- Email: [email protected]
Differentiation of embryonic chick brain cells in monolayer and reaggregate cultures: A potential model in vitro for neurotoxicity
Toxic. in Vitro Vol. 5, No. 5/6, pp. 419-425, 1991 Printed in Great Britain.All rights reserved
0887-2333/91 $3.00+ 0.00 Copyright © 1991PergamonPress pie
DIFFERENTIATION OF EMBRYONIC CHICK BRAIN CELLS IN MONOLAYER A N D REAGGREGATE CULTURES: A POTENTIAL MODEL IN VITRO FOR NEUROTOXICITY G. G. WYLE-GYURECHand C. A. REINHARDT* Cell Lab, Behavioral Biology, Federal Institute of Technology, ETH Center and Swiss Institute for Alternatives to Animal Testing (SIAT), Turnerstr. 1, CH-8092 Zurich, Switzerland Abstract--In order to develop a model for potential neurotoxicity and teratogenicity, chick brain cells (embryonic day 7) were mechanically dissociated and cultured for up to several months. Differentiation of nerve and glial cells (monolayers in petri dishes and reaggregates in suspension cultures) were monitored with monoclonal antibodies against 68 kDa neurofilament protein (anti-NF) and glial fibrillary acidic protein (anti-GFAP). Anti-NF stains neurons in vitro as they differentiate morphologically;at day 1 many nerve processes already are feebly stained. In monolayer cultures extensive neural networks develop at day 6, which are strongly stained by anti-NF. At day 8 staining fades, days before deterioration becomes visible. In reaggregates a stable differentiation of nerve cells could be observed by intensive anti-NF staining for as long as 3 wk in culture; 5-wk-old cultures still showed substantial staining. The differentiation of GFAP-positive cells takes 3 days in vitro in monolayers as well as in reaggregate cultures. Small groups of marked cells continually increase in number; after 2 wk in culture they are present throughout the reaggregates. The differentiation of astrocytes as measured by anti-GFAP immunostaining is considerably faster in vitro than in vivo. It is suggested that the expression of NF in nerve cells could be used as a sensitive cytotoxicity endpoint, whereas GFAP expression could serve as an endpoint for monitoring differentiation of neural tissue. Introduction Chick embryos have been widely used as a toxicological model by exposing the embryo to chemicals in vivo, using tissue explants or embryos in vitro (e.g. Jelinek et al., 1985; Romanoff, 1972). In vivo (in ovo, fertilized eggs), scoring generally was hampered by problems of even distribution of test chemicals within the albumen compartment of the egg. Kucera and Burnand (1988) used early chick embryos and cultured them in vitro for teratogenicity testing by scoring malformations after a 40-hr incubation period [embryonic day (ED) 1-3], which, however, were evaluated by subjective morphological criteria. Until now, cell cultures from chick embryos rarely have been used for toxicological studies (e.g. Davies and Vernadakis, 1984). However, many embryonic chick-cell types are used in developmental and cell biology, for studies on cell adhesion molecules for example (Grumet and Edelman, 1984). For these purposes various culture conditions for chick brain cells have been developed, some of which are on a serum-free basis. In suspension cultures subjected to constant gyration, brain cells aggregate spontaneously, giving rise to cell reaggregates of a few thousand cells within a few days. Their extraordinary differentiation capability is combined with long-
*To whom all correspondence should be addressed. Abbreviations: anti-GFAP and anti-NF = monoclonal antibodies against GFAP and NF, respectively;BM = Boehringer Mannheim; ED = embryonic day; GFAP = glial fibrillary acidic protein; NF-68 = neurofilament protein 68 kDa; PAP = peroxidase-antiperoxidase; S = Sigma; TBS = Tris buffered saline.
lasting stability and has been particularly well described for rat brain cells (Atterwill, 1987; Honegger, 1985). The strategy behind the development of this model in vitro has been developed bearing in mind the following: (1) current models in vivo for neurotoxicity and teratogenicity of chemicals, drugs and other xenobiotics such as environmental toxins are notoriously unreliable; (2) a model in vitro based on cells taken from chick embryos avoids treatment and killing of mother animals; (3) the use of a non-rodent species may supplement existing models for neurotoxicity in vitro based on rat and mouse brain cells (Atterwill, 1989; Honegger and Werffeli, 1988; Khera and Whalen, 1988); (4) the chick model seems to be particularly promising as the metabolism of xenobiotics by the chick is independent of any interfering maternal influence. Immunocytochemistry of critical cell markers permits the observation of toxicological effects on specific cell types and functions during differentiation. The purpose of this work was to standardize a cell-culture system for characterization of nervous and glial development by using a 68-kDa neurofilament marker (NF-68) and an astroglial marker (glial fibrillary acidic protein, GFAP). Monolayer cultures were used for short-term (6 days) and reaggregates for long-term (up to several months) studies in order to monitor functional differentiation in vitro.
Materials and Methods Chemicals. The culture medium consisted of a 1:1 mixture of Earle's minimum essential medium (Boehringer Mannheim, BM) and MCDB 201
419
420
G.G. WYLE-GYURECHand C. A. REISnAgDT
medium (Sigma, S) with the following supplements: 25 mM-Hepes, sodium bicarbonate (0.7 g/litre), galactose (4.0 g/litre) and D-glucose (4.84 g/litre). The following supplements were then added: 20nMprogesterone (S), 6 nM-triiodothyronine (S), 100 pMputrescine (S), 30 nM-sodium selenite (S), transferrin (BM) (30mg/ml), bovine serum albumin (S) (100mg/ml), linoleic acid (S) (10mg/ml), vitamins (100 x , BM), 5% foetal calf serum (BM). Shortly before use of the culture medium glutamine (4 mol/ litre) was added to the medium. Cell cultures. Fertile chicken eggs were obtained from a local vendor and kept at 37.2°C for 7 days (168 hr). Whole brains of 7-day-old chick embryos (stage 28-29: Hamburger and Hamilton, 1951) were isolated in cold Mg 2+- and Ca2+-free Hepes buffer. A single cell suspension was obtained by mechanical dissociation: the tissue was pipetted through nylon sieves of 200, 120 and 45/~m mesh width. The viability was 90-95% based on typan blue exclusion. Monolayer cultures were produced by plating 6 x 105 cells in 24-multiwell plates coated with polyL-lysine (BM) and 0.02% porcine skin collagen (Pentapharm, Basel). lmmunocytochemical staining was performed with cultures on coated glass slides in 24-multiwell plates. For reaggregate cultures the cell suspension was added to 3.0 ml culture medium in 25-ml Erlenmeyer flasks. The Erlenmeyer flasks were kept on a gyratory shaker (Infors AG, Bottmingen/Basel). The speed of rotation was increased daily during the first week in vitro from 70 to 82 rpm. For immunocytochemical staining, reaggregates were explanted (explant cultures) onto coated glass slides in multiwell plates as described for monolayer cultures. The cultures were incubated at 37°C in humidified air containing 5% CO2. Cryostat sections. Normal chick embryos were developed for 7 (stage 29), 12 (stage 35) and 13 days (stage 36-37) and prepared as follows: fresh whole chick embryo heads were frozen in isopentane cooled by liquid nitrogen. Sections l0/~m thick were cut at - 2 0 ° C and put onto gelatin-coated glass slides. The sections were air dried and stored at - 2 0 ° C . Immunocytochemistry and histochemistry. Monolayer and explanted reaggregates on glass coverslips were rinsed in cold Tris buffered saline (TBS) and fixed in acetone for 15 min at room temperature. Cryostat sections were defrosted and fixed in 4% paraformaldehyde in 0.1 M-phosphate buffer (pH 7.4) for l0 min and subsequently kept in absolute ethanol for 15 min. For staining, the unlabelled antibody-peroxidaseantiperoxidase (PAP) method after Sternberger (1986) was used. After fixation and washing (5 min in TBS) cultures and sections could be kept in TBS at 4°C up to 3 days. After TBS was removed, an incubation with normal rabbit serum (Dakopatts; diluted 1:25) for l0 min at room temperature served to block non-specific background. Excess serum was removed and the cultures and sections were incubated with the primary antibody, either against NF-68 (BM) or against G F A P (BM) for 30 min at 37°C in a moist chamber, then washed in TBS for 5 min, incubated with rabbit anti-mouse immunoglobulins IgG (Dakopatts) for 20 min at 37°C, rinsed again for
5 min in TBS and incubated with PAP for 20 min at 37°C. After rinsing in TBS for 5 min, incubation with diaminobenzidine (0.6 mg/ml in TBS) was carried out for 5-10min at room temperature. Sections were counterstained with methylene blue for 10 sec and differentiated with 0.1% acetic acid for another 10 sec. Results and Discussion
Morphology With regard to morphological development, neuronal cell outgrowth had already started after a few hours in monolayer cultures (Plates lb and 2b). Within the first 2 days, an extensive network of neurons with short neurites formed; the network contained very small cell clusters. After 3--4 days in vitro, the cell clusters grew into pseudoganglia, connected with a network of long, string-like cells, which may grow on top of substrate-attached cells (Plates 1d and 2d). Between days 6 and 7, pseudoganglia and neuronal networks become fully developed and, in particular, outgrown neurites become thicker. Monolayer cultures can be kept up to approximately 16 days, after which time they deteriorate. In reaggregate cultures, loose and irregular cell clusters form within the first few days, developing within the first 2 wk into solid and round reaggregates. As reaggregates grow older, they also grow larger and are stable for several months (Plates 3 and 4). Neurofilament polypeptide 68 kDa In monolayer cultures, NF-68 is expressed in many cell processes at day 1 (Plate la), although staining is still rather weak. From day 3 onwards, many (but not all) of the cell connections between the cell clusters stain intensely; this staining of the neural network is most intense between days 6 and 7 (Plate lc). Thereafter, cell connections may still be visible, but NFstaining diminishes with culture age. NF-positive nerve cells have a string-like, bipolar appearance. In reaggregate cultures, nerve cells also show positive staining for N F on day 1 (Plate 3a), but they differentiate to a much larger extent than in monolayer cultures. They can develop long and elaborate neurites (Plate 3c) and some cells have the typical shape of Purkinje cells (arrow, Plate 3b) which are stable for at least 3 wk, after which time the number of nerve cells decreases. Honegger (1985) found that reaggregate cultures of foetal rat brain cells form similar patterns of cell alignment. In addition, he observed myelinization of nerve fibres and full development of various types of synapses. Glial fibrillary acidic protein Cells stained for G F A P appear in monolayer cultures on day 3 in vitro (Plate 2a). They are arranged in small groups inside or around pseudoganglia which grow in number and size as the culture grows older (Plate 2c). GFAP-positive cells usually are of stellate type, as described for type II astrocyes (Raft et al., 1983). They seem to be quite robust cells, as they still show intensive staining, even when the cultures are clearly damaged.
Plate 1. Light micrographs of monolayer cultures of 7-day-old embryonic chick brain cells ( x 200). The cells are stained for neurofilament polypeptide 68 kDa (NF-68) by the peroxidase-antiperoxidase (PAP) method. (a) Bright light, and (b) phase contrast of l-day-old cultures; (c) bright light and (d) phase contrast of 6-day-old cultures.
421
Plate 2. Light micrographs of monolayer cultures of 7-day-old embryonic chick brain cells ( × 200). The cells are stained for gliai fibrillary acidic protein (GFAP) by the PAP method. (a) Bright light and (b) phase contrast of 1-day-old cultures; (c) bright light, and (d) phase contrast of 6-day-old cultures.
422
L
Plate 3 (left). Light micrographs of reaggregates of 7-day-old embryonic chick brain ceils, explanted onto glass slides I day before staining ( x 190). The ceils are stained for NF-68 by the PAP method. (a) 2-day-old reaggregates; (b) 7-day-old reaggregates; (c) 15-day-old reaggregates. Plate 4 (right). Light micrographs of reaggregates of 7-day-old embryonic chick brain cells, explanted onto glass slides I day before staining ( x 190). The cells are stained for GFAP by the PAP method. (a) 3-day-old reaggregates; (b) 7-day-old reaggregates; (c) 24-day-old reaggregates.
423
Embryonic chick brain-cell cultures In reaggregates, GFAP-positive cells appear after 3 days in small clusters, as in monolayer cultures. After 2 wk they are abundant throughout the entire reaggregate (Plate 4). Brain sections
Cryostat sections of brains of ED7 and ED 12 chick embryos show no GFAP-positive cells. Brains of EDI3 embryos, however, show faint staining in an outer layer of the brain stem and the hemisphere. NF-stained cells are already present in brains of E7 embryos in the optical lobe and brain stem, but not yet in the hemispheres. El2 and El3 embryos show NF-positive cells in most parts of the brain. Astrocytes and their marker GFAP develop very late in the chick brain (Lemmon, 1985). Sections of embryonic chick brain show a faint staining in a few small areas of the brain, beginning on EDI3, which corresponds to 6 days in vitro. In vitro, the first GFAP-positive cells have appeared by day 3; our culture conditions therefore trigger the development of astrocytes. Iacovitti et al. (1987) described the same phenomenon for the marker of tyrosine hydroxylase and concluded that cells are influenced by epigenetic factors (such as nerve growth factor) found in the culture medium. Thus, in contrast to NF, GFAP is not expressed at the start of the cultures but fully develops in vitro within a few days. Thus, GFAP could serve as a marker to trace any changes in development of astrocytes in vitro.
Conclusions
Thus, reaggregate cultures are appropriate for simulation of differentiation of neuronal cells because they represent a stable tissue-like cell complex. For this reason, they are well suited for studying longterm toxicological effects. Monolayer cultures, on the other hand, are a suitable system for short-term tests, as they are more sensitive to toxic insults owing to their direct exposure to a test substance. Moreover, morphological monitoring is easy in monolayer cultures and single cells can be followed up. When grown on microtitre plates, whole test series can be automatically performed and efficiently evaluated. Further analysis of cell markers for oligodendrocytes, endothelial cells, fibroblasts and ependymal cells, as well as functional analysis of the dopaminergic system in the reaggregates, is under way (Reinhardt et al., 1991) in order to develop fully a long-term assay system in vitro for neurotoxicological screening before animal testing is performed. Acknowledgements--This work has been supported by the Schweiz. Gesellschaft fuer Tierschutz (Zurich), the foundation Fonds fuer versuchstierfreie Forschung (Zurich) and
425
the Swiss National Science Foundation (project no. 31.8889.86). REFERENCES
Atterwill C. K. (1987) Brain reaggregate cultures in neurotoxicological investigations. In In Vitro Methods in Toxicology. Edited by C. K. Atterwill and C. E. Steele. pp. 133-164. Cambridge University Press, Cambridge. AtterwiH C K. (1989) Brain reaggregate cultures in neurotoxicological investigations: adaptational and neurodegenerative processes following lesions. Molecular Toxicology 1 (4), 489-502. Davies D. L. and Vernadakis A. (1984) Effects of ethanol on cultured glial cells: proliferation and glutamine synthetase activity. Developmental Brain Research 16, 27-35. Grumet M. and Edelman G. M. (1984) Heterotypic binding between neuronal membrane vesicles and glial cells is mediated by specific cell adhesion molecules. Journal of Cell Biology 96, 1746-1756. Hamburger V. and Hamilton H. L. (1951) A series of normal stages in the development of the chick embryo. Journal of Morphology 88, 49-92. Honegger P. (1985) Biochemical differentiation in serumfree aggregating brain cell cultures. In Cell Culture in the Neurosciences. Edited by J. E. Bottenstein and G. Sato. pp. 223-243. Plenum Press, New York. Honegger P. and Werffeli P. (1988) Use of aggregating cell cultures for toxicological studies. Experientia 44, 817-823. Iacovitti C., Lee J., Joh T. H. and Reis D. J. (1987) Expression of tyrosine hydroxylase in neurons of cultured cerebral cortex: evidence for phenotypic plasticity in neurons of the CNS. Journal of Neuroscience 7, 1264-1270. Jelinek R., Perka M. and Rychter Z. (1985) One hundred and thirty compounds tested with the chick embryo toxicity screeningtest (CHEST). Indian Journal of Experimental Biology 23, 588-595. Khera K. S. and Whalen C. (1988) Detection of neuroteratogens with an in vitro cytotoxicity assay using primary monolayers cultured from dissociated foetal rat brains. Toxicology in Vitro 2, 257-273. Kucera P. and Burnand M. B. (1988) Teratogenicity screening in standardized chick embryo culture: effects of dexamethasone and diphenylhydantoin. Experientia 44, 827-833. Lemmon V. (1985) Monoclonal antibodies specific for glia in the chick nervous system. Developmental Brain Research 23, 111-120. Raft M. C., Miller R. H. and Noble M. (1983) A glial progenitor cell that develops into an astrocyte or an oligodendrocyte depending on culture medium. Nature, London 303, 390-396. Reinhardt C. A., Bienz A., Romano-Diethelm T. and Wyle-Gyurech G. G. (1991) In vitro differentiation of embryonic chick brain cells: development of a neuroteratology test system. (Abstract). Experientia 47, A5. Romanoff A. (1972) The Avian Embryo. Wiley Interscience, New York. Sternberger L. A. (1986) Immunocytochemistry. pp. 90--200. John Wiley and Sons, New York.
0887-2333/91 $3.00+ 0.00 Copyright © 1991PergamonPress pie
DIFFERENTIATION OF EMBRYONIC CHICK BRAIN CELLS IN MONOLAYER A N D REAGGREGATE CULTURES: A POTENTIAL MODEL IN VITRO FOR NEUROTOXICITY G. G. WYLE-GYURECHand C. A. REINHARDT* Cell Lab, Behavioral Biology, Federal Institute of Technology, ETH Center and Swiss Institute for Alternatives to Animal Testing (SIAT), Turnerstr. 1, CH-8092 Zurich, Switzerland Abstract--In order to develop a model for potential neurotoxicity and teratogenicity, chick brain cells (embryonic day 7) were mechanically dissociated and cultured for up to several months. Differentiation of nerve and glial cells (monolayers in petri dishes and reaggregates in suspension cultures) were monitored with monoclonal antibodies against 68 kDa neurofilament protein (anti-NF) and glial fibrillary acidic protein (anti-GFAP). Anti-NF stains neurons in vitro as they differentiate morphologically;at day 1 many nerve processes already are feebly stained. In monolayer cultures extensive neural networks develop at day 6, which are strongly stained by anti-NF. At day 8 staining fades, days before deterioration becomes visible. In reaggregates a stable differentiation of nerve cells could be observed by intensive anti-NF staining for as long as 3 wk in culture; 5-wk-old cultures still showed substantial staining. The differentiation of GFAP-positive cells takes 3 days in vitro in monolayers as well as in reaggregate cultures. Small groups of marked cells continually increase in number; after 2 wk in culture they are present throughout the reaggregates. The differentiation of astrocytes as measured by anti-GFAP immunostaining is considerably faster in vitro than in vivo. It is suggested that the expression of NF in nerve cells could be used as a sensitive cytotoxicity endpoint, whereas GFAP expression could serve as an endpoint for monitoring differentiation of neural tissue. Introduction Chick embryos have been widely used as a toxicological model by exposing the embryo to chemicals in vivo, using tissue explants or embryos in vitro (e.g. Jelinek et al., 1985; Romanoff, 1972). In vivo (in ovo, fertilized eggs), scoring generally was hampered by problems of even distribution of test chemicals within the albumen compartment of the egg. Kucera and Burnand (1988) used early chick embryos and cultured them in vitro for teratogenicity testing by scoring malformations after a 40-hr incubation period [embryonic day (ED) 1-3], which, however, were evaluated by subjective morphological criteria. Until now, cell cultures from chick embryos rarely have been used for toxicological studies (e.g. Davies and Vernadakis, 1984). However, many embryonic chick-cell types are used in developmental and cell biology, for studies on cell adhesion molecules for example (Grumet and Edelman, 1984). For these purposes various culture conditions for chick brain cells have been developed, some of which are on a serum-free basis. In suspension cultures subjected to constant gyration, brain cells aggregate spontaneously, giving rise to cell reaggregates of a few thousand cells within a few days. Their extraordinary differentiation capability is combined with long-
*To whom all correspondence should be addressed. Abbreviations: anti-GFAP and anti-NF = monoclonal antibodies against GFAP and NF, respectively;BM = Boehringer Mannheim; ED = embryonic day; GFAP = glial fibrillary acidic protein; NF-68 = neurofilament protein 68 kDa; PAP = peroxidase-antiperoxidase; S = Sigma; TBS = Tris buffered saline.
lasting stability and has been particularly well described for rat brain cells (Atterwill, 1987; Honegger, 1985). The strategy behind the development of this model in vitro has been developed bearing in mind the following: (1) current models in vivo for neurotoxicity and teratogenicity of chemicals, drugs and other xenobiotics such as environmental toxins are notoriously unreliable; (2) a model in vitro based on cells taken from chick embryos avoids treatment and killing of mother animals; (3) the use of a non-rodent species may supplement existing models for neurotoxicity in vitro based on rat and mouse brain cells (Atterwill, 1989; Honegger and Werffeli, 1988; Khera and Whalen, 1988); (4) the chick model seems to be particularly promising as the metabolism of xenobiotics by the chick is independent of any interfering maternal influence. Immunocytochemistry of critical cell markers permits the observation of toxicological effects on specific cell types and functions during differentiation. The purpose of this work was to standardize a cell-culture system for characterization of nervous and glial development by using a 68-kDa neurofilament marker (NF-68) and an astroglial marker (glial fibrillary acidic protein, GFAP). Monolayer cultures were used for short-term (6 days) and reaggregates for long-term (up to several months) studies in order to monitor functional differentiation in vitro.
Materials and Methods Chemicals. The culture medium consisted of a 1:1 mixture of Earle's minimum essential medium (Boehringer Mannheim, BM) and MCDB 201
419
420
G.G. WYLE-GYURECHand C. A. REISnAgDT
medium (Sigma, S) with the following supplements: 25 mM-Hepes, sodium bicarbonate (0.7 g/litre), galactose (4.0 g/litre) and D-glucose (4.84 g/litre). The following supplements were then added: 20nMprogesterone (S), 6 nM-triiodothyronine (S), 100 pMputrescine (S), 30 nM-sodium selenite (S), transferrin (BM) (30mg/ml), bovine serum albumin (S) (100mg/ml), linoleic acid (S) (10mg/ml), vitamins (100 x , BM), 5% foetal calf serum (BM). Shortly before use of the culture medium glutamine (4 mol/ litre) was added to the medium. Cell cultures. Fertile chicken eggs were obtained from a local vendor and kept at 37.2°C for 7 days (168 hr). Whole brains of 7-day-old chick embryos (stage 28-29: Hamburger and Hamilton, 1951) were isolated in cold Mg 2+- and Ca2+-free Hepes buffer. A single cell suspension was obtained by mechanical dissociation: the tissue was pipetted through nylon sieves of 200, 120 and 45/~m mesh width. The viability was 90-95% based on typan blue exclusion. Monolayer cultures were produced by plating 6 x 105 cells in 24-multiwell plates coated with polyL-lysine (BM) and 0.02% porcine skin collagen (Pentapharm, Basel). lmmunocytochemical staining was performed with cultures on coated glass slides in 24-multiwell plates. For reaggregate cultures the cell suspension was added to 3.0 ml culture medium in 25-ml Erlenmeyer flasks. The Erlenmeyer flasks were kept on a gyratory shaker (Infors AG, Bottmingen/Basel). The speed of rotation was increased daily during the first week in vitro from 70 to 82 rpm. For immunocytochemical staining, reaggregates were explanted (explant cultures) onto coated glass slides in multiwell plates as described for monolayer cultures. The cultures were incubated at 37°C in humidified air containing 5% CO2. Cryostat sections. Normal chick embryos were developed for 7 (stage 29), 12 (stage 35) and 13 days (stage 36-37) and prepared as follows: fresh whole chick embryo heads were frozen in isopentane cooled by liquid nitrogen. Sections l0/~m thick were cut at - 2 0 ° C and put onto gelatin-coated glass slides. The sections were air dried and stored at - 2 0 ° C . Immunocytochemistry and histochemistry. Monolayer and explanted reaggregates on glass coverslips were rinsed in cold Tris buffered saline (TBS) and fixed in acetone for 15 min at room temperature. Cryostat sections were defrosted and fixed in 4% paraformaldehyde in 0.1 M-phosphate buffer (pH 7.4) for l0 min and subsequently kept in absolute ethanol for 15 min. For staining, the unlabelled antibody-peroxidaseantiperoxidase (PAP) method after Sternberger (1986) was used. After fixation and washing (5 min in TBS) cultures and sections could be kept in TBS at 4°C up to 3 days. After TBS was removed, an incubation with normal rabbit serum (Dakopatts; diluted 1:25) for l0 min at room temperature served to block non-specific background. Excess serum was removed and the cultures and sections were incubated with the primary antibody, either against NF-68 (BM) or against G F A P (BM) for 30 min at 37°C in a moist chamber, then washed in TBS for 5 min, incubated with rabbit anti-mouse immunoglobulins IgG (Dakopatts) for 20 min at 37°C, rinsed again for
5 min in TBS and incubated with PAP for 20 min at 37°C. After rinsing in TBS for 5 min, incubation with diaminobenzidine (0.6 mg/ml in TBS) was carried out for 5-10min at room temperature. Sections were counterstained with methylene blue for 10 sec and differentiated with 0.1% acetic acid for another 10 sec. Results and Discussion
Morphology With regard to morphological development, neuronal cell outgrowth had already started after a few hours in monolayer cultures (Plates lb and 2b). Within the first 2 days, an extensive network of neurons with short neurites formed; the network contained very small cell clusters. After 3--4 days in vitro, the cell clusters grew into pseudoganglia, connected with a network of long, string-like cells, which may grow on top of substrate-attached cells (Plates 1d and 2d). Between days 6 and 7, pseudoganglia and neuronal networks become fully developed and, in particular, outgrown neurites become thicker. Monolayer cultures can be kept up to approximately 16 days, after which time they deteriorate. In reaggregate cultures, loose and irregular cell clusters form within the first few days, developing within the first 2 wk into solid and round reaggregates. As reaggregates grow older, they also grow larger and are stable for several months (Plates 3 and 4). Neurofilament polypeptide 68 kDa In monolayer cultures, NF-68 is expressed in many cell processes at day 1 (Plate la), although staining is still rather weak. From day 3 onwards, many (but not all) of the cell connections between the cell clusters stain intensely; this staining of the neural network is most intense between days 6 and 7 (Plate lc). Thereafter, cell connections may still be visible, but NFstaining diminishes with culture age. NF-positive nerve cells have a string-like, bipolar appearance. In reaggregate cultures, nerve cells also show positive staining for N F on day 1 (Plate 3a), but they differentiate to a much larger extent than in monolayer cultures. They can develop long and elaborate neurites (Plate 3c) and some cells have the typical shape of Purkinje cells (arrow, Plate 3b) which are stable for at least 3 wk, after which time the number of nerve cells decreases. Honegger (1985) found that reaggregate cultures of foetal rat brain cells form similar patterns of cell alignment. In addition, he observed myelinization of nerve fibres and full development of various types of synapses. Glial fibrillary acidic protein Cells stained for G F A P appear in monolayer cultures on day 3 in vitro (Plate 2a). They are arranged in small groups inside or around pseudoganglia which grow in number and size as the culture grows older (Plate 2c). GFAP-positive cells usually are of stellate type, as described for type II astrocyes (Raft et al., 1983). They seem to be quite robust cells, as they still show intensive staining, even when the cultures are clearly damaged.
Plate 1. Light micrographs of monolayer cultures of 7-day-old embryonic chick brain cells ( x 200). The cells are stained for neurofilament polypeptide 68 kDa (NF-68) by the peroxidase-antiperoxidase (PAP) method. (a) Bright light, and (b) phase contrast of l-day-old cultures; (c) bright light and (d) phase contrast of 6-day-old cultures.
421
Plate 2. Light micrographs of monolayer cultures of 7-day-old embryonic chick brain cells ( × 200). The cells are stained for gliai fibrillary acidic protein (GFAP) by the PAP method. (a) Bright light and (b) phase contrast of 1-day-old cultures; (c) bright light, and (d) phase contrast of 6-day-old cultures.
422
L
Plate 3 (left). Light micrographs of reaggregates of 7-day-old embryonic chick brain ceils, explanted onto glass slides I day before staining ( x 190). The ceils are stained for NF-68 by the PAP method. (a) 2-day-old reaggregates; (b) 7-day-old reaggregates; (c) 15-day-old reaggregates. Plate 4 (right). Light micrographs of reaggregates of 7-day-old embryonic chick brain cells, explanted onto glass slides I day before staining ( x 190). The cells are stained for GFAP by the PAP method. (a) 3-day-old reaggregates; (b) 7-day-old reaggregates; (c) 24-day-old reaggregates.
423
Embryonic chick brain-cell cultures In reaggregates, GFAP-positive cells appear after 3 days in small clusters, as in monolayer cultures. After 2 wk they are abundant throughout the entire reaggregate (Plate 4). Brain sections
Cryostat sections of brains of ED7 and ED 12 chick embryos show no GFAP-positive cells. Brains of EDI3 embryos, however, show faint staining in an outer layer of the brain stem and the hemisphere. NF-stained cells are already present in brains of E7 embryos in the optical lobe and brain stem, but not yet in the hemispheres. El2 and El3 embryos show NF-positive cells in most parts of the brain. Astrocytes and their marker GFAP develop very late in the chick brain (Lemmon, 1985). Sections of embryonic chick brain show a faint staining in a few small areas of the brain, beginning on EDI3, which corresponds to 6 days in vitro. In vitro, the first GFAP-positive cells have appeared by day 3; our culture conditions therefore trigger the development of astrocytes. Iacovitti et al. (1987) described the same phenomenon for the marker of tyrosine hydroxylase and concluded that cells are influenced by epigenetic factors (such as nerve growth factor) found in the culture medium. Thus, in contrast to NF, GFAP is not expressed at the start of the cultures but fully develops in vitro within a few days. Thus, GFAP could serve as a marker to trace any changes in development of astrocytes in vitro.
Conclusions
Thus, reaggregate cultures are appropriate for simulation of differentiation of neuronal cells because they represent a stable tissue-like cell complex. For this reason, they are well suited for studying longterm toxicological effects. Monolayer cultures, on the other hand, are a suitable system for short-term tests, as they are more sensitive to toxic insults owing to their direct exposure to a test substance. Moreover, morphological monitoring is easy in monolayer cultures and single cells can be followed up. When grown on microtitre plates, whole test series can be automatically performed and efficiently evaluated. Further analysis of cell markers for oligodendrocytes, endothelial cells, fibroblasts and ependymal cells, as well as functional analysis of the dopaminergic system in the reaggregates, is under way (Reinhardt et al., 1991) in order to develop fully a long-term assay system in vitro for neurotoxicological screening before animal testing is performed. Acknowledgements--This work has been supported by the Schweiz. Gesellschaft fuer Tierschutz (Zurich), the foundation Fonds fuer versuchstierfreie Forschung (Zurich) and
425
the Swiss National Science Foundation (project no. 31.8889.86). REFERENCES
Atterwill C. K. (1987) Brain reaggregate cultures in neurotoxicological investigations. In In Vitro Methods in Toxicology. Edited by C. K. Atterwill and C. E. Steele. pp. 133-164. Cambridge University Press, Cambridge. AtterwiH C K. (1989) Brain reaggregate cultures in neurotoxicological investigations: adaptational and neurodegenerative processes following lesions. Molecular Toxicology 1 (4), 489-502. Davies D. L. and Vernadakis A. (1984) Effects of ethanol on cultured glial cells: proliferation and glutamine synthetase activity. Developmental Brain Research 16, 27-35. Grumet M. and Edelman G. M. (1984) Heterotypic binding between neuronal membrane vesicles and glial cells is mediated by specific cell adhesion molecules. Journal of Cell Biology 96, 1746-1756. Hamburger V. and Hamilton H. L. (1951) A series of normal stages in the development of the chick embryo. Journal of Morphology 88, 49-92. Honegger P. (1985) Biochemical differentiation in serumfree aggregating brain cell cultures. In Cell Culture in the Neurosciences. Edited by J. E. Bottenstein and G. Sato. pp. 223-243. Plenum Press, New York. Honegger P. and Werffeli P. (1988) Use of aggregating cell cultures for toxicological studies. Experientia 44, 817-823. Iacovitti C., Lee J., Joh T. H. and Reis D. J. (1987) Expression of tyrosine hydroxylase in neurons of cultured cerebral cortex: evidence for phenotypic plasticity in neurons of the CNS. Journal of Neuroscience 7, 1264-1270. Jelinek R., Perka M. and Rychter Z. (1985) One hundred and thirty compounds tested with the chick embryo toxicity screeningtest (CHEST). Indian Journal of Experimental Biology 23, 588-595. Khera K. S. and Whalen C. (1988) Detection of neuroteratogens with an in vitro cytotoxicity assay using primary monolayers cultured from dissociated foetal rat brains. Toxicology in Vitro 2, 257-273. Kucera P. and Burnand M. B. (1988) Teratogenicity screening in standardized chick embryo culture: effects of dexamethasone and diphenylhydantoin. Experientia 44, 827-833. Lemmon V. (1985) Monoclonal antibodies specific for glia in the chick nervous system. Developmental Brain Research 23, 111-120. Raft M. C., Miller R. H. and Noble M. (1983) A glial progenitor cell that develops into an astrocyte or an oligodendrocyte depending on culture medium. Nature, London 303, 390-396. Reinhardt C. A., Bienz A., Romano-Diethelm T. and Wyle-Gyurech G. G. (1991) In vitro differentiation of embryonic chick brain cells: development of a neuroteratology test system. (Abstract). Experientia 47, A5. Romanoff A. (1972) The Avian Embryo. Wiley Interscience, New York. Sternberger L. A. (1986) Immunocytochemistry. pp. 90--200. John Wiley and Sons, New York.