Dopaminergic differentiation of the Nurr1-expressing immortalized mesencephalic cell line CSM14.1 in vitro

Authors


Professor Andreas Wree, Institute of Anatomy, University of Rostock, Gertrudenstr. 9, PO Box 100888, D-18055 Rostock, Germany. Tel. +49 381494 8401; fax: +49 381494 8402, e-mail: andreas.wree@med.uni-rostock.de

Abstract

The use of neural stem cells as grafts is a potential treatment for Parkinson's disease, but the potential of stem cells to differentiate into dopaminergic neurones requires investigation. The present study examined the in vitro differentiation of the temperature-sensitive immortalized mesencephalic progenitor cell line CSM14.1 under defined conditions. Cells were derived from the mesencephalic region of a 14-day-old rat embryo, retrovirally immortalized with the Large T antigen and cultured at 33 °C in DMEM containing 10% fetal calf serum (FCS). For differentiation, the temperature was elevated at 39 °C and FCS was reduced (1%). Using histology, immunocytochemical detection of the stem cell marker Nestin and the neuronal marker MAP5 and, in addition, Western blotting to determine the presence of neurone-specific enolase and the neurone nuclei antigen we demonstrated a differentiation of these cells into neuronal cells accompanied by a decrease in Nestin production. In Western blots, we detected the orphan nuclear receptor Nurr1 in these cells. This was followed by a time-dependent up-regulation of the enzymes tyrosine hydroxylase and aldehyde dehydrogenase 2 characteristic of mature dopaminergic neurones. Our in vitro model of dopaminergic cell differentiation corroborates recent in vivo observations in the developing rodent brain.

Introduction

The intracerebral transplantation of conditionally immortalized progenitor cells could be a useful treatment in neurodegenerative disorders, such as Parkinson's disease (Martinez-Serrano & Björklund, 1997; Björklund & Lindvall, 2000). In the past, patients suffering from Parkinson's disease received autografts containing adrenal cells (Backlund et al. 1985; Olson et al. 1991; Dunnett & Björklund, 1999), fetal mesencephalic xenografts consisting of pig cells (Deacon et al. 1997; Dunnett & Björklund, 1999) or human mesencephalic progenitor cells derived from aborted fetuses (Herman & Abrous, 1994; Olanow et al. 1996; Dunnett & Björklund, 1999; Piccini et al. 1999; Freed et al. 2001).

The common idea of all these clinical trials was to reconstitute the dopaminergic innervation of the dopaminergically deafferentiated corpus striatum in these patients with cell grafts, as has been successful in animal models of Parkinson's disease (Björklund, 1992; Brundin & Wictorin, 1993; Herman & Abrous, 1994). However, the use of either adrenal medulla grafts (Dunnett & Björklund, 1999) and xenografts did not lead to a therapeutic breakthrough, since only single patients received such grafts in the past and possible virus contaminations are discussed (Deacon et al. 1997; Dunnett & Björklund, 1999). The use of human fetal grafts is not a routine therapy (Piccini et al. 1999), and raises ethical problems (Martinez-Serrano & Björklund, 1997; Björklund & Lindvall, 2000).

These technical and ethical problems could be diminished if conditionally immortalized progenitor cells could be generated from human fetal brain (Martinez-Serrano & Björklund, 1997; Björklund & Lindvall, 2000). In the past, several conditionally immortalized cell lines of neural progenitors or neural stem cells were established from brain tissue of rodent embryos (Frederiksen et al. 1988; Redies et al. 1991; Zhong et al. 1993; Snyder, 1994; Cattaneo & Conti, 1998). These cells can be expanded easily and screened in vitro for viral infections. They differentiate in vitro as well as in vivo into functionally integrated neurones and/or glial cells after transplantation into the central nervous system (Frederiksen et al. 1988; Redies et al. 1991; Zhong et al. 1993; Snyder, 1994; Martinez-Serrano & Björklund, 1997; Cattaneo & Conti, 1998; Auerbach et al. 2000). Taken together, they could serve as an ideal model in subsequent studies and clinical trials using human immortalized neural progenitors. Furthermore, immortalized rodent progenitor cells produced therapeutic benefits after transplantation in animal models of neurological disorders, i.e. Parkinson's disease (Anton et al. 1995), multiple sclerosis (Yandava et al. 1999) or stroke (Veizovic et al. 2001).

One focus in neural stem cell research is on dopaminergic differentiation because stem cell grafts could prove useful in the treatment of Parkinson's disease (Anton et al. 1995; Björklund & Lindvall, 2000). Nigrostriatal dopaminergic neurones are generated early during ontogenesis of rodents around embryonic day 8.5 under the influences of various secreted factors such as Sonic hedgehog and basic fibroblast growth factor 8 (Perrone-Capano & Di Porzio, 2000). Later during embryogenesis Nurr1, a member of the superfamily of orphan nuclear retinoic acid receptors, is essential for the development of ventral midbrain neuroblasts into dopaminergic neurones as has been shown previously by in vitro and in vivo studies (Zetterström et al. 1997; Wagner et al. 1999; Perrone-Capano & Di Porzio, 2000). Although Nurr1 lacks an identified ligand (Zetterström et al. 1997; Perrone-Capano & Di Porzio, 2000), exogenously applied retinoids are able to bind to the receptor (Zetterström et al. 1997; Wallen et al. 1999). In vitro experiments revealed that, in co-culture, mesencephalic astrocytes secreted a soluble factor able to bind to Nurr1 and to induce dopaminergic differentiation (Wagner et al. 1999). In the murine mesencephalon, the Nurr1-mRNA is detectable from embryonic day 10.5 (E10.5) on throughout adulthood (Bäckman et al. 1999). The transcription of the gene for Nurr1 is followed at E11.5, in the same cell population by the production of tyrosine hydroxylase (TH), the pacemaker-enzyme of the dopamine synthesis (Zetterström et al. 1997; Saucedo-Cardenas et al. 1998; Perrone-Capano & Di Porzio, 2000). Finally, the synthesis of a retinoic-aldehyde converting enzyme (ALDH2) in most TH-positive cells indicates a mature phenotype of nigrostriatal dopaminergic neurones (McCaffery & Dräger, 1994; Haque et al. 1997; Wagner et al. 1999; Wallen et al. 1999; Perrone-Capano & Di Porzio, 2000). From these in vivo observations, a sequential expression of the genes for Nurr1, TH and ALDH2 during development of dopaminergic neurones seems likely.

In the present study, we characterized the in vitro differentiation of the temperature-sensitive immortalized neuronal progenitor cell line CSM14.1 (Zhong et al. 1993) thought to be rather immature dopaminergic neuronal progenitors. This clonal cell line was derived from the ventral mesencephalic region of an E14 rat. At a well-defined temperature (39 °C, non-permissive temperature) and serum-deprived in vitro conditions we demonstrate decrease of Nestin, differentiation into a neuronal morphology, up-regulation of neuronal proteins, and the detection of Nurr1 and ALDH2 in these cells under conditions for differentiation. The time courses of the appearance of the respective markers established in vitro during differentiation largely corroborate the sequence of molecular events found in vivo.

Materials and methods

Cell culture and cell differentiation

Conditionally immortalized CSM14.1 cells (Zhong et al. 1993) were cultured and expanded in DMEM supplemented with 10% fetal calf serum (FCS), 100 U mL−1 penicillin, 100 µg mL−1 streptomycin in a humidified incubator (95% air/5% CO2, at 33 °C). The cells were passaged every third day. For in vitro differentiation, CSM14.1 cells were treated as described previously for primary neural stem cells (Winkler et al. 1998). In brief, CSM14.1 cells were passaged and cultured overnight in a humidified incubator (95% air/5% CO2, at 33 °C) in Petri dishes for Western blotting or on poly L-lysine-coated culture slides (Becton-Dickinson, Heidelberg, Germany) for immunocytochemistry. Thereafter, the medium was changed for DMEM supplemented with 1% FCS, 100 U mL−1 penicillin, 100 µg mL−1 streptomycin and the incubation continued at 39 °C (non-permissive temperature; 5% CO2, humidified environment). Cells grew for 3, 7, 10 and 14 days, respectively; the medium was changed every third day. All reagents for tissue culture were obtained from Gibco Life Technologies (Karlsruhe, Germany).

Immunocytochemical and Western blot analysis of the cultured cells were repeated at least three times in independent experiments performed under respective culture conditions with reproducible results.

Immunocytochemistry and histology

Cells grown on culture slides for 2 days at 33 °C in DMEM containing 10% FCS or for 3, 7, 10 and 14 days under the non-permissive conditions (see above) were washed with 0.9% sodium chloride and fixed with 4% formaldehyde (prepared freshly from paraformaldehyde dissolved in phosphate-buffered saline; PBS, pH 7.4) for 60 min at room temperature. Cultured cells were then rinsed three times for 5 min in 0.1 m Tris buffer (pH 7.4) and endogenous peroxidases were blocked by 20 min incubation in 3% H2O2 (dissolved in 0.1 m Tris buffer, pH 7.4). After three further washes in 0.1 m Tris buffer (pH 7.4) the specimens were pre-incubated for 1 h in Tris buffer containing 0.05% Triton-X100 (Sigma, Deisenhofen, Germany), 3% BSA (Sigma) and 1.25% normal horse serum (Dianova, Hamburg, Germany) to quench non-specific binding sites, and subsequently incubated with primary antibodies against microtubule-associated protein 5 (MAP5, mouse monoclonal, 1 : 500; Sigma), neural stem cell protein (Nestin, mouse monoclonal, 1 : 500, Becton-Dickinson), glial fibrillary acidic protein (GFAP, mouse monoclonal, 1 : 400, Sigma), neurone-specific enolase (NSE, rabbit polyclonal, 1 : 500, Chemicon, Hofheim, Germany) and tyrosine hydroxylase (TH, mouse monoclonal, 1 : 500, Sigma) overnight at 4 °C. The cell preparations were then washed three times for 10 min in 0.1 m Tris (pH 7.4), incubated again at 4 °C overnight with biotinylated secondary antibodies (polyclonal, 1 : 80, Dianova), washed three times in 0.1 m Tris buffer (pH 7.4), then incubated for 45 min in ABC-Complex (1 : 80, Dianova) at room temperature. After three further washes in 0.1 m Tris buffer (pH 7.4) the labelled binding sites were visualized with a 3,3′-Diaminobenzidine tetrahydrochloride substrate solution (Sigma). Finally, the cells were washed three times for 5 min in 0.1 m Tris buffer (pH 7.4), dehydrated in graded ethanol concentrations and mounted in DePeX (Serva, Heidelberg, Germany). After fixation, parallel culture slides were also stained for 3 min at room temperature with 0.1% Cresyl Violet acetate (Sigma), washed in water and also mounted in DePeX. Specimens were observed and documented with a Leitz Aristoplan microscope (Wetzlar, Germany).

Preparation of cell cultures for Western blotting

Cells grown on Petri dishes were rinsed with 0.1 m PBS (pH 7.4), trypsinized (Gibco Life Technologies), centrifuged (10 min at 400 g), washed and neutralized with DMEM containing 10% FCS. After a final centrifugation (10 min at 400 g), cells were resuspended in PBS. The viability and the number of cells were determined by trypan blue exclusion (Sigma). The cells were then lysed by several freeze/thaw cycles. Equal total cell protein concentrations were determined using a spectrophotometer (Model DU640, Beckman, Fullerton, CA, USA) and a bicinchoninic acid assay (Pierce Chemical Co, Rockford, IL, USA) according to the manufacturer's instructions. Comparable concentrations of whole protein lysates were boiled for 5 min in SDS (sodium dodecyl sulphate) sample buffer (Laemmli, 1970). Fifty micrograms of whole protein dissolved in 20 µL SDS sample buffer was loaded in each lane.

Western blotting

SDS polyacrylamide gel electrophoresis (SDS-PAGE) was performed with ready-to-use criterion mini gels (Biorad, München, Germany) consisting of 4–15% polyacrylamide. After electrophoresis, proteins were transferred onto nitrocellulose membranes (Amersham Pharmacia Biotech, Freiburg, Germany). Membranes were blocked for 2 h at room temperature in 0.1 m PBS (pH 7.4), 0.1% Tween 20 (PBS-T) and 1% BSA (Sigma). Primary antibodies directed against β-actin, the product of a housekeeping gene (1 : 3000, mouse monoclonal, Sigma), the immortalizing gene product SV40 Large T (SV40LT-ag, 1 : 500, mouse monoclonal, Becton-Dickinson), glial fibrillary acidic protein (GFAP, mouse monoclonal, 1 : 400, Sigma), the neurone-specific enolase (NSE, 1 : 1000, rabbit polyclonal, Chemicon), the neuronal nuclei antigen (NeuN, 1 : 5000, mouse monoclonal, Chemicon), the tyrosine hydroxylase (TH, 1 : 3000, mouse monoclonal, Sigma), the orphan nuclear receptor Nurr1 (Nurr1, 1 : 1000, mouse monoclonal, Becton-Dickinson) and aldehyde dehydrogenase 2 (ALDH2, 1 : 1500, rabbit polyclonal, Kathmann et al. 2000) were performed overnight at 4 °C. Membranes were washed in PBS-T (4 × 15 min) and incubated for 1 h at room temperature with secondary antibodies conjugated with horseradish peroxidase (anti-mouse, 1 : 5000 or anti-rabbit, 1 : 10 000, both Vector Laboratories, Burlingame, CA, USA). After washing the membranes in PBS-T (4 × 15 min), the peroxidase activity was visualized with an enhanced chemiluminescence kit (Amersham Pharmacia Biotech).

Results

Morphological changes during neuronal differentiation

Histology of cultured CSM14.1 cells stained with cresyl violet revealed morphological changes in differentiating CSM14.1 cells (Fig. 1). CSM14.1 cells cultured at a permissive temperature of 33 °C (Fig. 1A) had small cell bodies and a fibroblast-like phenotype. Their processes were short, and the cells formed an epithelial-like monolayer after reaching confluence. Cell divisions occurred every 18 h. Under differentiation conditions the CSM14.1 cells began to change their morphology (Fig. 1B,C). After one week at 39 °C and supplemented with 1% FCS only, cells with multipolar somata similar to those of cultured neurones could be found (Fig. 1B). After 14 days culture under differentiation conditions cell bodies and processes of these differentiating cells were considerably larger and began to form a connective network (Fig. 1C).

Figure 1.

Images of immortalized CSM14.1 cells, stained with Cresyl Violet (A–C) or immunostained for Nestin (D–F) or MAP5 (G–I), during differentiation caused by temperature elevation and serum deprivation. (A,D,G) Cells cultured for 2 days at 33 °C in DMEM containing 10% FCS. Undifferentiated cells were characterized by small cell bodies and short processes (A). Most were strongly stained for Nestin (D) but only weakly for MAP5 (G). (B,E,H) Cells cultured for 1 week at 39 °C in DMEM containing 1% FCS showed an enlargement of cell bodies and an elongation of processes developing a multipolar morphology (B). Cells stained weakly for Nestin (E), but showed increasing MAP5 immunostaining (G). (C,F,I) Cells cultured for 14 days at 39 °C had large cell somata, long processes and, furthermore, a connecting network appeared (C). Nestin was down-regulated (F), but the neuronal marker MAP5 was highly expressed by the CSM14.1 cells (I). Scale bar = 75 µm.

Undifferentiated CSM14.1 cells were highly immunoreactive for the neural stem cell marker Nestin (Fig. 1D), whereas in cells cultured for 7 and 14 days at 39 °C the amount of Nestin drastically decreased (Fig. 1E,F). In contrast, the neuronal protein MAP5, which was weakly detectable in cells cultured at 33 °C (Fig. 1G), increased during differentiation in the cell processes and somata of CSM14.1 cells (Fig. 1H,I).

Cell counts on cell suspensions and on the histological stainings (Fig. 1A–C) revealed a reduction of viable cells during differentiation. This could be explained by the down-regulation of the immortalizing temperature-sensitive Large T antigen whose gene is only expressed at the permissive temperature of 33 °C (Fig. 2A,B). When CSM14.1 cells were cultured at 39 °C, the immortalizing Large T antigen was not detectable after 3 days (7 d, 10 d and 14 d) (Fig. 2B).

Figure 2.

Western blots demonstrating various proteins in lysates of CSM14.1 cells cultured for 3 days at 33 °C (DMEM/10% FCS) and for 3 days, 7 days, 10 days or 14 days at 39 °C (DMEM/1% FCS). Fifty micrograms of whole protein lysates were loaded per lane. (A) Similar amounts of the housekeeping protein β-actin (∼42 kDa) were found in all lanes. (B) The immortalizing gene product Large T antigen (∼85 kDa) was only detectable in cells cultured at 33 °C. (C) NSE (∼39 kDa) showed a time- and differentiation-dependent increase, the highest protein content being detectable after 7 days at 39 °C. (D) For NeuN (∼46 and ∼48 kDa) the specific band of 46 kDa appeared only under differentiation conditions.

Using Western blotting to compare equal amounts of whole cell protein lysates (Fig. 2A) and further investigation of possible neuronal differentiation, we observed an increase of several neuronal markers in the differentiating CSM14.1 cells over time. An increased amount of the neuronal marker NSE protein was found in these cells (Fig. 2C). It was low at 33 °C, but increased when cultured at 39 °C, and reached a maximum at 7 d. NeuN, another marker for post-mitotic mature neurones, was also found to be increased in CSM14.1 cells over time (Fig. 2D). In Western blots, antibodies against this protein labelled two bands of ∼46 kDa and ∼48 kDa. No obvious changes in the NeuN content could be observed in the fraction of ∼48 kDa over time, whereas the band of ∼46 kDa was not detectable in cells cultured at 33 °C. Cells cultured at 39 °C for 3–14 days showed clearly detectable bands of NeuN in the 46-kDa range with a peak at 7 d similar to the temporal pattern of the synthesis of NSE.

In contrast, the glial marker GFAP could never be detected by immunocytochemistry or Western blotting in the CSM14.1 cells at any time in culture (not shown).

Dopaminergic differentiation

Undifferentiated CSM14.1 cells (i.e. at 33 °C) synthesized Nurr1 (Fig. 3). In contrast, when cultured at 39 °C, the amount of this putative receptor increased after 3 days and then decreased until, after 14 days in culture, it reached levels even below those observed in undifferentiated cells.

Figure 3.

Western blot showing the protein fraction containing Nurr1 (∼72 kDa) in lysates of CSM14.1 cells during differentiation and in tissue homogenates derived from two positive controls, i.e. adult rat substantia nigra (rt-SN) and lysed SW13-cells (PC). Nurr1 was produced in CSM14.1 cells at 33 °C and to a greater extent after 3 days of cultivation at 39 °C. Nurr1 decreased in long-term cultures, but remained detectable.

In Western blots, TH immunoreactivity was weak for protein extracts of CSM14.1 cells cultured at 33 °C, but strong when cells were exposed to differentiation conditions (Fig. 4). After 3 days at 39 °C, TH content in these cells was markedly increased and thereafter remained constant until 14 d. However, we were not able to demonstrate these changes by immunocytochemical staining of plated cells (not shown).

Figure 4.

Results of a Western blot showing the increase of TH protein (∼60 to ∼68 kDa) in CSM14.1 cell lysates. When cultured at 33 °C, TH immunoreactivity appeared weakly in the fraction of ∼68 kDa, but after 3 days at 39 °C, it was strong (two bands ranging from 60 to 68 kDa) and remained high.

In preliminary Western blotting studies, the expression of the enzyme ALDH2, a putative marker for mature dopaminergic neurones, in the homogenates of rat hippocampus (containing no dopaminergic perikarya) revealed only a very weak ALDH2 immunoreactivity, whereas it was intense in tissue homogenates from adult rat substantia nigra (containing abundant dopaminergic neurones) and in adult rat liver (according to Kathmann et al. 2000) (Fig. 5A). In further experiments with tissue homogenates from adult rat liver tissue, which served as a positive control, and cell lysates from undifferentiated and differentiating CSM14.1 cells (Fig. 5B), ALDH2 was not detectable in CSM14.1 cells when cultured at 33 °C, but when cultured at 39 °C the CSM14.1 cells showed a time-dependent increase of this enzyme, with the highest quantity at 14 d (Fig. 5B). Unfortunately, cellular diversity or the percentage of Nurr1-ir and ALDH2-ir neurones could not be evaluated because the respective antibodies were not suitable for immunocytochemistry.

Figure 5.

(A) Western blot demonstrating the protein fraction of ALDH2 (∼54 kDa) in protein lysates derived from different tissues. The adult rat hippocampus (rt-hipp), which contained no dopaminergic neurones, was negative for ALDH2, whereas the rat mesencephalon containing the substantia nigra (rt-SN) and, furthermore, rat liver tissue (rt-liver) were intensely stained. (B) Western blot demonstrating the immunoreactivity of rat liver tissue (rt-liver, positive control) and the CSM14.1 cells during differentiation. ALDH2 was not detectable at 33 °C in the CSM14.1 cells, whereas the immunoreactivity for ALDH2 in the ∼54 kDa band increased constantly with subsequent cultivation at 39 °C.

Discussion

The aim of this study was to characterize the potential of the CSM14.1 cells to differentiate under defined culture conditions, and to compare these data with those from in vivo studies concerning the differentiation of dopaminergic neurones. At the permissive temperature of 33 °C these cells proliferated, had a flat neuroepithelial-like morphology and possessed the neural stem cell marker Nestin indicating that, indeed, these cells have the character of neural progenitor (stem) cells when cultured at 33 °C (Lendahl et al. 1990). An elevation of the temperature to 39 °C, which is in the range of that of the rat brain (Gordon, 1990; Martinez-Serrano & Björklund, 1997), and which has been routinely used in previous studies dealing with neural stem cells immortalized with the temperature-sensitive Large T antigen (Frederiksen et al. 1988; Redies et al. 1991; Zhong et al. 1993; Snyder, 1994; Cattaneo & Conti, 1998), together with a deprivation of the serum content of the medium (Winkler et al. 1998) prevented them from expanding. The down-regulation of the Large T antigen was demonstrated in Western blots and observed in a reduction of cell numbers with extended cultivation. CSM14.1 cells then differentiated into a neuronal fate which was shown by morphological changes, a decrease of Nestin protein and, in contrast, an increase of various neuronal marker proteins. CSM14.1 cells were never immunoreactive for GFAP. This indicates that under serum reduced in vitro conditions this cell line derived from a single successfully immortalized cell clone (Zhong et al. 1993) was committed to a neuronal fate as has been shown for other neuroepithelial cells (Buc-Caron, 1995; Cattaneo & Conti, 1998). These phenomena are in line with both the in vivo situation of neuronal differentiation and the in vitro observations published by several groups working with reversibly immortalized stem cell lines under differentiation conditions (Frederiksen et al. 1988; Redies et al. 1991; Snyder, 1994; Cattaneo & Conti, 1998).

Until now, data concerning dopaminergic differentiation of the CSM14.1 cells were rare. CSM14.1 cells were shown to be TH-positive by RNA analysis in vitro (Zhong et al. 1993; Anton et al. 1994, 1995). After grafting into the brain of hemiparkinsonian animals, apomorphine-induced rotations were diminished compared to control animals; nevertheless, TH protein was not detectable in the cells (Anton et al. 1995). This could be due to the content of TH in these cells being too low to be detected when using immunocytochemistry (Anton et al. 1995), but probably high enough to produce some improvement in the grafted animals. No further data, however, were available about the time course of the synthesis of TH or other proteins indicating a dopaminergic differentiation.

One marker of early dopaminergic differentiation is the orphan nuclear retinoic acid receptor Nurr1 which is probably necessary for mesencephalic neurones to develop a dopaminergic fate in vivo (Zetterström et al. 1997; Saucedo-Cardenas et al. 1998; Wallen et al. 1999; Perrone-Capano & Di Porzio, 2000) and in vitro (Wagner et al. 1999). In mice the mRNA of this receptor is detectable at about embryonic day 10.5, whereas the enzyme TH is first found at E11.5 (Zetterström et al. 1997; Perrone-Capano & Di Porzio, 2000). Nurr1 is then present throughout the life of mesencephalic dopaminergic neurones (Bäckman et al. 1999). Nurr1 homozygous knock-out mice lack TH-positive dopaminergic neurones in the substantia nigra pars compacta and die within the first 24 h after birth due to motor deficits (Zetterström et al. 1997; Le et al. 1999). Heterozygous Nurr1 knock-out mice are capable of survival but the dopaminergic neurones in the substantia nigra are more vulnerable to MPTP-induced toxicity than those of wild type littermates (Le et al. 1999). Nurr1, whose endogenous ligand is unknown, seems to activate a promoter of the TH gene (Cazorla et al. 2000; Eells et al. 2001). In addition to the TH production, the subsequent appearance of ALDH2 in vivo found at least in a subpopulation of mesencephalic dopaminergic neurones is an indicator of a mature dopaminergic fate (McCaffery & Dräger, 1994; Haque et al. 1997; Wagner et al. 1999; Wallen et al. 1999). According to these in vivo observations of the development of dopaminergic neurones, a sequential expression of the genes for Nurr1, TH and finally ALDH2 is highly likely. In cultures of CSM14.1 cells used as a model for differentiation of dopaminergic cells in vitro, we observed a differentiation largely corroborating the sequence of molecular events found in vivo (McCaffery & Dräger, 1994; Haque et al. 1997; Zetterström et al. 1997; Saucedo-Cardenas et al. 1998; Wallen et al. 1999; Perrone-Capano & Di Porzio, 2000).

The fact that CSM14.1 cells contain Nurr1 and, after the onset of differentiation, at later time points, also the dopaminergic markers TH and ALDH2, is interesting in view of the study of Wagner et al. (1999). These authors transfected mouse cerebellar stem cells with a retrovirus which over expressed the gene of Nurr1. With their protocol used for differentiation, no TH and ALDH2 were observed in those cells. TH and also ALDH2 were exclusively detectable when Wagner et al. (1999) co-cultured the transfected cells with primary type 1 astrocytes derived from the mesencephalon of E16 mouse embryos. Wagner et al. (1999) supposed that yet unknown factors secreted by the astrocytes could have bound to Nurr1. Nevertheless, after grafting into the adult mouse striatum no TH was seen in the transplanted cerebellar stem cells over-expressing the gene of Nurr1 (Wagner et al. 1999). However, Nurr1 was present in the CSM14.1 cells cultured at 33 °C in DMEM containing 10% FCS, even without any co-cultivation of type 1 astrocytes. The mature dopaminergic marker proteins TH and ALDH2 were detectable with subsequent cultivation at 39 °C in medium supplemented only with 1% FCS. This might indicate that mesencephalic stem cells were committed to a dopaminergic fate whereas cerebellar stem cells were not.

In conclusion, by cultivating CSM14.1 cells under definite conditions we observed a sequentional differentiation largely similar to that reported for the embryonic development of dopaminergic neurones in the rodent mesencephalon (Zetterström et al. 1997; Saucedo-Cardenas et al. 1998; Perrone-Capano & Di Porzio, 2000) thus demonstrating the potential of this conditionally immortalized progenitor cell line to differentiate into dopaminergic neurones. The possibility to expand the CSM14.1 cells (Martinez-Serrano & Björklund, 1997; Björklund & Lindvall, 2000), their differentiation potential and the security of a temperature-sensitive immortalization (Frederiksen et al. 1988; Redies et al. 1991; Zhong et al. 1993; Snyder, 1994; Cattaneo & Conti, 1998) make them an ideal source for intracerebral grafts in animal models of Parkinson's disease (Anton et al. 1994, 1995). Further in vivo experiments to study the therapeutic benefit and the dopaminergic differentiation of grafted CSM14.1 cells in the rat hemiparkinsonian model are under investigation in our group.

Acknowledgments

We gratefully acknowledge Dr D. E. Bredesen (Neuroscience Department, University of California, San Diego) for providing us with the CSM14.1 cells, Dr J. J. Lipsky (Clinical Pharmacology Unit, Mayo Clinic and Foundation, Rochester) for the ALDH2-antibodies and Dr Christian Thode (Neuroscience and Signal Transduction Laboratory, Department of Life Sciences, The Nottingham Trent University, Nottingham) for his helpful comments on the manuscript.

Ancillary