- Email: [email protected]
Patterns of reaggregation and formation of linear aggregate chains in collagen-well cultures of dissociated mouse brain and spinal cord cells
Brain Research, 198 (1980) 167-182
167
© Elsevier/North-Holland Biomedical Press
P A T T E R N S OF R E A G G R E G A T I O N A N D F O R M A T I O N OF L I N E A R A G G R E G A T E C H A I N S I N C O L L A G E N - W E L L C U L T U R E S OF D I S S O C I A T E D M O U S E B R A I N A N D S P I N A L C O R D CELLS
ILYA V. VICTOROV and TERESA L. KRUKOFF* Nervous Tissue Culture Unit, Laboratory of Functional Synaptology, Brain Research Institute, U.S.S.R. Academy of Medical Sciences, Moscow 107120 (U.S.S.R.)
(Accepted April 3rd, 1980) Key words: embryonic mouse brain and spinal cord - - dissociation - - reaggregation - - formation
of aggregate systems - - collagen-well cultures
SUMMARY With the use of the newly developed collagen-well culture technique spontaneous reaggregation of cells from dissociated embryonic mouse brain and spinal cord were studied. Within 20 h in culture, aggregates are formed and settled onto collagen substrate. Two patterns of aggregate arrangement were observed: random and linear. Linear chains of aggregates appeared to be more characteristic of dissociated spinal cord cells, although the linear patterns were not uncommon in cultures of dissociated cortex. Formation of aggregate chains appeared to be dependent on the stage of neuronal and glial differentiation. After attachment to the collagen substrate, the general pattern of aggregate organization was not greatly altered except for changes which resulted from cellular migration and proliferation, the formation of connections between aggregates, or incorporation of small aggregates into larger ones. The number of non-aggregated cells in collagen-well cultures was small. Single, non-aggregated neurons seldom survived individually. Fiber connections between aggregates began to form after the first day in vitro, and by 2 or 3 days, the growing fibers formed neuritic bridges connecting aggregates. By the end of the first week growing fibers often organized compact bundles, but part connections between aggregates were presented by separate fibers in a diffused manner. Silver impregnation revealed that these connections were formed by the axons of neurons located in the aggregates. Thus, * Post-graduatestudent from Department of Anatomy, University of Saskatchewan Medical College, Saskatoon, Sask., Canada, as participant of Canada - - U.S.S.R. graduate student exchange program. Present adress: Health Science Center, Memorial University of Newfoundland, St. John's, Newfoundland, Canada.
168 progression of the above described processes resulted in the 'de novo' formation of linear organized or random systems of interconnected neuronal centers.
INTRODUCTION Cell dissociation and reaggregation techniques have provided a powerful analytic tool for examining the cell-cell interactions involved in the histogenesis of the central nervous system9-12. According to the dissociation and reaggregation procedure developed by Moscona20, 21, dissociated cells are allowed to reaggregate into tissue within gyratory Erlenmeyer flasks where three-dimensional freedom is afforded the cells. Crain and Bornstein 7 planted dissociated fetal mouse brain cells onto collagen-coated coverslips (in Maximow assemblies) and used electrophysiological techniques to study synapse formation between spontaneously formed aggregates. Trenkner and Sidman 35 have studied the reaggregation of mouse cerebellum in microwell cultures. Thus, depending on the specific reaggregation and cultivation methods employed, different aspects of nervous tissue development can be accurately studied. This paper describes a collagen-well culture technique that was used to study spontaneous reaggregation of dissociated cells from embryonic mouse cortex and spinal cord. Results obtained in the present study indicated that the dissociated cells in the collagen well quickly reorganized into aggregates before attaching to the substrate. An unexpected pattern of linear aggregate organization was discovered, especially in cultures of dissociated spinal cord cells. Cell differentiation, growth of neurites and development of fiber connections between aggregates in spatially organized and random systems are described. Preliminary data indicate that this culture system is applicable to the study of factors involved in spontaneous reaggregation of brain cells, formation of spatially arranged system of aggregates and development of connections between re-established neuronal centers. MATERIAL AND METHODS
Dissociation procedure and culture conditions C57BL/6J mice fetuses were used. Dissociated cells were prepared by the modified method of Moscona20, 21. In brief, small fragments of spinal cord (12-16-dayold embryos) or cerebral cortex (16-20-day-old embryos) were washed in Ca 2÷- and Mg2+-free Tyrode solution (CMF) and next incubated in trypsin solution (0.25 ~ in CMF) for 10-15 min at 37 °C. The trypsin was removed and the tissue fragments were washed in a mixture of equal parts of Eagle's minimal essential medium (Dulbecco's modification, DMEM), fetal calf serum and Simms X7 balanced salt solution (BSS). After one wash in nutrient medium (see below), the tissue was dissociated into single cells in a new portion of the medium by repeated passage through a fine-bore Pasteur pippete. The cell suspension was centrifuged and the supernatant was discarded. The cells were resuspended in nutrient medium. Successful disaggregation and more active
169 reaggregafion were also achieved when complete BSS was used instead of CMF and trypsin was dissolved in DMEM. Finally, one drop of the cell suspension (2.0 x 106-1.5 x 107 cells/ml) was placed on a coverslip that had been collagen-coated according to the technique described below. Each coverslip was then mounted in a Maximow double coverslip assembly and incubated at 35 4- 0.5 °C. Cultures were left undisturbed for 20 h and examined at regular daily intervals thereafter. Nutrient medium changes were made biweekly. Cultures were maintained for various time intervals, up to a maximum of 24 days. The living cultures were observed with brightfield microscopy or lateral illumination 8°, using an inverted tissue culture microscope (LOMO, MBI-13). In the histological studies the modified Holmes silver impregnation method a~ and the method of Bodian 17 were used.
Medium Nutrient medium consisted of human placental serum (20 ~), fetal calf serum (20 ~), Dulbecco's modification of Eagle's MEM (40 ~) and Simms X7 BSS (20 ~); the medium also contained 800 m g ~ glucose, 0.25 U/ml insulin and 10-2 M HEPES buffer.
Collagen-coated coverslips ('collagen wells') A modification of the collagen-coating techniquea was used in the following procedure. Two consistencies of dialyzed collagen were used: one (A) was dialyzed against water for 3--4 days and the other (B) was dialyzed for 1-2 days (collagen-B was therefore more fluid than collagen-A). Round glass coverslips (diameter 22 mm) were used. Using collagen-A, a complete collagen ring was made near the periphery of each coverslip. After exposing the ringed coverslips to ammonia vapor for 1 min, the coverslips were washed thoroughly in distilled water. Thorough washing is critical to avoid premature collagen reconstitution in the following step. After removing excess of water, a drop of collagen-B was placed within each collagen ring and evenly spread to meet all inside edges of the ring. Once again the coverslips were exposed to ammonia vapor, and collagen wells are formed. The collagen well has the concave profile versus the conventional convex profile of a coverslip coated with collagen according to the usual technique 3. The completed coverslips were washed in distilled water ( × 3) and in BSS ( × 3). The collagen-coated coverslips were stored in Columbia dishes containing 9 ml BSS, 0.5 ml fetal calf serum and 0.5 ml 2 0 ~ glucose. RESULTS
Aggregate formation Directly after seeding of the collagen wells, single isolated cells remained suspended in the medium (Fig. 1A). In spite of complete dissociation, cells were often observed to quickly adhere to one another to form small clumps which did not attach to the collagen during the initial period of cultivation. A comparatively large volume of medium allowed free movement of the single cells and cell clumps. Within 16-20 h
170
Fig. 1. Dissociated brain cells and the formation of aggregates in collagen wells during the initial 2-3 days in vitro. Living, unstained preparations. Bright-field illumination. Scale bars : 50 /m~. A: dissociated brain cells immediately after seeding. B: cell aggregates have formed and fixed to the collagen substrate. Only a few cells remain free, 20 h in vitro. C: initial growth of cell processes and formation of primary connections between aggregates, 2 days in vitro. D: flattening of aggregates and formation of fiber bundles. Note the active migration of glial cells along the fiber bundles: 3 days in vitro.
171 in vitro, spontaneous reaggregation of the single cells was near completion (Fig. 1B). In most cases, only a small percentage of cells did not become incorporated into aggregates. It should be stressed that reaggregation always occurred before the cells had attached to the collagen. Most aggregates attached to the substrate within the first day in vitro, and as early as 24-36 h cellular migration and initial outgrowth of cell processes were observed. By 2-3 days, distinct fiber outgrowth was clearly seen (Fig. 1C, D). The first fiber connections between aggregates were observed after the same period of time in vitro. It was difficult to characterize exactly the nature of the fiber outgrowth in living preparations (i.e. neuronal or glial); however, histological studies with silver impregnation showed that the number of nerve fibers in 2 or 3 day cultures was relatively low. The number of non-aggregated neuronal and glial cells decreased with aging of the cultures, either because of single cell incorporation into aggregates, or because of degeneration of some of the single cells (compare, for example, the number of nonaggregated cells in Figs. 2 and 3). Note that single astrocytes did occasionally grow isolated from the aggregates to form a network of glial fibers around the aggregates. As a rule, however, isolated neurons were not observed in mature reaggregate cultures grown in collagen wells.
Pattern formation The most interesting observation of reaggregate cultures in collagen wells was the spatial organization pattern of aggregate systems over the surface of the coverslip. During the first day of cultivation, newly formed aggregates attached to the collagen surface in two main configurations: random and linear. In some cases these two extreme types of arrangements were found in the same culture. It was found that the initial overall type of aggregate arrangement in each culture was generally maintained throughout the entire culture period. The only changes in each arrangement were due to processes of cell proliferation and death, aggregate flattening, cellular migration to and from aggregates, and incorporation of small aggregates into larger ones (see below).
Random organization of aggregates Random arrangement of aggregates was the simplest type of organization found. In this case, the aggregates of different sizes did not form specialized geometric patterns (Fig. 1B; see also Fig. 7). In general, the number of aggregates was large. Small aggregates were usually evenly distributed over the surface of the collagen wells, whereas larger aggregates more often formed separate groups. Aggregates within these groups maintained a random interrelationship. The 'random' arrangement of aggregates dictates the development of irregular interconnections between them. Thus, a reproducible pattern of fiber connection was not found in the system of randomly organized aggregates.
Linear organization of aggregates A completely different aggregate arrangement was discovered in some dissc-
172
A
!1
Fig. 2. Early formation of linear chains of aggregates. Living, unstained preparations. Scale bars , 200/~m. A: dissociated spinal cord; 12 days mouse embryo, 24 h in vitro; bright-field illumination. B: dissociated cerebral cortex: 18 day mouse embryo, 2 days in vitro; lateral illumination.
Fig. 3. Fragment of aggregate chain with well-established fiber bundle connections between aggregates. Living, unstained preparation. Dissociated spinal cord; 12 day mouse embryo; lateral illumination; 7 day in vitro. Scale bar = 100 #m.
174
Fig. 4. Myelinated nerve fibers formed in a reaggregated culture of dissociated spinal cord cells. Living, unstained preparations. Fourteen day mouse embryo, 18 days in vitro; bright-field illumination. Scale bar -- 20/~m. A: numerous myelinated fibers in the central area of an aggregate coursing in all directions at different planes of focus. Some of the fibers run above the cell body of neuron, which is below the focal plane (arrow). B: myelinated fiber going through a massive neuritic bridge, which connects adjacent aggregates.
ciated cultures of embryonic spinal cord (12-14-day embryos) and cerebral cortex (16-18-day embryos). Within the first day of cultivation, aggregates were aligned in a linear manner, often for large distances over the surface of the coverslip (Fig. 2A). In extreme cases, more typical in dissociated spinal cord cultures, most dissociated cells were incorporated into aggregates and the latter formed one long multi-aggregate chain (see, for example, Fig. 5A). The number of aggregates outside the chain was very small when the above phenomenon occurred. Often distinct chains were shorter in length and were interwoven amongst the randomly dispersed aggregates. Initial fiber outgrowth, clearly seen during the second day of cultivation (and probably glial in nature), was not oriented specifically in relation to the geometric configuration of the chains, and fibers radiated in all directions from the aggregates (Figs. 1C, D and 2B). However, even at this early stage of chain development, fiber connections in some areas of the chain tended to be predominantly located between adjacent aggregates (Fig. 2B, top). In later stages, these connections often formed solid, compact bundles (Fig. 3). After 2 weeks in vitro, myelination of axons within some aggregates of spinal cord dissociated cells was observed to begin (Fig. 4A). Later,
1~3 the process of myelination was extended to include axons situated in the fiber bundles which connected the aggregates (Fig. 4B). In the living cultures, it was difficult to distinguish between glial and neuronal components of fiber outgrowth, and especially of compact fiber bundles formed between aggregates. When silver impregnation techniques were used, however, axons connecting adjacent aggregates of the chains were clearly observed (Fig. 5A-D). Individual nerve fibers and nerve cell bodies, located in the dense aggregate centers, were not clearly distinguishable, but in small, flattened aggregates where the concentration of cells and fibers was lower, neurons were distinctly visible (Fig. 5E). As can be seen in silver-impregnated preparations (Fig. 5B-D), fiber connections between adjacent aggregates were formed by compact bundles ('neuritic bridges') consisting of numerous neurites (Fig. 5B, C) or by diffuse nets of single axons (Fig. 5D). When silver-impregnated preparations were observed under the phase contrast microscope, it became apparent that most neuritic bridges contained many non-impregnated fibers, probably glial in nature. It was also found that, in some cases, fiber bundles visible under phase contrast illumination were completely devoid of neurites. During the first days of cultivation, active cellular migration was observed. As development in culture continued, migration became more systematic (i.e. less random), and it is interesting to note that in many cases, cellular migration from the margins of the organized, compact aggregates was minimal. On the other hand, migration along the fiber bundles was very active (Fig. 6A; see also Fig. 1D). As can be seen in Figs. 3 and 6A, the spreading of aggregates was not generally typical of organized aggregates attached to a collagen substrate. Their hemisphere-like appearance was preserved throughout the initial period of cultivation up to at least 2 weeks in vitro (Fig. 6A). Occasionally some aggregates spread to a greater degree than others during the first week of cultivation and formed a monolayer of cells over the collagen (Fig. 6B). These flattened aggregates were mostly composed of glial cells, as was observed in living and silver-impregnated preparations. The changes in aggregate shape and size were found to be largely the result of glial cell proliferation and migration, and active fiber outgrowth. The constant dynamic alterations of aggregates were accompanied by changes with respect to the relationship between individual aggregates of the chain. An example of the confluence between some aggregates as a result of glial proliferation and migration is shown in the photograph of a silver-impregnated aggregate chain in Fig. 5A. Another factor responsible for confluence between aggregates was the active migration of entire aggregates. This phenomenon can be seen in the series of time-lapse photographs of randomly oriented aggregates of dissociated cerebral cortex cells after 3 days in vitro (Fig. 7A-C). It is clearly illustrated that after 6 h a small aggregate had become a part of the adjacent large aggregate (Fig. 7A-C, arrows). Another small aggregate, followed by fibers, was displaced toward the gap between two large, converging aggregates (Fig. 7A-C, arrowheads. Note the decrease in distance between the two large aggregates at the top of the photographs).
177
Fig. 6. Different types of aggregates formed from dissociated cells of cerebral cortex. Living, unstained preparations; 18 day mouse embryo. Bright-field illumination. Scale bars = 50 ktm. A: compact aggregate with its coarse fiber bundles. Glial migration along the fiber bundles has begun (arrows); 7 days in vitro. B: flattened aggregates without organized fiber bundles. This type of aggregate probably contained primarily glial cells; 7 days in v:tro. DISCUSSION The d i s s o c i a t i o n - r e a g g r e g a t i o n technique has been used by m a n y investigators to study the m a i n events which occur d u r i n g histogenesis o f the nervous system. Questions r e g a r d i n g the role o f cell m e m b r a n e s in cell-cell recognition, sorting out, m u t u a l cell affinities a n d histotypic association have been analyzed 10-12. It was shown t h a t s u s p e n d e d cells, u p o n m a k i n g contacts, f o r m small tissue-like fragments, so-called aggregates. Briefly, the d a t a concerning the d e v e l o p m e n t o f newly f o r m e d aggregates o f central nervous tissue in g y r a t o r y flasks 9-12,19,27,29,34 can be s u m m a r i z e d according to the following events: (1) f o r m a t i o n o f initial aggregates, c o m p o s e d o f loosely
Fig. 5. Aggregate chain in silver-impregnated preparations. A: part of the exceptionally long chain of aggregates, formed in culture of dissociated cells from 12-day mouse embryonic spinal cord, 11 days in vitro. Note that some areas of the chain composed of more widely separated aggregates connected by neuritic bridges, whereas the central part shows a confluency between many previously isolated aggregates. Scale bar = 1000 #m. B-D: different types of neuritic interconnections between aggregates shown in A. Scale bars = 100/~m. B, compact neuritic bridge; C, compact neuritic bridge and single nerve fibers connecting aggregates; D, diffuse connections formed by isolated nerve fibers. E: neuron in small flattened aggregate from the same preparation as A. Scale bar = 50/~m.
178
D
Fig. 7. Time-lapse frames of reaggregate culture of dissociated cerebral cortex cells illustrating movements of aggregates. Living, unstained preparation; 18-day-old mouse embryo, 3 days in vitro. Bright-field illumination. Scale bar 50/ma. A: 0 h. B : ~ 3 h. C : ~ 6 h. A C, arrows: movement and confluence of aggregates; arrowheads: displacement of a small aggregate, followed by fibers. D: schematic drawing of A and C frames. Dots represent 0 h and solid lines represent position of aggregates after 6 h of observation. Note the changes in size and position of aggregates and resulting confluence between some. Arrows indicate direction of aggregate moverr.ent. p a c k e d a n d r a n d o m l y dispersed cells; (2) m i g r a t i o n a n d segregation o f cells within the aggregates which result in (3) histotypic pattern f o r m a t i o n , often characteristic o f the b r a i n structures sampled (e.g. h i p p o c a m p u s , bulbus olfactorius, etc.). The histotypic d e v e l o p m e n t o f aggregates is a c c o m p a n i e d by biochemical differentiation o f cells t6,26, synaptogenesis 1,28,3°,a2,34 a n d myelin f o r m a t i o n ~,28. The G o l g i i m p r e g n a t i o n technique, used in the study o f aggregates, d e m o n s t r a t e s the high level o f n e u r o n a l differentiation a n d process f o r m a t i o n 24. The a b o v e - m e n t i o n e d events were f o u n d to occur in s p o n t a n e o u s l y f o r m e d aggregates cultured on fiat substrates in M a x i m o w c h a m b e r assembliesS, 7, L e i g h t o n tubes 14 a n d microwell c h a m b e r s 35. However, in such
179 two-dimensional systems, histotypic pattern formations characteristic of the sampled brain structures were not achieved to the same extent as they were in gyratory reaggregate systems (see, for example, ref. 35). At the same time it was shown that, in Maximow chamber cultures of dissociated embryonic chick tectum opticum, separate cells formed reproducible, linearly organized networks3L The present study revealed that the main events involved in the formation and organization of aggregates, their attachment to the collagen substrate, and development of fiber connections between aggregates, were similar to those described in earlier papers which dealt with reaggregation of trypsin-dissociated CNS cells in stationary culturesS,7,14,35. The most interesting observation not found to date, was the formation of linear aggregate chains during the cultivation of dissociated cells in collagen wells. It must be emphasized that, in cultures of dissociated spinal cord cells, linear aggregate chain formation occurred more often than in cultures of dissociated cerebral cortex, where the appearance of a random pattern of aggregate arrangement was more common. However, in parts of these cultures linear chains dispersed among the random aggregate systems were observed. The formation of these shorter aggregate chains may have been the result of any of three factors: first, the immediate linear arrangement of some aggregates; second, subsequent confluency between only some of the adjacent aggregates; and third, later incorporation of small aggregates into larger ones as a result of active aggregate migration. The fact of active migration of entire aggregates observed in these cultures was surprising, although the migration of an array of dorsal root ganglion neurons away from spinal cord explant has been describedS,25. Factors involved in the initial stages of aggregate chain formation in collagenwell cultures are not clear. These factors, biological or physical, have yet to be elucidated. It is possible that one of the factors involved in the formation of linear aggregate chains is the physical properties of collagen substrate in collagen wells. In the same regard, however, aggregate chain formation appears to be dependent on the age of the embryo, i.e. the stage of nerve and glial cell development, and the structure of the CNS from which the cells are taken. For example, aggregate chains were most often seen in the dissociated spinal cord cultures of 12-14-day-old mouse embryos. Thus, formation of aggregate chains with linear fiber interconnections probably reflects the age-dependent histogenetic potentials of cells from different brain structures. It is generally accepted that aggregates are formed as complexes of undifferentiated nerve and glial cellslS,34. In the present study the ratio of these two cell types appeared to be very important for the development of aggregates and fiber connections between them. The absence or death of neurons in newly formed aggregates resulted in the instability of aggregates, their excessive spreading, and even their occasional disappearance and formation of glial cell monolayer. In contrast, the compact aggregates which remained stable throughout the culture period and which were active in the formation of organized fiber interconnections, were always composed of a proportionally large number of neurons (as revealed by silver impregnation studies). Thus, the presence of neurons appeared to be the stabilizing factor in the organization and development of aggregates,
180 Formation of fiber connections also played an important role in the structural stabilization of aggregates and in the consolidation of aggregate chains. Initial fiber outgrowth was mainly of a glial nature and during the early stages of culture development was oriented radially from the aggregates. When the growth of neurites began, however, fiber bundles showed growth directed to neighboring aggregates. Often the fiber bundles growing out of aggregates had leading glial cells located at their tips j 3,15. In mature aggregate chains, neuritic bridges were represented by compact fiber bundles, which occasionally contained myelinated axons, or by a diffuse net of nervous fibers, so that the entire chain could be described as a system of synaptically interconnected neuronal centers. The latter speculation is based on data concerning synaptogenesis in reaggregate cultures available in the literaturel,'5,7,28, 3° -3'~,34,~,5. In conclusion, the newly developed collagen-well coverslip technique described in this paper thus offers many advantages for use in the study of histogenetic processes involved in the development of the CNS, especially with regard to the factors which govern the ensemble formation of neuronal centers. In effect, the technique is a variant of the basic Maximow double coverslip culture system and therefore takes advantage of the favorable features which the latter culture system offers4,22, ~3, in particular for dissociated-reaggregated cultures 5,7,aa. The main difference between the reaggregation patterns of dissociated CNS cells cultures in a Maximow assembly on coverslips coated with reconstituted collagen according to the technique of Bornstein 3, and the technique described in the present paper, is the formation in collagen wells of linear systems of aggregates interconnected with the neuritic bridges. As a result of the present study, it has also been shown that the collagen-well technique, similarly to the microwell technique ~5, offers the following advantages when used for the cultivation of dissociated and reaggregated cells. The concave nature of the well promotes the spontaneous reaggregation of free-floating dissociated cells. In addition, relatively small amounts of tissue are required, little time is needed for the cells to attain effective survival conditions, culture parameters are optimal, and a small volume of cell suspension (1-2 drops) is required, making a large number of experiments economically feasible. In contrast to the microwell technique, however, the collagen wells have a much greater surface area upon which the systems of aggregates can grow. Therefore, the interrelationships between aggregates and the pattern of reaggregate organization over the surface of the coverslip may be studied more effectively.
REFERENCES 1 Bird, M. and James, D. W., The development of synaptic profiles between previously dissociated spinal cord cells in vitro, Z. Zellforsch., 140 (1973) 202-216. 2 Bird, M. and James D. W., Myelin formation in cultures of previously dissociated mouse spinal cord, Cell Tiss. Res., 162 (1975) 93-106. 3 Bornstein, M. B., Reconstituted rat-tail collagen used as substrate for tissue cultures on coverslips in Maximow slides and roller tubes, Lab. Invest., 7 (1958) 134-137. 4 Bornstein, M. B., Organotypic mammalian central and peripheral nerve tissue, lnP. F. Kruseand M. K. Peterson (Eds.), Tissue Culture Methods and Applications, Academic Press, New York, 1973, pp. 86-92.
181 5 Bornstein, M. B. and Model, P. G., Development of synapses and myelin in cultures of dissociated embryonic mouse spinal cord, medulla and cerebrum, Brain Research, 37 (1972) 287-293. 6 Bornstein, M. B. and Murray, M. R., Serial observations on patterns of growth, myelin formation, maintenance and degeneration in cultures of new-born rat and kitten cerebellum, J. Biophys. Biochem. Cytol., 5 (1958) 499-504. 7 Crain, S. M. and Bornstein, M. B., Organotypic bioelectric activity in cultured reaggregates of dissociated rodent brain cells, Science, 176 (1972) 182-184. 8 Crain, S. M. and Peterson, E. R., Enhanced afferent synaptic functions in fetal mouse spinal cordsensory ganglion explants following NGF-induced ganglion hypertrophy, Brain Research, 79 (1974) 145-152. 9 DeLong, G. R., Histogenesis of fetal mouse isocortex and hippocampus in reaggregating cell cultures, Develop. Biol., 22 (1970) 563-583. 10 Garber, B. B., Cell aggregation and recognition in the self-assembly of brain tissues. In S. Fedoroff and L. Hertz (Eds.), Cell, Tissue and Organ Cultures in Neurobiology, Academic Press, New York, 1977, pp. 515-537. 11 Garber, B. B. and Moscona, A. A., Reconstruction of brain tissue from cell suspensions. I. Aggregation patterns of cells dissociated from different regions of the developing brain, Develop. Biol., 27 (1972) 217-234. 12 Garber, B. B. and Moscona, A. A., Reconstruction of brain tissue from cell suspensions. II. Specific enhancement of aggregation of embryonic cerebral cells by supernatant from homologous cell cultures, Develop. Biol., 27 (1972) 235-243. 13 Grainger, F., James, D. W. and Tressman, R. L., An electron microscopic study of the early outgrowth from chick spinal cord in vitro, Z. Zellforsch., 90 (1968) 53-67. 14 Greenham, L. W., Hill, T. J., Moss, C. A. and Peacock, D. B., Preparation of disaggregated mouse brain cultures and observations on early neurite and axon formation therein, Exp. Cell Res., 84 (1974) 287-299. 15 Guillery, R. W., Sobkowicz, H. M. and Scott, G. L., Relationships between glial and neuronal elements in the development of long term cultures of the spinal cord of the fetal mouse, J. comp. Neurol., 140 (1970) 1-34. 16 Honegger, P. and Richelson, E., Biochemical differentiation of mechanically dissociated mammalian brain in aggregating cell cultures, Brain Research, 109 (1976) 335-354. 17 Kim, S. U., Observations on cerebellar granule cells in tissue culture. A silver and electron microscopic study, Z. Zellforsch., 107 (1970) 454-456. 18 Kozak, L. P., Eppig, J. J., Dahl, D. and Bignami, A., Ultrastructural and immunohistochemical characterization of a cell culture model for the study of neuronal-glial interactions, Develop. Biol., 59 (1977) 206-227. 19 Levitt, P., Moore, R. Y. and Garber, B. B., Selective cell association of catecholamine neurons in brain aggregates in vitro, Brain Research, 111 (1976) 311-320. 20 Moscona, A. A., Rotation-mediated histogenetic aggregation of dissociated cells: a quantifiable approach to cell interactions in vitro. Exp. Cell Res., 22 (1961) 455-475. 21 Moscona, A. A., Recombination of dissociated cells and the development of cell aggregates. In E. N. Willmer (Ed.), Cells and Tissues in Culture, Vol. 1, Academic Press, New York, 1965, pp. 489529 22 Murray, M. R., Nervous tissues in vitro. In E. N. Willmer (Ed.), Cells and Tissues in Culture, Vol.2, Academic Press, New York, 1965, pp. 373-455. 23 Murray, M. R., Nervous tissues isolated in culture. In A. Lajtha (Ed.), Handbook of Neurochemistry, Iiol. 5, Part ,4, Plenum Press, New York, 1971, pp. 373-438. 24 Palacios, E. B., Garber, B. B. and Larramendi, L. M. H., Silver chromate impregnation of chick embryo cell aggregates, Brain Research, 66 (1974) 173-178. 25 Peterson, E. R., Crain, S. M. and Murray, M. R., Differentiation and prolonged maintenance of bioelectrically active spinal cord cultures (rat, chick and human), Z. Zellforsch., 66 (1965) 130-154. 26 Seeds, N. W., Biochemical differentiation in reaggregating brain cell cultures, Proc. nat. ,4cad. Sci. (Wash.), 68 (1971) 1858-1861. 27 Seeds, N. W., Differentiation of aggregating brain cell cultures. In G. Sato (Ed.), Tissue Culture of the Nervous System, Plenum Press, New York, 1973, pp. 35-53. 28 Seeds, N. W. and Vatter, A. E., Synaptogenesis in reaggregating brain cell culture, Proc. nat. ,4cad. Sci. (Wash.), 68 (1971) 3219-3222. 29 Sheffield, J. B. and Moscona, A. A., Early stages in the reaggregation of embryonic chick neural retina cells, Exp. Cell Res., 57 (1969) 462-466.
182 30 Stefanelli, A., Cataldi, E. and leradi, L. A., Specific synaptic systems in reaggregated spherules from dissociated chick cerebellum cultivated in vitro, Cell Tiss. Res., 182 (1977) 311-325. 31 Stefanelli, A., Zacchei, A. M., Caravita, S., Cataldi, A. and leradi, L.A., New-forming retinal synapses in vitro, Experientia (Basel), 23 (1967) 199. 32 Suburo, A. M. and Adler, R., Neuronal and glial differentiation in reaggregation cultures, Cell Tiss. Res., 176 (1977) 407-416. 33 Svanidze, I. K., Didimova, E. V., Aggregation of dissociated cells in cultured chicken embryo midbrain, Cytologia (USSR), 21 (1979)9~93. 34 Trapp, B. D., Honegger, P., Richelson, E. and Webster, H. deF., Morphological differentiation of mechanically dissociated fetal rat brain in aggregating cell cultures, Brain Research, 160 (1979) 117-130. 35 Trenkner, E. and Sidman, R. L., Histogenesis of mouse cerebellum in microwell cultures. Cell reaggregation and migration, fiber and synapse formation, J. Cell Biol., 75 (1977) 915-940. 36 Wolf, M. K., Differentiation of neuronal types and synapses in myelinating cultures of mouse cerebellum. J. Cell Biol., 22 (1964) 259-279.
167
© Elsevier/North-Holland Biomedical Press
P A T T E R N S OF R E A G G R E G A T I O N A N D F O R M A T I O N OF L I N E A R A G G R E G A T E C H A I N S I N C O L L A G E N - W E L L C U L T U R E S OF D I S S O C I A T E D M O U S E B R A I N A N D S P I N A L C O R D CELLS
ILYA V. VICTOROV and TERESA L. KRUKOFF* Nervous Tissue Culture Unit, Laboratory of Functional Synaptology, Brain Research Institute, U.S.S.R. Academy of Medical Sciences, Moscow 107120 (U.S.S.R.)
(Accepted April 3rd, 1980) Key words: embryonic mouse brain and spinal cord - - dissociation - - reaggregation - - formation
of aggregate systems - - collagen-well cultures
SUMMARY With the use of the newly developed collagen-well culture technique spontaneous reaggregation of cells from dissociated embryonic mouse brain and spinal cord were studied. Within 20 h in culture, aggregates are formed and settled onto collagen substrate. Two patterns of aggregate arrangement were observed: random and linear. Linear chains of aggregates appeared to be more characteristic of dissociated spinal cord cells, although the linear patterns were not uncommon in cultures of dissociated cortex. Formation of aggregate chains appeared to be dependent on the stage of neuronal and glial differentiation. After attachment to the collagen substrate, the general pattern of aggregate organization was not greatly altered except for changes which resulted from cellular migration and proliferation, the formation of connections between aggregates, or incorporation of small aggregates into larger ones. The number of non-aggregated cells in collagen-well cultures was small. Single, non-aggregated neurons seldom survived individually. Fiber connections between aggregates began to form after the first day in vitro, and by 2 or 3 days, the growing fibers formed neuritic bridges connecting aggregates. By the end of the first week growing fibers often organized compact bundles, but part connections between aggregates were presented by separate fibers in a diffused manner. Silver impregnation revealed that these connections were formed by the axons of neurons located in the aggregates. Thus, * Post-graduatestudent from Department of Anatomy, University of Saskatchewan Medical College, Saskatoon, Sask., Canada, as participant of Canada - - U.S.S.R. graduate student exchange program. Present adress: Health Science Center, Memorial University of Newfoundland, St. John's, Newfoundland, Canada.
168 progression of the above described processes resulted in the 'de novo' formation of linear organized or random systems of interconnected neuronal centers.
INTRODUCTION Cell dissociation and reaggregation techniques have provided a powerful analytic tool for examining the cell-cell interactions involved in the histogenesis of the central nervous system9-12. According to the dissociation and reaggregation procedure developed by Moscona20, 21, dissociated cells are allowed to reaggregate into tissue within gyratory Erlenmeyer flasks where three-dimensional freedom is afforded the cells. Crain and Bornstein 7 planted dissociated fetal mouse brain cells onto collagen-coated coverslips (in Maximow assemblies) and used electrophysiological techniques to study synapse formation between spontaneously formed aggregates. Trenkner and Sidman 35 have studied the reaggregation of mouse cerebellum in microwell cultures. Thus, depending on the specific reaggregation and cultivation methods employed, different aspects of nervous tissue development can be accurately studied. This paper describes a collagen-well culture technique that was used to study spontaneous reaggregation of dissociated cells from embryonic mouse cortex and spinal cord. Results obtained in the present study indicated that the dissociated cells in the collagen well quickly reorganized into aggregates before attaching to the substrate. An unexpected pattern of linear aggregate organization was discovered, especially in cultures of dissociated spinal cord cells. Cell differentiation, growth of neurites and development of fiber connections between aggregates in spatially organized and random systems are described. Preliminary data indicate that this culture system is applicable to the study of factors involved in spontaneous reaggregation of brain cells, formation of spatially arranged system of aggregates and development of connections between re-established neuronal centers. MATERIAL AND METHODS
Dissociation procedure and culture conditions C57BL/6J mice fetuses were used. Dissociated cells were prepared by the modified method of Moscona20, 21. In brief, small fragments of spinal cord (12-16-dayold embryos) or cerebral cortex (16-20-day-old embryos) were washed in Ca 2÷- and Mg2+-free Tyrode solution (CMF) and next incubated in trypsin solution (0.25 ~ in CMF) for 10-15 min at 37 °C. The trypsin was removed and the tissue fragments were washed in a mixture of equal parts of Eagle's minimal essential medium (Dulbecco's modification, DMEM), fetal calf serum and Simms X7 balanced salt solution (BSS). After one wash in nutrient medium (see below), the tissue was dissociated into single cells in a new portion of the medium by repeated passage through a fine-bore Pasteur pippete. The cell suspension was centrifuged and the supernatant was discarded. The cells were resuspended in nutrient medium. Successful disaggregation and more active
169 reaggregafion were also achieved when complete BSS was used instead of CMF and trypsin was dissolved in DMEM. Finally, one drop of the cell suspension (2.0 x 106-1.5 x 107 cells/ml) was placed on a coverslip that had been collagen-coated according to the technique described below. Each coverslip was then mounted in a Maximow double coverslip assembly and incubated at 35 4- 0.5 °C. Cultures were left undisturbed for 20 h and examined at regular daily intervals thereafter. Nutrient medium changes were made biweekly. Cultures were maintained for various time intervals, up to a maximum of 24 days. The living cultures were observed with brightfield microscopy or lateral illumination 8°, using an inverted tissue culture microscope (LOMO, MBI-13). In the histological studies the modified Holmes silver impregnation method a~ and the method of Bodian 17 were used.
Medium Nutrient medium consisted of human placental serum (20 ~), fetal calf serum (20 ~), Dulbecco's modification of Eagle's MEM (40 ~) and Simms X7 BSS (20 ~); the medium also contained 800 m g ~ glucose, 0.25 U/ml insulin and 10-2 M HEPES buffer.
Collagen-coated coverslips ('collagen wells') A modification of the collagen-coating techniquea was used in the following procedure. Two consistencies of dialyzed collagen were used: one (A) was dialyzed against water for 3--4 days and the other (B) was dialyzed for 1-2 days (collagen-B was therefore more fluid than collagen-A). Round glass coverslips (diameter 22 mm) were used. Using collagen-A, a complete collagen ring was made near the periphery of each coverslip. After exposing the ringed coverslips to ammonia vapor for 1 min, the coverslips were washed thoroughly in distilled water. Thorough washing is critical to avoid premature collagen reconstitution in the following step. After removing excess of water, a drop of collagen-B was placed within each collagen ring and evenly spread to meet all inside edges of the ring. Once again the coverslips were exposed to ammonia vapor, and collagen wells are formed. The collagen well has the concave profile versus the conventional convex profile of a coverslip coated with collagen according to the usual technique 3. The completed coverslips were washed in distilled water ( × 3) and in BSS ( × 3). The collagen-coated coverslips were stored in Columbia dishes containing 9 ml BSS, 0.5 ml fetal calf serum and 0.5 ml 2 0 ~ glucose. RESULTS
Aggregate formation Directly after seeding of the collagen wells, single isolated cells remained suspended in the medium (Fig. 1A). In spite of complete dissociation, cells were often observed to quickly adhere to one another to form small clumps which did not attach to the collagen during the initial period of cultivation. A comparatively large volume of medium allowed free movement of the single cells and cell clumps. Within 16-20 h
170
Fig. 1. Dissociated brain cells and the formation of aggregates in collagen wells during the initial 2-3 days in vitro. Living, unstained preparations. Bright-field illumination. Scale bars : 50 /m~. A: dissociated brain cells immediately after seeding. B: cell aggregates have formed and fixed to the collagen substrate. Only a few cells remain free, 20 h in vitro. C: initial growth of cell processes and formation of primary connections between aggregates, 2 days in vitro. D: flattening of aggregates and formation of fiber bundles. Note the active migration of glial cells along the fiber bundles: 3 days in vitro.
171 in vitro, spontaneous reaggregation of the single cells was near completion (Fig. 1B). In most cases, only a small percentage of cells did not become incorporated into aggregates. It should be stressed that reaggregation always occurred before the cells had attached to the collagen. Most aggregates attached to the substrate within the first day in vitro, and as early as 24-36 h cellular migration and initial outgrowth of cell processes were observed. By 2-3 days, distinct fiber outgrowth was clearly seen (Fig. 1C, D). The first fiber connections between aggregates were observed after the same period of time in vitro. It was difficult to characterize exactly the nature of the fiber outgrowth in living preparations (i.e. neuronal or glial); however, histological studies with silver impregnation showed that the number of nerve fibers in 2 or 3 day cultures was relatively low. The number of non-aggregated neuronal and glial cells decreased with aging of the cultures, either because of single cell incorporation into aggregates, or because of degeneration of some of the single cells (compare, for example, the number of nonaggregated cells in Figs. 2 and 3). Note that single astrocytes did occasionally grow isolated from the aggregates to form a network of glial fibers around the aggregates. As a rule, however, isolated neurons were not observed in mature reaggregate cultures grown in collagen wells.
Pattern formation The most interesting observation of reaggregate cultures in collagen wells was the spatial organization pattern of aggregate systems over the surface of the coverslip. During the first day of cultivation, newly formed aggregates attached to the collagen surface in two main configurations: random and linear. In some cases these two extreme types of arrangements were found in the same culture. It was found that the initial overall type of aggregate arrangement in each culture was generally maintained throughout the entire culture period. The only changes in each arrangement were due to processes of cell proliferation and death, aggregate flattening, cellular migration to and from aggregates, and incorporation of small aggregates into larger ones (see below).
Random organization of aggregates Random arrangement of aggregates was the simplest type of organization found. In this case, the aggregates of different sizes did not form specialized geometric patterns (Fig. 1B; see also Fig. 7). In general, the number of aggregates was large. Small aggregates were usually evenly distributed over the surface of the collagen wells, whereas larger aggregates more often formed separate groups. Aggregates within these groups maintained a random interrelationship. The 'random' arrangement of aggregates dictates the development of irregular interconnections between them. Thus, a reproducible pattern of fiber connection was not found in the system of randomly organized aggregates.
Linear organization of aggregates A completely different aggregate arrangement was discovered in some dissc-
172
A
!1
Fig. 2. Early formation of linear chains of aggregates. Living, unstained preparations. Scale bars , 200/~m. A: dissociated spinal cord; 12 days mouse embryo, 24 h in vitro; bright-field illumination. B: dissociated cerebral cortex: 18 day mouse embryo, 2 days in vitro; lateral illumination.
Fig. 3. Fragment of aggregate chain with well-established fiber bundle connections between aggregates. Living, unstained preparation. Dissociated spinal cord; 12 day mouse embryo; lateral illumination; 7 day in vitro. Scale bar = 100 #m.
174
Fig. 4. Myelinated nerve fibers formed in a reaggregated culture of dissociated spinal cord cells. Living, unstained preparations. Fourteen day mouse embryo, 18 days in vitro; bright-field illumination. Scale bar -- 20/~m. A: numerous myelinated fibers in the central area of an aggregate coursing in all directions at different planes of focus. Some of the fibers run above the cell body of neuron, which is below the focal plane (arrow). B: myelinated fiber going through a massive neuritic bridge, which connects adjacent aggregates.
ciated cultures of embryonic spinal cord (12-14-day embryos) and cerebral cortex (16-18-day embryos). Within the first day of cultivation, aggregates were aligned in a linear manner, often for large distances over the surface of the coverslip (Fig. 2A). In extreme cases, more typical in dissociated spinal cord cultures, most dissociated cells were incorporated into aggregates and the latter formed one long multi-aggregate chain (see, for example, Fig. 5A). The number of aggregates outside the chain was very small when the above phenomenon occurred. Often distinct chains were shorter in length and were interwoven amongst the randomly dispersed aggregates. Initial fiber outgrowth, clearly seen during the second day of cultivation (and probably glial in nature), was not oriented specifically in relation to the geometric configuration of the chains, and fibers radiated in all directions from the aggregates (Figs. 1C, D and 2B). However, even at this early stage of chain development, fiber connections in some areas of the chain tended to be predominantly located between adjacent aggregates (Fig. 2B, top). In later stages, these connections often formed solid, compact bundles (Fig. 3). After 2 weeks in vitro, myelination of axons within some aggregates of spinal cord dissociated cells was observed to begin (Fig. 4A). Later,
1~3 the process of myelination was extended to include axons situated in the fiber bundles which connected the aggregates (Fig. 4B). In the living cultures, it was difficult to distinguish between glial and neuronal components of fiber outgrowth, and especially of compact fiber bundles formed between aggregates. When silver impregnation techniques were used, however, axons connecting adjacent aggregates of the chains were clearly observed (Fig. 5A-D). Individual nerve fibers and nerve cell bodies, located in the dense aggregate centers, were not clearly distinguishable, but in small, flattened aggregates where the concentration of cells and fibers was lower, neurons were distinctly visible (Fig. 5E). As can be seen in silver-impregnated preparations (Fig. 5B-D), fiber connections between adjacent aggregates were formed by compact bundles ('neuritic bridges') consisting of numerous neurites (Fig. 5B, C) or by diffuse nets of single axons (Fig. 5D). When silver-impregnated preparations were observed under the phase contrast microscope, it became apparent that most neuritic bridges contained many non-impregnated fibers, probably glial in nature. It was also found that, in some cases, fiber bundles visible under phase contrast illumination were completely devoid of neurites. During the first days of cultivation, active cellular migration was observed. As development in culture continued, migration became more systematic (i.e. less random), and it is interesting to note that in many cases, cellular migration from the margins of the organized, compact aggregates was minimal. On the other hand, migration along the fiber bundles was very active (Fig. 6A; see also Fig. 1D). As can be seen in Figs. 3 and 6A, the spreading of aggregates was not generally typical of organized aggregates attached to a collagen substrate. Their hemisphere-like appearance was preserved throughout the initial period of cultivation up to at least 2 weeks in vitro (Fig. 6A). Occasionally some aggregates spread to a greater degree than others during the first week of cultivation and formed a monolayer of cells over the collagen (Fig. 6B). These flattened aggregates were mostly composed of glial cells, as was observed in living and silver-impregnated preparations. The changes in aggregate shape and size were found to be largely the result of glial cell proliferation and migration, and active fiber outgrowth. The constant dynamic alterations of aggregates were accompanied by changes with respect to the relationship between individual aggregates of the chain. An example of the confluence between some aggregates as a result of glial proliferation and migration is shown in the photograph of a silver-impregnated aggregate chain in Fig. 5A. Another factor responsible for confluence between aggregates was the active migration of entire aggregates. This phenomenon can be seen in the series of time-lapse photographs of randomly oriented aggregates of dissociated cerebral cortex cells after 3 days in vitro (Fig. 7A-C). It is clearly illustrated that after 6 h a small aggregate had become a part of the adjacent large aggregate (Fig. 7A-C, arrows). Another small aggregate, followed by fibers, was displaced toward the gap between two large, converging aggregates (Fig. 7A-C, arrowheads. Note the decrease in distance between the two large aggregates at the top of the photographs).
177
Fig. 6. Different types of aggregates formed from dissociated cells of cerebral cortex. Living, unstained preparations; 18 day mouse embryo. Bright-field illumination. Scale bars = 50 ktm. A: compact aggregate with its coarse fiber bundles. Glial migration along the fiber bundles has begun (arrows); 7 days in vitro. B: flattened aggregates without organized fiber bundles. This type of aggregate probably contained primarily glial cells; 7 days in v:tro. DISCUSSION The d i s s o c i a t i o n - r e a g g r e g a t i o n technique has been used by m a n y investigators to study the m a i n events which occur d u r i n g histogenesis o f the nervous system. Questions r e g a r d i n g the role o f cell m e m b r a n e s in cell-cell recognition, sorting out, m u t u a l cell affinities a n d histotypic association have been analyzed 10-12. It was shown t h a t s u s p e n d e d cells, u p o n m a k i n g contacts, f o r m small tissue-like fragments, so-called aggregates. Briefly, the d a t a concerning the d e v e l o p m e n t o f newly f o r m e d aggregates o f central nervous tissue in g y r a t o r y flasks 9-12,19,27,29,34 can be s u m m a r i z e d according to the following events: (1) f o r m a t i o n o f initial aggregates, c o m p o s e d o f loosely
Fig. 5. Aggregate chain in silver-impregnated preparations. A: part of the exceptionally long chain of aggregates, formed in culture of dissociated cells from 12-day mouse embryonic spinal cord, 11 days in vitro. Note that some areas of the chain composed of more widely separated aggregates connected by neuritic bridges, whereas the central part shows a confluency between many previously isolated aggregates. Scale bar = 1000 #m. B-D: different types of neuritic interconnections between aggregates shown in A. Scale bars = 100/~m. B, compact neuritic bridge; C, compact neuritic bridge and single nerve fibers connecting aggregates; D, diffuse connections formed by isolated nerve fibers. E: neuron in small flattened aggregate from the same preparation as A. Scale bar = 50/~m.
178
D
Fig. 7. Time-lapse frames of reaggregate culture of dissociated cerebral cortex cells illustrating movements of aggregates. Living, unstained preparation; 18-day-old mouse embryo, 3 days in vitro. Bright-field illumination. Scale bar 50/ma. A: 0 h. B : ~ 3 h. C : ~ 6 h. A C, arrows: movement and confluence of aggregates; arrowheads: displacement of a small aggregate, followed by fibers. D: schematic drawing of A and C frames. Dots represent 0 h and solid lines represent position of aggregates after 6 h of observation. Note the changes in size and position of aggregates and resulting confluence between some. Arrows indicate direction of aggregate moverr.ent. p a c k e d a n d r a n d o m l y dispersed cells; (2) m i g r a t i o n a n d segregation o f cells within the aggregates which result in (3) histotypic pattern f o r m a t i o n , often characteristic o f the b r a i n structures sampled (e.g. h i p p o c a m p u s , bulbus olfactorius, etc.). The histotypic d e v e l o p m e n t o f aggregates is a c c o m p a n i e d by biochemical differentiation o f cells t6,26, synaptogenesis 1,28,3°,a2,34 a n d myelin f o r m a t i o n ~,28. The G o l g i i m p r e g n a t i o n technique, used in the study o f aggregates, d e m o n s t r a t e s the high level o f n e u r o n a l differentiation a n d process f o r m a t i o n 24. The a b o v e - m e n t i o n e d events were f o u n d to occur in s p o n t a n e o u s l y f o r m e d aggregates cultured on fiat substrates in M a x i m o w c h a m b e r assembliesS, 7, L e i g h t o n tubes 14 a n d microwell c h a m b e r s 35. However, in such
179 two-dimensional systems, histotypic pattern formations characteristic of the sampled brain structures were not achieved to the same extent as they were in gyratory reaggregate systems (see, for example, ref. 35). At the same time it was shown that, in Maximow chamber cultures of dissociated embryonic chick tectum opticum, separate cells formed reproducible, linearly organized networks3L The present study revealed that the main events involved in the formation and organization of aggregates, their attachment to the collagen substrate, and development of fiber connections between aggregates, were similar to those described in earlier papers which dealt with reaggregation of trypsin-dissociated CNS cells in stationary culturesS,7,14,35. The most interesting observation not found to date, was the formation of linear aggregate chains during the cultivation of dissociated cells in collagen wells. It must be emphasized that, in cultures of dissociated spinal cord cells, linear aggregate chain formation occurred more often than in cultures of dissociated cerebral cortex, where the appearance of a random pattern of aggregate arrangement was more common. However, in parts of these cultures linear chains dispersed among the random aggregate systems were observed. The formation of these shorter aggregate chains may have been the result of any of three factors: first, the immediate linear arrangement of some aggregates; second, subsequent confluency between only some of the adjacent aggregates; and third, later incorporation of small aggregates into larger ones as a result of active aggregate migration. The fact of active migration of entire aggregates observed in these cultures was surprising, although the migration of an array of dorsal root ganglion neurons away from spinal cord explant has been describedS,25. Factors involved in the initial stages of aggregate chain formation in collagenwell cultures are not clear. These factors, biological or physical, have yet to be elucidated. It is possible that one of the factors involved in the formation of linear aggregate chains is the physical properties of collagen substrate in collagen wells. In the same regard, however, aggregate chain formation appears to be dependent on the age of the embryo, i.e. the stage of nerve and glial cell development, and the structure of the CNS from which the cells are taken. For example, aggregate chains were most often seen in the dissociated spinal cord cultures of 12-14-day-old mouse embryos. Thus, formation of aggregate chains with linear fiber interconnections probably reflects the age-dependent histogenetic potentials of cells from different brain structures. It is generally accepted that aggregates are formed as complexes of undifferentiated nerve and glial cellslS,34. In the present study the ratio of these two cell types appeared to be very important for the development of aggregates and fiber connections between them. The absence or death of neurons in newly formed aggregates resulted in the instability of aggregates, their excessive spreading, and even their occasional disappearance and formation of glial cell monolayer. In contrast, the compact aggregates which remained stable throughout the culture period and which were active in the formation of organized fiber interconnections, were always composed of a proportionally large number of neurons (as revealed by silver impregnation studies). Thus, the presence of neurons appeared to be the stabilizing factor in the organization and development of aggregates,
180 Formation of fiber connections also played an important role in the structural stabilization of aggregates and in the consolidation of aggregate chains. Initial fiber outgrowth was mainly of a glial nature and during the early stages of culture development was oriented radially from the aggregates. When the growth of neurites began, however, fiber bundles showed growth directed to neighboring aggregates. Often the fiber bundles growing out of aggregates had leading glial cells located at their tips j 3,15. In mature aggregate chains, neuritic bridges were represented by compact fiber bundles, which occasionally contained myelinated axons, or by a diffuse net of nervous fibers, so that the entire chain could be described as a system of synaptically interconnected neuronal centers. The latter speculation is based on data concerning synaptogenesis in reaggregate cultures available in the literaturel,'5,7,28, 3° -3'~,34,~,5. In conclusion, the newly developed collagen-well coverslip technique described in this paper thus offers many advantages for use in the study of histogenetic processes involved in the development of the CNS, especially with regard to the factors which govern the ensemble formation of neuronal centers. In effect, the technique is a variant of the basic Maximow double coverslip culture system and therefore takes advantage of the favorable features which the latter culture system offers4,22, ~3, in particular for dissociated-reaggregated cultures 5,7,aa. The main difference between the reaggregation patterns of dissociated CNS cells cultures in a Maximow assembly on coverslips coated with reconstituted collagen according to the technique of Bornstein 3, and the technique described in the present paper, is the formation in collagen wells of linear systems of aggregates interconnected with the neuritic bridges. As a result of the present study, it has also been shown that the collagen-well technique, similarly to the microwell technique ~5, offers the following advantages when used for the cultivation of dissociated and reaggregated cells. The concave nature of the well promotes the spontaneous reaggregation of free-floating dissociated cells. In addition, relatively small amounts of tissue are required, little time is needed for the cells to attain effective survival conditions, culture parameters are optimal, and a small volume of cell suspension (1-2 drops) is required, making a large number of experiments economically feasible. In contrast to the microwell technique, however, the collagen wells have a much greater surface area upon which the systems of aggregates can grow. Therefore, the interrelationships between aggregates and the pattern of reaggregate organization over the surface of the coverslip may be studied more effectively.
REFERENCES 1 Bird, M. and James, D. W., The development of synaptic profiles between previously dissociated spinal cord cells in vitro, Z. Zellforsch., 140 (1973) 202-216. 2 Bird, M. and James D. W., Myelin formation in cultures of previously dissociated mouse spinal cord, Cell Tiss. Res., 162 (1975) 93-106. 3 Bornstein, M. B., Reconstituted rat-tail collagen used as substrate for tissue cultures on coverslips in Maximow slides and roller tubes, Lab. Invest., 7 (1958) 134-137. 4 Bornstein, M. B., Organotypic mammalian central and peripheral nerve tissue, lnP. F. Kruseand M. K. Peterson (Eds.), Tissue Culture Methods and Applications, Academic Press, New York, 1973, pp. 86-92.
181 5 Bornstein, M. B. and Model, P. G., Development of synapses and myelin in cultures of dissociated embryonic mouse spinal cord, medulla and cerebrum, Brain Research, 37 (1972) 287-293. 6 Bornstein, M. B. and Murray, M. R., Serial observations on patterns of growth, myelin formation, maintenance and degeneration in cultures of new-born rat and kitten cerebellum, J. Biophys. Biochem. Cytol., 5 (1958) 499-504. 7 Crain, S. M. and Bornstein, M. B., Organotypic bioelectric activity in cultured reaggregates of dissociated rodent brain cells, Science, 176 (1972) 182-184. 8 Crain, S. M. and Peterson, E. R., Enhanced afferent synaptic functions in fetal mouse spinal cordsensory ganglion explants following NGF-induced ganglion hypertrophy, Brain Research, 79 (1974) 145-152. 9 DeLong, G. R., Histogenesis of fetal mouse isocortex and hippocampus in reaggregating cell cultures, Develop. Biol., 22 (1970) 563-583. 10 Garber, B. B., Cell aggregation and recognition in the self-assembly of brain tissues. In S. Fedoroff and L. Hertz (Eds.), Cell, Tissue and Organ Cultures in Neurobiology, Academic Press, New York, 1977, pp. 515-537. 11 Garber, B. B. and Moscona, A. A., Reconstruction of brain tissue from cell suspensions. I. Aggregation patterns of cells dissociated from different regions of the developing brain, Develop. Biol., 27 (1972) 217-234. 12 Garber, B. B. and Moscona, A. A., Reconstruction of brain tissue from cell suspensions. II. Specific enhancement of aggregation of embryonic cerebral cells by supernatant from homologous cell cultures, Develop. Biol., 27 (1972) 235-243. 13 Grainger, F., James, D. W. and Tressman, R. L., An electron microscopic study of the early outgrowth from chick spinal cord in vitro, Z. Zellforsch., 90 (1968) 53-67. 14 Greenham, L. W., Hill, T. J., Moss, C. A. and Peacock, D. B., Preparation of disaggregated mouse brain cultures and observations on early neurite and axon formation therein, Exp. Cell Res., 84 (1974) 287-299. 15 Guillery, R. W., Sobkowicz, H. M. and Scott, G. L., Relationships between glial and neuronal elements in the development of long term cultures of the spinal cord of the fetal mouse, J. comp. Neurol., 140 (1970) 1-34. 16 Honegger, P. and Richelson, E., Biochemical differentiation of mechanically dissociated mammalian brain in aggregating cell cultures, Brain Research, 109 (1976) 335-354. 17 Kim, S. U., Observations on cerebellar granule cells in tissue culture. A silver and electron microscopic study, Z. Zellforsch., 107 (1970) 454-456. 18 Kozak, L. P., Eppig, J. J., Dahl, D. and Bignami, A., Ultrastructural and immunohistochemical characterization of a cell culture model for the study of neuronal-glial interactions, Develop. Biol., 59 (1977) 206-227. 19 Levitt, P., Moore, R. Y. and Garber, B. B., Selective cell association of catecholamine neurons in brain aggregates in vitro, Brain Research, 111 (1976) 311-320. 20 Moscona, A. A., Rotation-mediated histogenetic aggregation of dissociated cells: a quantifiable approach to cell interactions in vitro. Exp. Cell Res., 22 (1961) 455-475. 21 Moscona, A. A., Recombination of dissociated cells and the development of cell aggregates. In E. N. Willmer (Ed.), Cells and Tissues in Culture, Vol. 1, Academic Press, New York, 1965, pp. 489529 22 Murray, M. R., Nervous tissues in vitro. In E. N. Willmer (Ed.), Cells and Tissues in Culture, Vol.2, Academic Press, New York, 1965, pp. 373-455. 23 Murray, M. R., Nervous tissues isolated in culture. In A. Lajtha (Ed.), Handbook of Neurochemistry, Iiol. 5, Part ,4, Plenum Press, New York, 1971, pp. 373-438. 24 Palacios, E. B., Garber, B. B. and Larramendi, L. M. H., Silver chromate impregnation of chick embryo cell aggregates, Brain Research, 66 (1974) 173-178. 25 Peterson, E. R., Crain, S. M. and Murray, M. R., Differentiation and prolonged maintenance of bioelectrically active spinal cord cultures (rat, chick and human), Z. Zellforsch., 66 (1965) 130-154. 26 Seeds, N. W., Biochemical differentiation in reaggregating brain cell cultures, Proc. nat. ,4cad. Sci. (Wash.), 68 (1971) 1858-1861. 27 Seeds, N. W., Differentiation of aggregating brain cell cultures. In G. Sato (Ed.), Tissue Culture of the Nervous System, Plenum Press, New York, 1973, pp. 35-53. 28 Seeds, N. W. and Vatter, A. E., Synaptogenesis in reaggregating brain cell culture, Proc. nat. ,4cad. Sci. (Wash.), 68 (1971) 3219-3222. 29 Sheffield, J. B. and Moscona, A. A., Early stages in the reaggregation of embryonic chick neural retina cells, Exp. Cell Res., 57 (1969) 462-466.
182 30 Stefanelli, A., Cataldi, E. and leradi, L. A., Specific synaptic systems in reaggregated spherules from dissociated chick cerebellum cultivated in vitro, Cell Tiss. Res., 182 (1977) 311-325. 31 Stefanelli, A., Zacchei, A. M., Caravita, S., Cataldi, A. and leradi, L.A., New-forming retinal synapses in vitro, Experientia (Basel), 23 (1967) 199. 32 Suburo, A. M. and Adler, R., Neuronal and glial differentiation in reaggregation cultures, Cell Tiss. Res., 176 (1977) 407-416. 33 Svanidze, I. K., Didimova, E. V., Aggregation of dissociated cells in cultured chicken embryo midbrain, Cytologia (USSR), 21 (1979)9~93. 34 Trapp, B. D., Honegger, P., Richelson, E. and Webster, H. deF., Morphological differentiation of mechanically dissociated fetal rat brain in aggregating cell cultures, Brain Research, 160 (1979) 117-130. 35 Trenkner, E. and Sidman, R. L., Histogenesis of mouse cerebellum in microwell cultures. Cell reaggregation and migration, fiber and synapse formation, J. Cell Biol., 75 (1977) 915-940. 36 Wolf, M. K., Differentiation of neuronal types and synapses in myelinating cultures of mouse cerebellum. J. Cell Biol., 22 (1964) 259-279.