reversible after 24 h at 39 °C, but not after incubation at 39 °C for 1 week.
Integration and expression of the large T antigen
In order to estimate the number of copies of the immortalizing gene integrated into the genome, the DNA was isolated and digested with BamHI and XhoI restriction endonucleases. The restriction fragments were analyzed by Southern blot hybridization with a 32P-labelled SVLT probe (Fig. 4). The results showed that 2–5 and 1–2 copies of the SVLT gene were integrated into the genome of the CR3a1 and CR1b4 cells, respectively.
At the permissive temperature, Schwann cell lines expressed the large T antigen predominantly in the nucleus, as shown by indirect immunofluorescence (Fig. 5B). Upon being shifted to the non-permissive temperature, the level of fluorescence diminished and became nearly undetectable after 48 h of incubation at 39 °C (data not shown).
Expression of Schwann cell markers
Indirect immunofluorescence experiments showed that, at 34 °C, CR3a1, CR1b4 and secondary SCs expressed SC markers such as S-100 (Jessen et al. 1989), GFAP (Jessen et al. 1990) and laminin (McGarvey et al. 1984). CR1b4 cells and secondary SCs were recognized by anti-LNGFR (low-affinity nerve growth factor receptor) antibody, whereas CR3a1 cells were not (Fig. 5; see also Table 1). As might be expected, in the absence of neurones, the Po protein was not detected in the two clones. In contrast, 2 weeks after plating on DRG neurones, CR3a1 and CR1b4 cells expressed type IV collagen (Fig. 5E), whereas only CR1b4 cells expressed Po protein, which was revealed both by immunofluorescence (Fig. 6A) and by western blot analysis (Fig. 6B). Addition of forskolin did not increase Po expression, probably because of the presence of DRG neurones (Lemke and Chao, 1988). At the non-permissive temperature (39 °C), both immortalized cell lines were also able to express several Schwann cell markers such as S-100, GFAP and laminin, whereas only CR1b4 cells expressed LNGFR. At this temperature, neurones did not survive well enough after 2 weeks to allow the absence of Po expression to be interpreted unambiguously.
Myelination
In vitro myelination
To determine whether SC lines could differentiate at 34 °C and form myelin in the presence of axons, the cell lines were co-cultured with DRG neurones. These co-cultures were analyzed by double-immunofluorescence experiments: neurones were visualized with an anti-NF antibody, whereas SCs were revealed by an antibody against SVLT antigen, in order to ensure that the cells interacting with neurones were immortalized SCs. No myelin sheath was detected in co-cultured DRG neurones and CR3a1 cells, even though the nuclei were seen aligned along neurites (Fig. 7A,B).
In contrast, a few short myelinated segments were observed in co-cultures of CR1b4 cells and DRG neurones (Fig. 8B). However, the length of the myelinated segments was 0.5 % (75 μm mm−2) of lengths detected in control cultures (14 900 μm mm−2, Fig. 8A). As shown in Fig. 9, in parallel cultures treated for electron microscopy, non-myelinated axons appeared to be surrounded by Schwann cell processes. Some myelinated axons were also observed (Fig. 9A), and basal lamina was seen around SCs (Fig. 9B). Myelin ovoids were present in the SC cytoplasm (Fig. 9B). In the axoplasm of a few axons, lamellar bodies and electron-dense bodies were also observed (Fig. 9C). Such features were never seen in control myelinating cultures.
In vivo myelination
To determine whether the CR1b4LacZ cells could myelinate regenerating axons, we transected the left sciatic nerve and injected the cells in a silicone tube which bridged the ends of the nerve. After incubation for 6 weeks, regenerated nerves from the silicone tube were removed, fixed and stained as described in Materials and methods. A small proportion of the grafted cells showed LacZ staining in the middle part of the regenerated nerve segment. Some of these β-galactosidase-expressing cells were associated with myelinated axons, which were visualized with Cresol Violet to reveal the myelin sheath surrounding the axons (Fig. 10).
In control experiments (transected sciatic nerves were allowed to regenerate in empty silicone tubes), the number of myelinated fibres was ten times higher and the diameter of regenerated nerves was greater (15 μm compared with 6 μm)
Discussion
This study shows that transfection of rat SCs with the pJC-SVLTtsA vector (Bernard et al. 1994) is an efficient method of deriving clonal SC lines that are conditional for growth. This vector carries the enhancer-promoter region of the GS/B gliotropic strain of JC virus fused to the gene for the large T antigen of the tsA58 temperature-sensitive mutant of SV40. This enhancer-promoter region is responsible for the gliotropism of this viral strain (Loeber and Dörries, 1988), whereas the role of the tsA58 thermosensitive allelic form of SVLT gene is to switch off the oncogene function at the non-permissive temperature.
It has been shown that neonatal SCs have a doubling time of 7 days (Raff et al. 1978a,b; Brockes et al. 1979). However, in the presence of cholera toxin, a Schwann cell mitogen, the doubling time of secondary SCs is reduced to 48 h (Raff et al. 1978a,b; Brockes et al. 1979). The immortalized SC lines described here did not need the addition of mitogens other than those present in the serum. Indeed, the doubling times of CR3a1 and CR1b4 cells, at the permissive temperature, were 48 h and 76 h, respectively. The reduction in generation time of these cells compared with that of secondary SCs is probably due to the integration of the tsa form of the large T gene and to the continued expression of the large T antigen at 34 °C, which maintains the cells in a proliferative state (Jat and Sharp, 1989).
CR1b4 cells stopped dividing within 48 h after transfer to 39 °C, and incubation of these cells for 1 week at the non-permissive temperature suppressed irreversibly their ability to resume growth at 34 °C. In comparison, CR3a1 cells, which divided very slowly at 39 °C (Fig. 3), were able to resume growth at 34 °C after 5 days at the non-permissive temperature. This result shows that the proliferation potential of SCs immortalized with temperature-sensitive mutants of large T antigen gene is highly dependent on temperature.
The immortalized SCs expressed several antigenic SC markers, at 34 °C as well as at 39 °C, as shown by immunofluorescence (Table 1). They were positive for S-100, GFAP, laminin and type IV collagen; CR1b4 cells are also positive for LNGFR (Table 1).
As might be expected, when immortalized SCs were cultivated alone, they did not express Po protein, as described for secondary SCs in vitro (Brockes et al. 1980; Mirsky et al. 1980). In the presence of DRG neurones, CR1b4 cells expressed Po protein (Fig. 6B). In contrast, the CR3a1 cells did not express this myelin protein. An explanation could be that the large T antigen, which is strongly expressed in this line, might alter the expression of myelin genes, as has been suggested by Tennekoon et al. (1987) and Bharucha et al. (1994). For the CR1b4 cell line, the low level of large T antigen might allow the expression of Po protein when the cells were co-cultured with DRG neurones.
Even though the CR1b4 cells are able to express Po protein, we obtained only a little myelination with these cells compared with secondary SCs. This may also be partly due to the expression of the large T antigen, even at low levels. However, this does not explain why only some cells respond to axonal signals for myelination while other cells do not. A possible explanation, which remains to be demonstrated, might be an incompatibility, at least partial, between the response of the immortalized SCs cells to axonal signals for myelination and the expression of the large T antigen associated with a proliferative state of the cells.
In co-cultures of DRG neurones and CR1b4 cells, the expression of the SVLT antigen in the nucleus of the Schwann cells, which form myelin, is hardly detectable. Formally, we cannot completely exclude the possibility that, even after treatment with AraC, some endogenous SCs could migrate from the DRG neurone mass and myelinate axons. However, this hypothesis does not fit with the presence of demyelination figures and the accumulation of dense material in the axoplasms in such cultures. Indeed, we have never observed such figures in myelinated control cultures. Such abnormal figures have been reported in sciatic nerves of the dysmyelinating mutant mouse Trembler (Ayers and Anderson, 1976; Low, 1977), in canine neuroaxonal dystrophy of Rottweiler dogs (Cork et al. 1983) and in rat demyelinated peripheral nerve axons after intoxication by Taxol (Vuorinen et al. 1989) or by diphtheria toxin (Hildebrand, 1989). In addition, it has been reported that transgenic mice expressing the SV40 large T antigen develop a peripheral demyelinating neuropathy (Messing et al. 1985). Other transgenic mice carrying the early region (including the large T gene) of the papovavirus JC showed a dysmyelination in the central nervous system (Small et al. 1986). These observations are in agreement with the above-proposed hypothesis which assumes an incompatibility between the response of SCs to axonal signals for myelination and the expression of the large T antigen.
Experiments in which the cells were allowed to interact with transected sciatic nerve were carried out in order to determine whether CR1b4LacZ cells could participate in the regeneration of the nerve. Although only a fraction of the cell population (20 %) expressed β-galactosidase at the time of transplantation, our observations suggest that some grafted cells could make myelin when associated with regenerated axons. Indeed, immortalized cells visualized by lacZ staining were found in close association with the regenerating axons.
Surprisingly, the diameter of the regenerated nerves in the presence of CR1b4 cells was smaller than in control experiments without a contribution from exogenous cells. Moreover, the number of myelinated axons was significantly decreased. These observations suggest that the grafted cells might release factors which partially inhibit regeneration by preventing the migration of endogenous SCs from the proximal part of the transected sciatic nerves. Once again, this could be due to expression of the large T antigen, although we cannot exclude the possibility that the number of grafted cells could have been too low for efficient regeneration. Nevertheless, the present SC lines should be considered with care for nerve regeneration experiments.
However, these lines should be available for studying the molecular mechanisms involved in the terminal differentiation of SCs and in the expression of Po protein. More particularly, CR1b4 cells display a clear-cut expression of this protein, whereas CR3a1 cells do not. This differential expression might facilitate the search for regulatory genes that are differentially expressed during terminal differentiation of SCs and that could be involved in the onset of Po expression. Indeed, by studying gene expression in other immortalized neural precursor cells, we have identified a novel gene, NPDC-1, which encodes a protein involved in the control of proliferation and differentiation of neural cells (Galiana et al. 1995).
Acknowledgement
We thank Professors François Gros and Herbert Koenig for helpful discussion and Drs Annie Ressouches and Renée Picart for their help in electron microscopy. This work was supported by a grant from Association Française contre les Myopathies (AFM) and from the Association pour la Recherche contre le Cancer (ARC).