Brain Research, 169 (1979) 231-246 .?) Elsevier/North-Holland Biomedical Press
231
M O R P H O L O G Y OF D I S S O C I A T E D H I P P O C A M P A L C U L T U R E S F R O M FETAL MICE
JOHN H. PEACOCK, DAPHNE F. RUSH and LAWRENCE H. MATHERS Department of Neurology, Stan/brd University Medical Center, Stanford, Cali/i 94305 (U.S.A.)
(Accepted October 5th, 1978)
SUMMARY Dissociated hippocampal cultures from fetal mice (13-18 days gestational age) can be maintained for up to two months in culture. Cells grow as either isolated neurons or in small neuronal aggregates. Neurons remain small with a soma diameter of 15-20 ,urn even in mature cultures and develop extensively branched processes during the first two weeks in culture. After this time, processes become more difficult to visualize with phase-contrast optics because of a tendency to grow within the underlying non-neuronal cells. However, the presence of processes has been proved by silver-staining which demonstrates an organizational complexity ranging from a loosely reticulated neuropil to fascicles containing many fibers. More detailed study of individual neuronal morphology was carried out in cells filled with the fluorescent dye, Lucifer Yellow CH, in conjunction with the intracellular recording of synaptic and action potentials from dye-containing micropipettes. Dye-filled cells show a well-developed branching morphology. Process specializations include spines, beading, and basket-like endings. Processes tend to emanate from one side of the soma, either originating at the cell body or from a single trunk. Commonly there are 2-4 orders of branching, but up to 6 orders can occur (counted centrifugally from the soma). Electron microscopy revealed synapses distributed predominantly on dendrites with a smaller number on somata. Dendritic spines are present and are contacted principally by asymmetric synaptic junctions. Symmetric synapses are relatively more common on somata and proximal dendrites.
INTRODUCTION
Differentiated properties of hippocampal neurons develop or are preserved to a remarkable extent in organotypic cultures from both morphologicga ° and intracellular electrophysiologic t9 standpoints, although the latter has been studied less extensively.
232 Further electrophysiotogic analysis of cultured hippocampal neurons is of considerable interest and would be greatly facilitated if hippocampal neurons could be grown in a dissociated, surface culture system where cells are visually accessible for intracellular penetration. This has been accomplished for short-term Cultures (up to two weeks) by Banker and Cowan ~ who give evidence that the development of neurons in dissociated hippocampal cell cultures is consistent with that of pyramidal cells in the intact animal. We have developed a long-term (up to two months) dissociated hippocampal cell culture system for electrophysiology and in this report describe the major morphologic features of these cultures. In addition we show single neurons whose dendritic morphology has been visually isolated from surrounding cells after intracellular fluorescent dye-staining with the new fluorescent probe Lucifer Yellow CH is and, from which, spontaneous electrical activity has been recorded while staining is visually monitored. MATERIALS AND METHODS
Dissection, dissociation and culture procedures Whole brains from 4 to 6 Swiss Webster mouse fetuses, 13-18 days gestationaf age, were removed under sterile conditions and placed in chilled Minimal Essential Eagle's Medium (MEM) containing 6.0 g/liter glucose and 12.5 m M Hepes buffer. Each hemisphere was separated from the thalamus, the hippocampal fissure was identified by the vascular pattern (Fig. I A, left hemisphere, arrows) and the meninges removed in older fetuses (16--18 days in utero). The hippocampus then appeared as a dark band when viewed under oblique transillumination. With iridectomy scissors, a transversely oriented incision was made into the rostral and caudal parts of the hippocampus (Fig. IB) and the hippocampus was cut away from the hemisphere along its
Fig. 1. Photomicrograph of medial surfaces of the left and right cerebral hemispheres. A : the meninges have been left intact to demonstrate the vessels which mark the hippocampal fissure (arrowsJ. B'. the meninges have been stripped away and two incisions (arrows}have been made through thehippoeampus which now appears as a dark band of tissue. Fetal gestational age was 16 days.
233 longitudinal axis. The dissected pieces of hippocampus were collected in a 35 mm collagen-coated plastic dish (Falcon 3001) which contained 1 ml of growth medium (see below) at 20 °C. After 6-12 hippocampi had been pooled, the tissue was aspirated twice using a I ml plastic syringe and a 27-gauge needle. Cells were distributed from the syringe into 35 mm collagen-coated dishes at a density of 0.5-1 x 106 cells per dish (0.5-1.0 ~,i 10~ ceIls/sq, cm) in a volume of 1.5 ml medium. Collagen was prepared by suspending 150 mg of calf skin collagen (Worthington) in 100 ml double distilled water and autoclaving to dissolve. The suspension was filtered through No. 2 Whatman paper and sterilized by reautoclaving the clarified collagen. Growth medium consisted of MEM or Dulbecco's Modified Eagle's Medium (DMEM) added in equal volume to conditioned medium from spinal cord cultures. Conditioned medium was MEM or DM EM containing 10 ~ fetal calf serum and 10 ~ horse serum removed every 3--4 days from mouse spinal cord cell cultures at least 14 days old. Growth medium, with a final serum concentration of 5 ~ each for fetal calf serum and horse serum, was then filtered through a 0.2/zm Millipore filter, and stored at 4 °C until use. The first medium change was made after 4 days in culture and subsequent changes every 7 days. In some cases 10-20/zg/ml 5-fluoro-2'-deoxyuridine (FdU) and 25-50/zg/ml uridine were added during days 7-I 1 ofculture~L No antibiotics were used.
Histology and electron microscopy Silver-staining of cultures was carried out according to a modification of the procedure of Sevier and Munger 17 after fixation in 2 ~ cacodylate in buffered formalin (4"/~ 16- or 17-day / O J" For light microscopy, the hippocampus was dissected from fetuses, fixed in 2 IV,,glutaraldehyde, postfixed in 1 ~ OsO4, dehydrated, embedded in Spurr, sectioned at I /zm thickness, and stained with hot 1 o/methylene blue in I ~i sodium borate buffer. For electron microscopy, cultures were fixed in 2 ~ glutaraldehyde in 0.1 M Sorensen's phosphate buffer plus 0.1 M sucrose for a final osmolarity of 392 mOsm at pH 7.3 for at least 24 h at 4 °C, stored in Sorensen's phosphate buffer at 4 °C, postfixed in I ~ OsO4, dehydrated in ethanol, and embedded in Epon. Embedded cultures were removed from the plastic dish and designated areas were sectioned in a plane parallel or perpendicular to the surface of the culture. Sections were stained with lead acetate and in some cases with uranyl acetate.
Electrophysiology and ,fluorescent dye staining Culture dishes were placed in a chamber on the stage of a Zeiss inverted phase contrast microscope where they were kept at 37 °C by heating the air above and below the chamber, at pH 7.2, by flowing 5 ~,, COz in air over the culture, and at constant osmolarity by spreading a layer of light mineral oil on the surface of the medium. Calcium concentration was increased from 1.8 mM to 3.5 raM. A Zeiss epilluminator (IV FL) containing an ultraviolet light source (50 watt DC mercury bulb) was mounted between the body of the microscope and the eyepieces for simultaneous or sequential phase contrast and fluorescent microscopy. For fluorescence, Zeiss excita-
234 tion filters B G 12 and U G 5 (peak transmission of about 395 nm), a 460 nm dichroic reflector, and a 500 nm barrier filter were used. Microelectrodes were pulled from microfiber glass capillary tubes, filled with a 4 ~ aqueous solution of Lucifer Yellow C H and bevelled until tip resistances of 135-175 M R were achieved. After cells were penetrated using phase-contrast optics, the incandescent light source was turned off and neurons were visually monitored using ultraviolet optics until fluorescent dye staining was complete, about 3 5 min. Lucifer Yellow C H is negatively charged and was ejected from the micropipette by applying a steady negative DC current of 0.5-3 nA to the dye solution. A bridge circuit was used to simultaneously monitor stimulating currents and voltage responses from the same microelectrode. Currents and voltages were always recorded on a cathode ray oscilloscope, a chart recorder, and in some cases on FM tape. RESULTS
Morphological appearance of the hippocampus at time of culture The portion of the hippocampus that was removed for culture wa~ compared with the ablated area of the brain from which it was dissected. The dissection included the fimbria and adjacent zones corresponding to CA2, CA3, and CA4 pyramidal cells along with a portion of the developing dentate area. The CA~ pyramidal zone was present to a variable extent and it is a possibility that cells from the subicutum were included in some dissections. Light microscopy observauons suggested that many pyramidal neurons in the zone corresponding to CA3 were postmitotic with their pale staining nuclei, two or more nucleoli, and orientation of large proximal dendrites toward the interior of the hippocampal formation. Darkly staining, and presumably proliferating, pyramidal cells were present along the ventricular zone and. similarly, intensely stained cells were found in the tmmature dentate area.
Fig. 2. Silver-stained neuron in 2,day-01d culture from fetus of 18 days gestational age.
235 Gestational age at time o f culture Cultures were prepared from fetuses of 13-18 days gestational age. Dissection of 13-15-day fetuses presents some difficulty due to reduced contrast in the density of t he cell layers and the fragility of the tissue. It is difficult to remove the meninges completely until fetuses reach 16 days in utero. However, cultures prepared from animals over this fetal age range display a similar morphology at various stages of culture growth. Neurons with a major bifurcating process are frequently found in young cultures from a time shortly after the initial plating (Fig. 2) up to about two weeks as shown in Fig. 3A and B from a 13-day-old fetus, and Fig. 3C and D from 17day-old fetuses. There does not appear to be a difference in terms of neuronal survival or in the rate of successful cultures over this fetal age range. However, cultures from 18-day-old fetuses are subject to greater cell death than cultures from the younger fetuses and, on this empirical basis, the hippocampus from 18-day fetuses was not routinely cultured. The use of conditioned medium from spinal cord cultures contributes to neuronal survival for periods in culture longer than two weeks either when used from the beginning of the culture or to rescue cultures which had not been grown with conditioned medium until deterioration was detected at about two weeks in culture. In these cases the results have been sufficiently dramatic that we routinely use conditioned medium in long-term cultures. However, it should be noted that we have on occasion observed long-term survival without conditioned medium.
Fig. 3. Phase contrast photomicrographs ofhippocampal neurons in week old cultures from fetuses of 14 days gestational age (A and B) and 17 days gestational age (C and D).
236
Fig. 4. Phase contrast photomicrographs from hippocampal cultures after 5 days in culture (A), 10 days in culture (B), and 4 weeks in culture (C). Fetal gestational age 13 days.
Culture development Cultures progress t h r o u g h two m a j o r m o r p h o l o g i c a n d d e v e l o p m e n t a l stages. The first stage is shown in Fig. 4 A and B a n d the second in Fig. 4C: all examples are
237 from the same culture dish. Cells attach to the surface of the culture dish within the first 24 h and extend processes. Frequently but not always they grow upon several contiguous flat non-neuronal cells. After 3 4 days in culture (Fig. 4A) it is evident whether a culture is going to survive for the next two weeks. Over the course of 10 days the cells become more phase bright, reach a relatively uniform size of 15-20/zm, and develop many interlacing processes (Fig. 4B). At this stage non-neuronal background cells cover the surface of this dish. Some of these are ependymal cells which beat rhythmically and are frequently found in association with neurons. After two weeks in culture the processes become difficult to visualize with phase contrast optics (Fig. 4C, left and right hand cell groups) and in some areas of a culture, or even throughout entire cultures, seem to disappear. But hints of the existence of processes are given by occasional thin neuritic bridges which interconnect small groups of cells (Fig. 4C) or short segments of processes which are visible for several cell diameters away from a cell body before merging with background cells. Treatment of the cultures with FdU plus uridine did not have a major effect on the morphology of hippocampal neurons compared to the demonstrated effect of these agents in our control cultures of spinal cord neurons and in other reports~, 15.
Fig. 5. Silver-stained hippocampal culture. A includes an enlarged inset from lower third of fiber bundle. B and C are two focal planes within the same microscopicfield. Age of culture was 54 days and fetal gestational age was 14 days. Left hand calibrations apply to A and the right hand calibration to B and C.
238
Fig. 6. Silver-stained hippocampal culture shown at two levels of microscopic focus. Age of culture was 54 days and fetal gestational age was 14 days. Morphology of processes in older cultures
Silver-staining reveals that neuronal processes in cultures older than two weeks do not degenerate; in fact a loosely interconnected network of silver-stained fibers is present just beneath the nerve cell bodies (compare Fig. 5B focused on the cell bodies with Fig. 5C at a deeper focal plane). In some areas of the culture, these fibers are organized into fascicles consisting of lo-20 fibers (Fig. 5A with inset which shows the fascicle at a higher magnification). In sparser cultures it is possible to identify single neurons with a major process exiting from one pole of the cell (Fig. 6, arrow). Again, at deeper focal planes (Fig. 6B) many additional fibers are seen. Some fibers interdigitate with the branches of this cell and obscure the full extent of its branching pattern. In order to visualize more completely the morphology of hippocampal neurons, they have been filled with the fluorescent dye Lucifer Yellow CH under direct microscopic observation while simultaneously recording spontaneously occurring synaptic or action potentials. An example of this combined morphologic and electrophysiologic data is shown in Fig. ‘7. Here an isolated neuron (Fig. 7A), which under phase contrast optics has only a short process visible, has in fact a relatively well-developed branching pattern when filled with Lucifer Yellow CH dye (Fig. 7B). In
239
C SECONDS •
B
. ,..,j
L_,
r
D | L
I
IO0/JM
"~
Fig. 7. Spontaneous electrical activity recorded from Lucifer Yellow stained cell. A shows phase-contrast photomicrograph of neuron before injection of Lucifer Yellow. B printed as a positive (black on white) image. C and D are chartwriter and oscilloscope recordings, respectively, of ongoing electrical activity. The culture was 19 days old and prepared from fetuses of 14 and 16 days gestational age. this cell there are at least 4 orders o f branching which can be counted in a centrifugal direction with respect to the cell body. While the neuron was filling with dye, short depolarizing bursts (Fig. 7C) occurred at a rather regular frequency of about 1.4 Hz over a period o f about 9 rain. This depolarizing activity appears to consist of a combination of synaptic and local action potentials (Fig. 7D). Processes tended to emanate from one side of the soma in 22 out of 28 fluorescent-stained cells, either originating at the cell b o d y (11 cells) or from a single trunk (11 cells). Several examples are shown o f a remarkably well-developed branching pattern attained by these deceptively simple round cells when viewed under incandescent as c o m p a r e d to ultraviolet illumination in Fig. 8A1 and 2, 8B1 and 2, and 8C~,2 and ~, respectively). Both kinds of optical illumination were used together in Fig. 8A1 to establish how the r a n d o m l y oriented processes of the cell were related to the
240
241 surrounding culture. Such r a n d o m orientation was unusual and m a y be compared with the cell in Fig. 8B1 (arrow) which has multiple processes with a definite orientation relative to the cell b o d y (Fig. 8Bz) as did the other dye-stained cells in Fig. 8. Although dye-staining revealed most of the processes lateral to the cell body, processes immediately beneath the cell body could not always be visualized until the cell body was pulled partly away from the surface of the culture by the microelectrode. Such was the case for the cell in Fig. 8C2 which was resting upon a small tangle of previously hidden processes (Fig. 8C3), Some processes exhibited specialized basketlike ending (Fig. 8A3), beading (Fig. 8Bz ~nd 3), and a suggestion of spines (Fig. 8A~ and C~).
Electron microscopic observations Examination of cultures was performed both in sections cut parallel and perpendicular to the plane of the culture dish. While m a n y neural somata were located at or near the upper surface of the culture, m a n y were embedded within small clusters o f neurons (see Fig. 9a). Glial cells were almost completely confined to the interior of the cell cluster. Neuronal somata appeared as similar to neurons from whole animals, with no remarkable cytologic differences. As many as 7 synaptic terminals could be seen abutting upon one cell body in one section (Fig. 9a), although the usual number was 1-3. The majority of these terminals formed symmetric synapses (presumably inhibitory), though in most cases at least one asymmetric junction was found (presumably excitatory, see Fig. 9a, b, c). Dendrites with numerous synapses along them occurred frequently (Fig. 10a). Spines originating from these dendrites were regularly observed, and were surrounded with synaptic terminals, usually o f the asymmetric variety (Fig. 10a, b). N u m e r o u s synapses upon dendritic shafts were also seen. The large majority of axodendritic synapses was asymmetric. No clear examples o f mossy fiber terminals were seen. Large synaptic terminals were observed, making multiple contacts upon dendrites (Fig. 10c), but we failed to observe the typical pattern of multiple postsynaptic dendrites s. Whether this indicates an absence of real mossy fibers, perhaps due to the immaturity of granule cells at the time of culture z,3, or whether the altered geometry of the culture system disturbs the usual morphologic patterns is u n k n o w n to us.
Fig. 8. Lucifer Yellow stained neurons from several hippocampal cultures. Either action or synaptic potentials were recorded from each cell at the time of fluorescent staining. Phase contrast photomicrographs in A1, BI, and C1 correspond respectively to fluorescent stained cells in A2, B2, and C2,3. A1 is a double exposure of the fluorescent and phase-contrast views. A3 and B~ are separate cells for which a phase-contrast counterpart has not been illustrated. C3 shows the same cell as C2, but with the cell body pulled away to reveal underlying branches bidden by the cell body. The fluorescent shadow of the microelectrode is seen in C2 and Ca. For AI,2: culture age 19 days; fetal ages 14 and 17 days. For A3: culture age 44 days; fetal age 18 days. For B1._9:culture age 51 days; fetal age 13 days. For B3: culture age 36 days; fetal age 18 days. For Cl,_9,:3:culture age 31 days; fetal age 14 days.
Fig. 9, a: a n e u r o n i s s h o w n at low magnification ( x 2400) to illustrate the location o f several a x o s o m a t i c synapses. D a r k arrows indicate t h e loci of s y m m e t r i c synapses, a n d o p e n arrows the loci o f a s y m m e t r i c s y n a p s e s (nucl ~-: nucleolus). T h e s y m m e t r i c s y n a p s e m a r k e d by a n asterisk in a is enlarged in b a n d the synaptic zone is indicated by a n arrow. >~ 48,500. c: a n a x o s o m a t i c s y n a p s e f r o m a different n e u r o n t h a n in a with arrow indicating the s y m m e t r i c contact zone. - 48,500. T h e culture wa~ 22 days old a n d f r o m a fetus of 17 days gestationat age.
243
Fig. 10. a: a dendrite ( D E N D ) receiving a synapse upon its shaft (dark arrow) and forming a spine (*). The spine is receiving an asymmetric synapse with at least two contact zones (open arrows). × 31,500. b: a large, isolated asymmetric synaptic contact (arrow) on what is probably a dendritic spine. × 36,000. c: a large dendritic spine is receiving three synapses (1,2,3) with at least 5 contact points (dark arrows). Synapse l appears to make symmetric contact, while 2 and 3 appear asymmetric, even though some o f the vesicles are flattened (open arrow). × 32,500. Same culture as for Fig. 9.
244 DISCUSSION The major conclusion from the present work is that a marked degree o f morphologic differentiation has developed or been maintained in dissociated cultures of mouse hippocampus for 1-2 months. At the time of culture the prenatal hippocampus is highly enriched for pyramidal cells in various stages of proliferation z.:~ and it is likely that many, if not most, of the neurons in this study are of pyramidal cell origin. Whether these cells are already postmitotic at the time of culture remains to be demonstrated. Nevertheless, it is improbable that a neuronal population of different origin would have proliferated within these primary cultures to any substantial degree. The cultures in this report are rather distinctively different from those described by Banker and Cowan 3. Some of these differences reflect the different methodology that was developed to provide long-term cultures for electrophysiology. These methodologic differences include mechanical and not trypsin dissociation procedures. collagen-coated and not polylysine-treated surfaces, a COz atmosphere which was permissive for the growth of non-neuronal cells, relatively high plating densities, and the deliberate inclusion of small aggregates of cells with single cells in the initial cell dispersion. The presence of small neuronal aggregates and reaggregates very likely contributed to the maintenance of cultures beyond the survival barrier of about two weeks reported for dissociated hippocampus 3 and cerebellumlL In addition, the use of conditioned medium from spinal cord cultures is beneficial, if not necessarily essential, in obtaining a reasonable yield of long-term cultures. The conditioned medium was collected from at least 2-week-old spinal cord cultures which themselves have a capability for long-term survival 13,15. Whether spinal cord cultures produce a long-term survival-promoting factor and, if so. whether this factor has nutritional or trophic effects, or protective effects against toxic or infectious agents is speculative at this time. The morphologic isolation of individual neurons from surrounding cells by selective fluorescent dye-filling with Lucifer Yellow C H is should allow substantial progress in identification of specific pyramidal cell types by using a combination of morphological and electrophysiological criteria. In this regard the fluorescent dye, Lucifer Yellow CH, appears particularly promising. It would be predicted from Golgi studies of young mice 4,11 and kittens 14 that a considerable diversity of dendritic branching patterns would be encountered in culture, assuming thai cultured neurons could morphologically differentiate to some extent as might have occurred in situ. That such differentiation has occurred is strongly suggested by the appearance of fluorescent dye-filled cells in Fig. 8A3, Be, and Ba which have developed an orderly pattern of branching with a definite orientation relative to the cell body. Because of the age of the cultures, it is probable that this orientation is actively maintained to counter the movement or proliferation of surrounding non-neuronal cells which could otherwise distort the orientation of neuronal processes as possibly was the case in Fig. 8A1. Nevertheless there is a striking similarity between some features of dye-filled neurons in these mouse hippocampal cultures and Golgi-stained hippocampal neurons from young m~ce illustrated by Lorente de No: compare the basket type ending shown in Fig. 8An with the pyramidal basket cells of Lorente do No. Fig. 9, cell 9~L
245 Ultrastructural observations of these cultures are generally similar to observations made on hippocampus in vivo~, v,s in the following respects. (1) Synapses are distributed largely on dendrites, with a smaller number on somata: (2) dendritic spines are present, and are contacted principally by asymmetric synaptic junctions; and (3) symmetric synapses are relatively more common on somata and proximal dendrites. An occasional symmetric junction was found on a spine (Fig. 10c), which may correspond to similar rare observations by Gottlieb and Cowan 7, or may reflect an immaturity in the placement of inhibitory synapses, as discussed by Schwartz et a156. The conditions of tissue culture may have been responsible for producing a larger number of axosomatic synapses than was reported for rat and cat hippocampal CAs neurons in vivoL That is, the flattened, somewhat two-dimensional environment may have produced an apparent change in the number of such synapses per neuron when in fact it is their distribution which is altered. Whether this was the case or not, the basic pattern of synaptic distribution closely resembles that in vivo. Previous studies of hippocampal explants in vitro9, '0 have shown that many of the ultrastructural features of hippocampal organization (cell layers, synaptic distributions, gliogenesis, myelin formation) persist in these preparations over periods of several weeks. Even when hippocampal cells are dissociated, they can at certain ages reaggregate into a histologic pattern typical of the normal hippocampusZ. The present cultures have shown no tendency to form lamination similar to that of the hippocampus, even though partial reaggregation has occurred. Again, this may be due to the two-dimensional constraints of the preparation, as contrasted with De Long's5 rotary liquid cultures. Our cultures were plated at a period of embryonic life when only some dentate neurons are postmitotic 2. This may account for the paucity of mossy fiber endings in our material. At embryonic day 18, the latest day when cells for our cultures were removed, most dentate neurons have not yet become postmitotic, and therefore have not extended axons toward CA3. It is of some interest that synaptic terminals we observed upon spines were large and often formed multiple contacts. If these are not mossy fibers, as we believe they are not, it suggests that the spines and their recipient membrane zones may be able to influence the shape of the presynaptic terminals they attach to, of dentate origin or not. The remarkable extent to which these neurons have been able to develop symmetric and asymmetric synaptic structures would predict a correspondingly welldeveloped functional activity. This is confirmed in the succeeding report which describes an abundance of inhibitory and excitatory postsynaptic potentials recorded from these neurons. ACKNOWLEDGEMENTS
This work was supported by NIH Grant NS 12151 to J.H.P. and NS 11669 to L.K.M. We are grateful to Diane Long, Robin Lawson, and John Oehlert for technical assistance, to Reed Pike for help with photography, and to Cheryl Joo for typing the
246 m a n u s c r i p t . W e a l s o t h a n k D r . D a v i d P r i n c e f o r h e l p f u l s u g g e s t i o n s a n d his r e v i e w o f t h e m a n u s c r i p t . D r . W a l t e r S t e w a r d p r o v i d e d t h e g e n e r o u s gift o f L u c i f e r Y e l l o w C H t o us.
REFERENCES 1 Andersen, P., Organization of hippocampal neurons and their interconnections, in R. L. Isaacson and K. H. Pribram (Eds.), The Hippocampus, Vol. 1, Plenum Press. New York. 1975. p p 155-175. 2 Angevine, J. B., Development of the hippocampal region. In R. L. Isaacson and K. H. Pribram (Eds.), The Hippocarnpus, I1ol. 1. Plenum Press, New York. 1975. pp. 61-90. 3 Banker, G. A. and Cowan, W. M.. Rat hippocampal neurons in dispersed cell culture. Brain Research, 126 (1977) 397425. 4 Cajal, S. R., Estructura del Asta de Ammon. Anal. Soc. Esp. hist. nat. Madrid, 22 (1893J 53-113. 5 DeLong, G. R., Histogenesis of fetal mouse tsocortex and hippocampus in reaggregating cell cultures, Develop. BioL, 22 11970) 563-583. 6 Godfrey, E. W., Nelson, P. G., Schrier, B, K., Breuer, A. C. and Ransom, B. R., Neurons from fetal rat brain in a new cell culture system: a multidisciplinary analysis, Brain Research. 90 (1975) 1 -21. 7 Gottlieb. D. I. and Cowan, W. M.. On the distribution of axonal terminals containing spheroidal and flattened synaptic vesicles in the hippocampus and dentate gyrus of the rat and cat, Z. Zellforsch., 129 (1972) 413-429. 8 Hamlyn, L. H.. The fine structure of the mossy fibre endings in the hippocampus of the rabbit. J. Anat. Lond.), 96 (1962) 112-120. 9 Kim, S. U., Morphological development of neonatal mouse hippocampus cultured in vitro. L~p. Neurol., 41 (1973) 150 t62. 10 LaVail. J. H. and Wolf. M. K,, Postnatal development of the mouse dentate gyrus m organotypic cultures of the hippocampal formation, Amer. J. Anat.. 137 (1973) 47-66. 11 Lorente de No, R., Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system, J. Psychol. Neurol. (Lpz.), 46 (1934) 113-177. 12 Nelson. P. G. and Peacock, J. H,, Electrical activity of dissociated cell cultures from fetal mouse cerebellum, Brain Research, 61 (1973) 163-174. 13 Peacock, J. H., Nelson, P. G. and Goldstone. M. W.. Electrophysiologic study of cultured neurons dissociated from spinal cords and dorsal root ganglia of fetal mice, Develop. Biol., 30 (1973) 137-152. 14 Purpura, D. P. and Pappas, G. D., Structural characteristics of neurons in the feline hippocampus during postnatal ontogenesis, Exp. Neurol., 22 I1968) 379-393. 15 Ransom, B. R., Neale. E., Henkart, M., Bullock, P. N. and Nelson, P. G., Mouse spinal cord in cell culture. 1. Morphology and intrinsic neuronal electrophysiologic properties, J. NeurophysioL. 40 (1977) 1132-1150. 16 Schwartz. 1. R., Pappas, G. D. and Purpura, D. P., Fine structure of neurons and synapses in the feline hippocampus during postnatal ontogenesis, Exp. NeuroL, 22 (1968) 394-407. 17 Sevier. A. C. and Munger, B. L.. A silver stain for paraffin sections of neural tissue, J. Neuropath. exp. NeuroL. 24 (1965) 130-135. 18 Stewart. W. W., Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphthalamide tracer, Cell, 14 119781 741-759. 19 Zipser. B.. Crain, S. M. and Bornstein. M. B., Directly evoked 'paroxysmal' depolarizations of mouse hippocampal neurons in synaptically organized explants in long-term culture, Brain Research. 60 (1973) 489495.
231
M O R P H O L O G Y OF D I S S O C I A T E D H I P P O C A M P A L C U L T U R E S F R O M FETAL MICE
JOHN H. PEACOCK, DAPHNE F. RUSH and LAWRENCE H. MATHERS Department of Neurology, Stan/brd University Medical Center, Stanford, Cali/i 94305 (U.S.A.)
(Accepted October 5th, 1978)
SUMMARY Dissociated hippocampal cultures from fetal mice (13-18 days gestational age) can be maintained for up to two months in culture. Cells grow as either isolated neurons or in small neuronal aggregates. Neurons remain small with a soma diameter of 15-20 ,urn even in mature cultures and develop extensively branched processes during the first two weeks in culture. After this time, processes become more difficult to visualize with phase-contrast optics because of a tendency to grow within the underlying non-neuronal cells. However, the presence of processes has been proved by silver-staining which demonstrates an organizational complexity ranging from a loosely reticulated neuropil to fascicles containing many fibers. More detailed study of individual neuronal morphology was carried out in cells filled with the fluorescent dye, Lucifer Yellow CH, in conjunction with the intracellular recording of synaptic and action potentials from dye-containing micropipettes. Dye-filled cells show a well-developed branching morphology. Process specializations include spines, beading, and basket-like endings. Processes tend to emanate from one side of the soma, either originating at the cell body or from a single trunk. Commonly there are 2-4 orders of branching, but up to 6 orders can occur (counted centrifugally from the soma). Electron microscopy revealed synapses distributed predominantly on dendrites with a smaller number on somata. Dendritic spines are present and are contacted principally by asymmetric synaptic junctions. Symmetric synapses are relatively more common on somata and proximal dendrites.
INTRODUCTION
Differentiated properties of hippocampal neurons develop or are preserved to a remarkable extent in organotypic cultures from both morphologicga ° and intracellular electrophysiologic t9 standpoints, although the latter has been studied less extensively.
232 Further electrophysiotogic analysis of cultured hippocampal neurons is of considerable interest and would be greatly facilitated if hippocampal neurons could be grown in a dissociated, surface culture system where cells are visually accessible for intracellular penetration. This has been accomplished for short-term Cultures (up to two weeks) by Banker and Cowan ~ who give evidence that the development of neurons in dissociated hippocampal cell cultures is consistent with that of pyramidal cells in the intact animal. We have developed a long-term (up to two months) dissociated hippocampal cell culture system for electrophysiology and in this report describe the major morphologic features of these cultures. In addition we show single neurons whose dendritic morphology has been visually isolated from surrounding cells after intracellular fluorescent dye-staining with the new fluorescent probe Lucifer Yellow CH is and, from which, spontaneous electrical activity has been recorded while staining is visually monitored. MATERIALS AND METHODS
Dissection, dissociation and culture procedures Whole brains from 4 to 6 Swiss Webster mouse fetuses, 13-18 days gestationaf age, were removed under sterile conditions and placed in chilled Minimal Essential Eagle's Medium (MEM) containing 6.0 g/liter glucose and 12.5 m M Hepes buffer. Each hemisphere was separated from the thalamus, the hippocampal fissure was identified by the vascular pattern (Fig. I A, left hemisphere, arrows) and the meninges removed in older fetuses (16--18 days in utero). The hippocampus then appeared as a dark band when viewed under oblique transillumination. With iridectomy scissors, a transversely oriented incision was made into the rostral and caudal parts of the hippocampus (Fig. IB) and the hippocampus was cut away from the hemisphere along its
Fig. 1. Photomicrograph of medial surfaces of the left and right cerebral hemispheres. A : the meninges have been left intact to demonstrate the vessels which mark the hippocampal fissure (arrowsJ. B'. the meninges have been stripped away and two incisions (arrows}have been made through thehippoeampus which now appears as a dark band of tissue. Fetal gestational age was 16 days.
233 longitudinal axis. The dissected pieces of hippocampus were collected in a 35 mm collagen-coated plastic dish (Falcon 3001) which contained 1 ml of growth medium (see below) at 20 °C. After 6-12 hippocampi had been pooled, the tissue was aspirated twice using a I ml plastic syringe and a 27-gauge needle. Cells were distributed from the syringe into 35 mm collagen-coated dishes at a density of 0.5-1 x 106 cells per dish (0.5-1.0 ~,i 10~ ceIls/sq, cm) in a volume of 1.5 ml medium. Collagen was prepared by suspending 150 mg of calf skin collagen (Worthington) in 100 ml double distilled water and autoclaving to dissolve. The suspension was filtered through No. 2 Whatman paper and sterilized by reautoclaving the clarified collagen. Growth medium consisted of MEM or Dulbecco's Modified Eagle's Medium (DMEM) added in equal volume to conditioned medium from spinal cord cultures. Conditioned medium was MEM or DM EM containing 10 ~ fetal calf serum and 10 ~ horse serum removed every 3--4 days from mouse spinal cord cell cultures at least 14 days old. Growth medium, with a final serum concentration of 5 ~ each for fetal calf serum and horse serum, was then filtered through a 0.2/zm Millipore filter, and stored at 4 °C until use. The first medium change was made after 4 days in culture and subsequent changes every 7 days. In some cases 10-20/zg/ml 5-fluoro-2'-deoxyuridine (FdU) and 25-50/zg/ml uridine were added during days 7-I 1 ofculture~L No antibiotics were used.
Histology and electron microscopy Silver-staining of cultures was carried out according to a modification of the procedure of Sevier and Munger 17 after fixation in 2 ~ cacodylate in buffered formalin (4"/~ 16- or 17-day / O J" For light microscopy, the hippocampus was dissected from fetuses, fixed in 2 IV,,glutaraldehyde, postfixed in 1 ~ OsO4, dehydrated, embedded in Spurr, sectioned at I /zm thickness, and stained with hot 1 o/methylene blue in I ~i sodium borate buffer. For electron microscopy, cultures were fixed in 2 ~ glutaraldehyde in 0.1 M Sorensen's phosphate buffer plus 0.1 M sucrose for a final osmolarity of 392 mOsm at pH 7.3 for at least 24 h at 4 °C, stored in Sorensen's phosphate buffer at 4 °C, postfixed in I ~ OsO4, dehydrated in ethanol, and embedded in Epon. Embedded cultures were removed from the plastic dish and designated areas were sectioned in a plane parallel or perpendicular to the surface of the culture. Sections were stained with lead acetate and in some cases with uranyl acetate.
Electrophysiology and ,fluorescent dye staining Culture dishes were placed in a chamber on the stage of a Zeiss inverted phase contrast microscope where they were kept at 37 °C by heating the air above and below the chamber, at pH 7.2, by flowing 5 ~,, COz in air over the culture, and at constant osmolarity by spreading a layer of light mineral oil on the surface of the medium. Calcium concentration was increased from 1.8 mM to 3.5 raM. A Zeiss epilluminator (IV FL) containing an ultraviolet light source (50 watt DC mercury bulb) was mounted between the body of the microscope and the eyepieces for simultaneous or sequential phase contrast and fluorescent microscopy. For fluorescence, Zeiss excita-
234 tion filters B G 12 and U G 5 (peak transmission of about 395 nm), a 460 nm dichroic reflector, and a 500 nm barrier filter were used. Microelectrodes were pulled from microfiber glass capillary tubes, filled with a 4 ~ aqueous solution of Lucifer Yellow C H and bevelled until tip resistances of 135-175 M R were achieved. After cells were penetrated using phase-contrast optics, the incandescent light source was turned off and neurons were visually monitored using ultraviolet optics until fluorescent dye staining was complete, about 3 5 min. Lucifer Yellow C H is negatively charged and was ejected from the micropipette by applying a steady negative DC current of 0.5-3 nA to the dye solution. A bridge circuit was used to simultaneously monitor stimulating currents and voltage responses from the same microelectrode. Currents and voltages were always recorded on a cathode ray oscilloscope, a chart recorder, and in some cases on FM tape. RESULTS
Morphological appearance of the hippocampus at time of culture The portion of the hippocampus that was removed for culture wa~ compared with the ablated area of the brain from which it was dissected. The dissection included the fimbria and adjacent zones corresponding to CA2, CA3, and CA4 pyramidal cells along with a portion of the developing dentate area. The CA~ pyramidal zone was present to a variable extent and it is a possibility that cells from the subicutum were included in some dissections. Light microscopy observauons suggested that many pyramidal neurons in the zone corresponding to CA3 were postmitotic with their pale staining nuclei, two or more nucleoli, and orientation of large proximal dendrites toward the interior of the hippocampal formation. Darkly staining, and presumably proliferating, pyramidal cells were present along the ventricular zone and. similarly, intensely stained cells were found in the tmmature dentate area.
Fig. 2. Silver-stained neuron in 2,day-01d culture from fetus of 18 days gestational age.
235 Gestational age at time o f culture Cultures were prepared from fetuses of 13-18 days gestational age. Dissection of 13-15-day fetuses presents some difficulty due to reduced contrast in the density of t he cell layers and the fragility of the tissue. It is difficult to remove the meninges completely until fetuses reach 16 days in utero. However, cultures prepared from animals over this fetal age range display a similar morphology at various stages of culture growth. Neurons with a major bifurcating process are frequently found in young cultures from a time shortly after the initial plating (Fig. 2) up to about two weeks as shown in Fig. 3A and B from a 13-day-old fetus, and Fig. 3C and D from 17day-old fetuses. There does not appear to be a difference in terms of neuronal survival or in the rate of successful cultures over this fetal age range. However, cultures from 18-day-old fetuses are subject to greater cell death than cultures from the younger fetuses and, on this empirical basis, the hippocampus from 18-day fetuses was not routinely cultured. The use of conditioned medium from spinal cord cultures contributes to neuronal survival for periods in culture longer than two weeks either when used from the beginning of the culture or to rescue cultures which had not been grown with conditioned medium until deterioration was detected at about two weeks in culture. In these cases the results have been sufficiently dramatic that we routinely use conditioned medium in long-term cultures. However, it should be noted that we have on occasion observed long-term survival without conditioned medium.
Fig. 3. Phase contrast photomicrographs ofhippocampal neurons in week old cultures from fetuses of 14 days gestational age (A and B) and 17 days gestational age (C and D).
236
Fig. 4. Phase contrast photomicrographs from hippocampal cultures after 5 days in culture (A), 10 days in culture (B), and 4 weeks in culture (C). Fetal gestational age 13 days.
Culture development Cultures progress t h r o u g h two m a j o r m o r p h o l o g i c a n d d e v e l o p m e n t a l stages. The first stage is shown in Fig. 4 A and B a n d the second in Fig. 4C: all examples are
237 from the same culture dish. Cells attach to the surface of the culture dish within the first 24 h and extend processes. Frequently but not always they grow upon several contiguous flat non-neuronal cells. After 3 4 days in culture (Fig. 4A) it is evident whether a culture is going to survive for the next two weeks. Over the course of 10 days the cells become more phase bright, reach a relatively uniform size of 15-20/zm, and develop many interlacing processes (Fig. 4B). At this stage non-neuronal background cells cover the surface of this dish. Some of these are ependymal cells which beat rhythmically and are frequently found in association with neurons. After two weeks in culture the processes become difficult to visualize with phase contrast optics (Fig. 4C, left and right hand cell groups) and in some areas of a culture, or even throughout entire cultures, seem to disappear. But hints of the existence of processes are given by occasional thin neuritic bridges which interconnect small groups of cells (Fig. 4C) or short segments of processes which are visible for several cell diameters away from a cell body before merging with background cells. Treatment of the cultures with FdU plus uridine did not have a major effect on the morphology of hippocampal neurons compared to the demonstrated effect of these agents in our control cultures of spinal cord neurons and in other reports~, 15.
Fig. 5. Silver-stained hippocampal culture. A includes an enlarged inset from lower third of fiber bundle. B and C are two focal planes within the same microscopicfield. Age of culture was 54 days and fetal gestational age was 14 days. Left hand calibrations apply to A and the right hand calibration to B and C.
238
Fig. 6. Silver-stained hippocampal culture shown at two levels of microscopic focus. Age of culture was 54 days and fetal gestational age was 14 days. Morphology of processes in older cultures
Silver-staining reveals that neuronal processes in cultures older than two weeks do not degenerate; in fact a loosely interconnected network of silver-stained fibers is present just beneath the nerve cell bodies (compare Fig. 5B focused on the cell bodies with Fig. 5C at a deeper focal plane). In some areas of the culture, these fibers are organized into fascicles consisting of lo-20 fibers (Fig. 5A with inset which shows the fascicle at a higher magnification). In sparser cultures it is possible to identify single neurons with a major process exiting from one pole of the cell (Fig. 6, arrow). Again, at deeper focal planes (Fig. 6B) many additional fibers are seen. Some fibers interdigitate with the branches of this cell and obscure the full extent of its branching pattern. In order to visualize more completely the morphology of hippocampal neurons, they have been filled with the fluorescent dye Lucifer Yellow CH under direct microscopic observation while simultaneously recording spontaneously occurring synaptic or action potentials. An example of this combined morphologic and electrophysiologic data is shown in Fig. ‘7. Here an isolated neuron (Fig. 7A), which under phase contrast optics has only a short process visible, has in fact a relatively well-developed branching pattern when filled with Lucifer Yellow CH dye (Fig. 7B). In
239
C SECONDS •
B
. ,..,j
L_,
r
D | L
I
IO0/JM
"~
Fig. 7. Spontaneous electrical activity recorded from Lucifer Yellow stained cell. A shows phase-contrast photomicrograph of neuron before injection of Lucifer Yellow. B printed as a positive (black on white) image. C and D are chartwriter and oscilloscope recordings, respectively, of ongoing electrical activity. The culture was 19 days old and prepared from fetuses of 14 and 16 days gestational age. this cell there are at least 4 orders o f branching which can be counted in a centrifugal direction with respect to the cell body. While the neuron was filling with dye, short depolarizing bursts (Fig. 7C) occurred at a rather regular frequency of about 1.4 Hz over a period o f about 9 rain. This depolarizing activity appears to consist of a combination of synaptic and local action potentials (Fig. 7D). Processes tended to emanate from one side of the soma in 22 out of 28 fluorescent-stained cells, either originating at the cell b o d y (11 cells) or from a single trunk (11 cells). Several examples are shown o f a remarkably well-developed branching pattern attained by these deceptively simple round cells when viewed under incandescent as c o m p a r e d to ultraviolet illumination in Fig. 8A1 and 2, 8B1 and 2, and 8C~,2 and ~, respectively). Both kinds of optical illumination were used together in Fig. 8A1 to establish how the r a n d o m l y oriented processes of the cell were related to the
240
241 surrounding culture. Such r a n d o m orientation was unusual and m a y be compared with the cell in Fig. 8B1 (arrow) which has multiple processes with a definite orientation relative to the cell b o d y (Fig. 8Bz) as did the other dye-stained cells in Fig. 8. Although dye-staining revealed most of the processes lateral to the cell body, processes immediately beneath the cell body could not always be visualized until the cell body was pulled partly away from the surface of the culture by the microelectrode. Such was the case for the cell in Fig. 8C2 which was resting upon a small tangle of previously hidden processes (Fig. 8C3), Some processes exhibited specialized basketlike ending (Fig. 8A3), beading (Fig. 8Bz ~nd 3), and a suggestion of spines (Fig. 8A~ and C~).
Electron microscopic observations Examination of cultures was performed both in sections cut parallel and perpendicular to the plane of the culture dish. While m a n y neural somata were located at or near the upper surface of the culture, m a n y were embedded within small clusters o f neurons (see Fig. 9a). Glial cells were almost completely confined to the interior of the cell cluster. Neuronal somata appeared as similar to neurons from whole animals, with no remarkable cytologic differences. As many as 7 synaptic terminals could be seen abutting upon one cell body in one section (Fig. 9a), although the usual number was 1-3. The majority of these terminals formed symmetric synapses (presumably inhibitory), though in most cases at least one asymmetric junction was found (presumably excitatory, see Fig. 9a, b, c). Dendrites with numerous synapses along them occurred frequently (Fig. 10a). Spines originating from these dendrites were regularly observed, and were surrounded with synaptic terminals, usually o f the asymmetric variety (Fig. 10a, b). N u m e r o u s synapses upon dendritic shafts were also seen. The large majority of axodendritic synapses was asymmetric. No clear examples o f mossy fiber terminals were seen. Large synaptic terminals were observed, making multiple contacts upon dendrites (Fig. 10c), but we failed to observe the typical pattern of multiple postsynaptic dendrites s. Whether this indicates an absence of real mossy fibers, perhaps due to the immaturity of granule cells at the time of culture z,3, or whether the altered geometry of the culture system disturbs the usual morphologic patterns is u n k n o w n to us.
Fig. 8. Lucifer Yellow stained neurons from several hippocampal cultures. Either action or synaptic potentials were recorded from each cell at the time of fluorescent staining. Phase contrast photomicrographs in A1, BI, and C1 correspond respectively to fluorescent stained cells in A2, B2, and C2,3. A1 is a double exposure of the fluorescent and phase-contrast views. A3 and B~ are separate cells for which a phase-contrast counterpart has not been illustrated. C3 shows the same cell as C2, but with the cell body pulled away to reveal underlying branches bidden by the cell body. The fluorescent shadow of the microelectrode is seen in C2 and Ca. For AI,2: culture age 19 days; fetal ages 14 and 17 days. For A3: culture age 44 days; fetal age 18 days. For B1._9:culture age 51 days; fetal age 13 days. For B3: culture age 36 days; fetal age 18 days. For Cl,_9,:3:culture age 31 days; fetal age 14 days.
Fig. 9, a: a n e u r o n i s s h o w n at low magnification ( x 2400) to illustrate the location o f several a x o s o m a t i c synapses. D a r k arrows indicate t h e loci of s y m m e t r i c synapses, a n d o p e n arrows the loci o f a s y m m e t r i c s y n a p s e s (nucl ~-: nucleolus). T h e s y m m e t r i c s y n a p s e m a r k e d by a n asterisk in a is enlarged in b a n d the synaptic zone is indicated by a n arrow. >~ 48,500. c: a n a x o s o m a t i c s y n a p s e f r o m a different n e u r o n t h a n in a with arrow indicating the s y m m e t r i c contact zone. - 48,500. T h e culture wa~ 22 days old a n d f r o m a fetus of 17 days gestationat age.
243
Fig. 10. a: a dendrite ( D E N D ) receiving a synapse upon its shaft (dark arrow) and forming a spine (*). The spine is receiving an asymmetric synapse with at least two contact zones (open arrows). × 31,500. b: a large, isolated asymmetric synaptic contact (arrow) on what is probably a dendritic spine. × 36,000. c: a large dendritic spine is receiving three synapses (1,2,3) with at least 5 contact points (dark arrows). Synapse l appears to make symmetric contact, while 2 and 3 appear asymmetric, even though some o f the vesicles are flattened (open arrow). × 32,500. Same culture as for Fig. 9.
244 DISCUSSION The major conclusion from the present work is that a marked degree o f morphologic differentiation has developed or been maintained in dissociated cultures of mouse hippocampus for 1-2 months. At the time of culture the prenatal hippocampus is highly enriched for pyramidal cells in various stages of proliferation z.:~ and it is likely that many, if not most, of the neurons in this study are of pyramidal cell origin. Whether these cells are already postmitotic at the time of culture remains to be demonstrated. Nevertheless, it is improbable that a neuronal population of different origin would have proliferated within these primary cultures to any substantial degree. The cultures in this report are rather distinctively different from those described by Banker and Cowan 3. Some of these differences reflect the different methodology that was developed to provide long-term cultures for electrophysiology. These methodologic differences include mechanical and not trypsin dissociation procedures. collagen-coated and not polylysine-treated surfaces, a COz atmosphere which was permissive for the growth of non-neuronal cells, relatively high plating densities, and the deliberate inclusion of small aggregates of cells with single cells in the initial cell dispersion. The presence of small neuronal aggregates and reaggregates very likely contributed to the maintenance of cultures beyond the survival barrier of about two weeks reported for dissociated hippocampus 3 and cerebellumlL In addition, the use of conditioned medium from spinal cord cultures is beneficial, if not necessarily essential, in obtaining a reasonable yield of long-term cultures. The conditioned medium was collected from at least 2-week-old spinal cord cultures which themselves have a capability for long-term survival 13,15. Whether spinal cord cultures produce a long-term survival-promoting factor and, if so. whether this factor has nutritional or trophic effects, or protective effects against toxic or infectious agents is speculative at this time. The morphologic isolation of individual neurons from surrounding cells by selective fluorescent dye-filling with Lucifer Yellow C H is should allow substantial progress in identification of specific pyramidal cell types by using a combination of morphological and electrophysiological criteria. In this regard the fluorescent dye, Lucifer Yellow CH, appears particularly promising. It would be predicted from Golgi studies of young mice 4,11 and kittens 14 that a considerable diversity of dendritic branching patterns would be encountered in culture, assuming thai cultured neurons could morphologically differentiate to some extent as might have occurred in situ. That such differentiation has occurred is strongly suggested by the appearance of fluorescent dye-filled cells in Fig. 8A3, Be, and Ba which have developed an orderly pattern of branching with a definite orientation relative to the cell body. Because of the age of the cultures, it is probable that this orientation is actively maintained to counter the movement or proliferation of surrounding non-neuronal cells which could otherwise distort the orientation of neuronal processes as possibly was the case in Fig. 8A1. Nevertheless there is a striking similarity between some features of dye-filled neurons in these mouse hippocampal cultures and Golgi-stained hippocampal neurons from young m~ce illustrated by Lorente de No: compare the basket type ending shown in Fig. 8An with the pyramidal basket cells of Lorente do No. Fig. 9, cell 9~L
245 Ultrastructural observations of these cultures are generally similar to observations made on hippocampus in vivo~, v,s in the following respects. (1) Synapses are distributed largely on dendrites, with a smaller number on somata: (2) dendritic spines are present, and are contacted principally by asymmetric synaptic junctions; and (3) symmetric synapses are relatively more common on somata and proximal dendrites. An occasional symmetric junction was found on a spine (Fig. 10c), which may correspond to similar rare observations by Gottlieb and Cowan 7, or may reflect an immaturity in the placement of inhibitory synapses, as discussed by Schwartz et a156. The conditions of tissue culture may have been responsible for producing a larger number of axosomatic synapses than was reported for rat and cat hippocampal CAs neurons in vivoL That is, the flattened, somewhat two-dimensional environment may have produced an apparent change in the number of such synapses per neuron when in fact it is their distribution which is altered. Whether this was the case or not, the basic pattern of synaptic distribution closely resembles that in vivo. Previous studies of hippocampal explants in vitro9, '0 have shown that many of the ultrastructural features of hippocampal organization (cell layers, synaptic distributions, gliogenesis, myelin formation) persist in these preparations over periods of several weeks. Even when hippocampal cells are dissociated, they can at certain ages reaggregate into a histologic pattern typical of the normal hippocampusZ. The present cultures have shown no tendency to form lamination similar to that of the hippocampus, even though partial reaggregation has occurred. Again, this may be due to the two-dimensional constraints of the preparation, as contrasted with De Long's5 rotary liquid cultures. Our cultures were plated at a period of embryonic life when only some dentate neurons are postmitotic 2. This may account for the paucity of mossy fiber endings in our material. At embryonic day 18, the latest day when cells for our cultures were removed, most dentate neurons have not yet become postmitotic, and therefore have not extended axons toward CA3. It is of some interest that synaptic terminals we observed upon spines were large and often formed multiple contacts. If these are not mossy fibers, as we believe they are not, it suggests that the spines and their recipient membrane zones may be able to influence the shape of the presynaptic terminals they attach to, of dentate origin or not. The remarkable extent to which these neurons have been able to develop symmetric and asymmetric synaptic structures would predict a correspondingly welldeveloped functional activity. This is confirmed in the succeeding report which describes an abundance of inhibitory and excitatory postsynaptic potentials recorded from these neurons. ACKNOWLEDGEMENTS
This work was supported by NIH Grant NS 12151 to J.H.P. and NS 11669 to L.K.M. We are grateful to Diane Long, Robin Lawson, and John Oehlert for technical assistance, to Reed Pike for help with photography, and to Cheryl Joo for typing the
246 m a n u s c r i p t . W e a l s o t h a n k D r . D a v i d P r i n c e f o r h e l p f u l s u g g e s t i o n s a n d his r e v i e w o f t h e m a n u s c r i p t . D r . W a l t e r S t e w a r d p r o v i d e d t h e g e n e r o u s gift o f L u c i f e r Y e l l o w C H t o us.
REFERENCES 1 Andersen, P., Organization of hippocampal neurons and their interconnections, in R. L. Isaacson and K. H. Pribram (Eds.), The Hippocampus, Vol. 1, Plenum Press. New York. 1975. p p 155-175. 2 Angevine, J. B., Development of the hippocampal region. In R. L. Isaacson and K. H. Pribram (Eds.), The Hippocarnpus, I1ol. 1. Plenum Press, New York. 1975. pp. 61-90. 3 Banker, G. A. and Cowan, W. M.. Rat hippocampal neurons in dispersed cell culture. Brain Research, 126 (1977) 397425. 4 Cajal, S. R., Estructura del Asta de Ammon. Anal. Soc. Esp. hist. nat. Madrid, 22 (1893J 53-113. 5 DeLong, G. R., Histogenesis of fetal mouse tsocortex and hippocampus in reaggregating cell cultures, Develop. BioL, 22 11970) 563-583. 6 Godfrey, E. W., Nelson, P. G., Schrier, B, K., Breuer, A. C. and Ransom, B. R., Neurons from fetal rat brain in a new cell culture system: a multidisciplinary analysis, Brain Research. 90 (1975) 1 -21. 7 Gottlieb. D. I. and Cowan, W. M.. On the distribution of axonal terminals containing spheroidal and flattened synaptic vesicles in the hippocampus and dentate gyrus of the rat and cat, Z. Zellforsch., 129 (1972) 413-429. 8 Hamlyn, L. H.. The fine structure of the mossy fibre endings in the hippocampus of the rabbit. J. Anat. Lond.), 96 (1962) 112-120. 9 Kim, S. U., Morphological development of neonatal mouse hippocampus cultured in vitro. L~p. Neurol., 41 (1973) 150 t62. 10 LaVail. J. H. and Wolf. M. K,, Postnatal development of the mouse dentate gyrus m organotypic cultures of the hippocampal formation, Amer. J. Anat.. 137 (1973) 47-66. 11 Lorente de No, R., Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system, J. Psychol. Neurol. (Lpz.), 46 (1934) 113-177. 12 Nelson. P. G. and Peacock, J. H,, Electrical activity of dissociated cell cultures from fetal mouse cerebellum, Brain Research, 61 (1973) 163-174. 13 Peacock, J. H., Nelson, P. G. and Goldstone. M. W.. Electrophysiologic study of cultured neurons dissociated from spinal cords and dorsal root ganglia of fetal mice, Develop. Biol., 30 (1973) 137-152. 14 Purpura, D. P. and Pappas, G. D., Structural characteristics of neurons in the feline hippocampus during postnatal ontogenesis, Exp. Neurol., 22 I1968) 379-393. 15 Ransom, B. R., Neale. E., Henkart, M., Bullock, P. N. and Nelson, P. G., Mouse spinal cord in cell culture. 1. Morphology and intrinsic neuronal electrophysiologic properties, J. NeurophysioL. 40 (1977) 1132-1150. 16 Schwartz. 1. R., Pappas, G. D. and Purpura, D. P., Fine structure of neurons and synapses in the feline hippocampus during postnatal ontogenesis, Exp. NeuroL, 22 (1968) 394-407. 17 Sevier. A. C. and Munger, B. L.. A silver stain for paraffin sections of neural tissue, J. Neuropath. exp. NeuroL. 24 (1965) 130-135. 18 Stewart. W. W., Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphthalamide tracer, Cell, 14 119781 741-759. 19 Zipser. B.. Crain, S. M. and Bornstein. M. B., Directly evoked 'paroxysmal' depolarizations of mouse hippocampal neurons in synaptically organized explants in long-term culture, Brain Research. 60 (1973) 489495.
