~~
Previously, embryonic tissues have been used to produce monolayer cultures containing mammalian spinal cord neurons (SCN) and motoneurons (MN) for studies of the pathophysiology of motoneuron diseases. We demonstrate here that viable SCN and MN were observed in dissociated cultures from neonatal rat and mouse. These SCN and MN produced neurites and expressed acetylcholinesterase,neuron-specific enolase and neurofilament protein. These results indicate that cultured postnatal SCN and MN are capable of survival, neurite extension, and phenotypic expression in culture. Key words: spinal cord neuron motoneuron cell culture fluorescence activated cell sorting MUSCLE & NERVE 14~14-21 1991
NEONATAL MAMMALIAN SPINAL CORD NEURONS AND MOTONEURONS IN MONOLAYER CULTURE CHARLES KRIEGER, MD, PhD, FRCP(C) and SEUNG U. KIM, MD, PhD
T h e pathophysiology of diseases affecting motoneurons (MN) are poorly understood. The technique of neural tissue culture permits the study of M N in a controlled environment and has been used for studies relevant to motoneuron diseases. 14,17 Neural tissue culture involves the mechanical or enzymatic dissociation of neural tissue into cells which are then maintained in vitro. Spinal cord cultures (SC cultures) are often monolayer cultures containing neurons and non-neuronal cellular elements of the SC which include ciand y-motoneurons, interneurons, and dorsal horn neurons. l 7 Recently, cultures enriched in their content of motoneurons (MN cultures) have been developed using the retrograde labeling of M N by fluorescent markers followed by cell sorting and these cultures consist largely of ci- and y~~.2,9,12 Generally, mammalian SC cultures or MN cultures have been prepared from embryonic material. Only a few studies have examined the viabil-
From the Division of Neurology, Department of Medicine, Health Sciences Centre Hospital, University of British Columbia, Vancouver, British Columbia, Canada. Supported by MRC (Canada) and the ALS Society of Canada. We express our thanks to M. Kim, D. Osborne, and D Zecchini. Address reprint requests to C. Krieger, Division of Neurology, Department of Medicine, Health Sciences Centre Hospital, University of British Columbia, 221 1 Westbrook Mall, Vancouver, B.C , V6T 1W5, Canada Accepted for publication November 30, 1989. CCC 0148-639X/91/01014-08 $04 00 0 1991 John Wiley & Sons, Inc
14
Cultured Spinal Neurons
ity or phenotypic expression of monolayer cultures of postnatal spinal neurons"" and these have met with little success. An evaluation of the viability of postnatal spinal cord neurons (SCN) and MN in culture is of importance as embryonic neurons differ from postnatal neurons with the most notable difference being a period of naturally occurring cell death during embryogenesi~~; in the mouse this be ins at embryonic day 13 and is completd at birth! Since there is no significant postnatal motoneuron cell death in the rat" or mouse7 investigations of cell death in neurons from neonatal animals may be more relevant to an understanding of the mechanism of cell death in adult human neurons than comparable studies using embryonic cells. In this report we demonstrate that cultures of spinal cord cells from neonatal rodents can be prepared which contain viable SCN and MN. To confirm the presence of M N in these cultures we have retrograde-labeled these cells in situ followed by dissociation and fluorescence-activated cell sorting. MATERIALS AND METHODS
Neonatal rats (Sprague-Dawley) or mice (CU-1), aged between 1 to 6 days were cold anesthetized, decapitated, and the spinal cords were removed. Cords were minced, incubated in 1.0% trypsin (Gibco, Grand Island, NY) and 0.01% DNAse (Sigma, St. Louis, MO) in Dulbecco's phosphate buffered saline with 10 mg % D-glucose (PBS-D) for 10 min at 37°C followed Preparation of S c Cultures.
MUSCLE & NERVE
January 1991
Cultured Spinal Neurons
MUSCLE & NERVE
January 1991
15
FIGURE 1. (Opposite page and above) Murine spinal cord neurons from newborn animals in culture. (A) phase, (B) NF immunofluorescence, (C) NSE immunocytochernistry, (D)AchE histochemistry. Calibration bars, 20 wm.
by incubation in 0.1% trypsin for 10 min. Dissociated cells were obtained by repeated gentle trituration. T h e supernatant was collected in a tube containing 0.6 mL of heat-inactivated horse serum (Gibco), passed through a 50 pm nylon mesh filter, and concentrated by centifugation at 800 rpm for 10 min. Cells were then washed in PBS-D, suspended in feeding medium and plated onto poly-l-lysinecoated coverslips in concentrations of 2 to 8 x lo6 cells/mL of which over 90% were viable as determined by trypan blue exclusion. The feeding medium consisted of 90% Eagle’s minimum essential medium with 10% fetal calf serum (Gibco), 5 mgi mL D-glucose, 50 pg/mL streptomycin, and 50 U/ mL penicillin. Coverslips were placed in dishes maintained at 36°C in a 5% co2 incubator with media changes twice per week. Preparation of Cultures Containing Identified MN (MN Cultures). For preparation of cultures en-
riched in their content of M N or containing identified MN all 4 extremities of cold anesthetized neonatal or 2 day-old animals were injected with 3.0 to 4.0 pL of 0.4% rhodamine isothiocyanate (RITC, Sigma) or 10% horseradish peroxidase (HRP, Toyobo, Osaka, Japan). Injections were made into the muscle masses of the proximal portions of each limb using a Hamilton microsyringe (Reno, NV).
16
Cultured Spinal Neurons
After 24 to 48 hours had elapsed, spinal cords were removed and dissociated as described above. In some experiments the cell suspension was then sorted using a Becton-Dickinson FAGS IV flow cytometer operating at 488 nm for light scatter and 568 nm for fluorescence measurements with a power output of 300 mW.I5 Cells were analyzed at rates of 500 to 1000 eventsisec, where events constituted live and dead cells as well as cell fragments. Sorted cells were collected in sterile glass tubes containing feeding medium and were plated onto glass coverslips in Petri dishes coated with poly-L-lysine alone, poly-L-lysine and EHS laminin (BKL, Bethesda, MD), or feeder layers of rat brain astrocytes6 After 2 to 14 days in vitro, cultures were examined for the presence of neurofilament protein (NF), neuronspecific enolase (NSE), and acetylcholinesterase (AChE). NF8 immunoreaction was performed by fixation in acid alcohol and incubation with a primary antibody and rhodamine or fluorescein-conjugated second antibodies (Cappel, West Chester, PA) used at 1 : 100 dilution. NSEI3 was detected after fixation in 4% paraformaldehyde and incubation in primary antibody using a biotin-avidin method (Vectastain rabbit IgG ABC kit, Vector Labs, Burlingame, CA). AChE was determined as described previously16 but with the addition of 24 Cytochemistry and Immunocytochemistry.
MUSCLE & NERVE
January 1991
FIGURE 2. (A) and (6):rnotoneurons labelled by retrograde transport of HRP from muscle. Cells from 2-day-old mice plated 3 days prior to fixation. Calibration bars 20 p.
p M tetraisopropylpyrophosphoramide (Sigma) to block pseudocholinesterase activity. The presence of intracellular HRP was detected using the diaminobenzidine method after paraformaldehyde f i ~ a t i o n .Monoclonal ~ antibody specific for 1501 200 kda neurofilament triplet protein was provided by Dr. V. Lee' and rabbit antiserum monospecific for NSE was obtained from Dako (Santa Barbara, CA). RESULTS
Spinal cord cultures contained numerous large, phase bright, processbearing cells with a multipolar appearance (Fig. 1A). Generally about 1% of the cells in culture had this morphology and were identified as neurons using neuronal markers such as NF (Figs. 1A and B), NSE (Fig. lC), or AChE (Fig. ID). Neurons could be identified in cultures for at least 2 weeks. To detect the presence of MN within some of these cultures HRP was injected intramuscularly prior to spinal cord dissociation and plating. After 3 to 4 days in culture HRP was detected within M N as a black reaction product within the cytoplasm (Figs. 2A and B). Labeled cells demonstrated short processes which also contained black reaction product. Less than 0.1% of cultured cells had such labeling and all of these cells had multipolar cell bodies and processes. There was no apparent difference between the number of surviving neurons when mice o r rats were used for culturing. Older animals (day 5 and 6) produced slightly fewer surviving neurons and only newSpinal Cord Cultures.
Cultured Spinal Neurons
borns and 2 day-old animals were used for preparation of motoneuron cultures. Cultures enriched in their content of M N were prepared by cell sorting after the retrograde labeling. T o detect viable M N , as opposed to identification after fixation, M N were labeled specifically by the retrograde transport of RITC from muscle as shown in Figure 3A. Labeling was confined to the MN pool on the injected side and in the appropriate segments of the cord. T h e dissociation protocol was optimized to obtain cells having both high levels of fluorescence as well as large light scatter which corresponds to a viable labeled cell population.' 1x12215Fluorescence and light scatter ditributions were highly reproducible for experimental (Fig. 3B) and control (Fig. 3C) spinal cord tissue. Control experiments were done using segments of cord that did not contain M N innervating the injected muscles. Dissociations performed without trypsin and using only mechanical trituration produced fewer live cells and higher amounts of debris (not shown). Sorting windows were chosen to maximize the number of fluorescent cells with large light scatter and these sorted cells comprised about 10% of the total number of events analysed by the cell sorter. The sorted cells included both live MN as well as some dead cells; these two populations were distinguishable as relatively distinct peaks in the light scatter distributions when the sorted cells were reanalyzed by the FACS (Fig. 3D). Estimates from the Light scatter distributions of sorted cells as well as trypan blue Yotoneuron Cultures.
MUSCLE & NERVE
January 1991
17
D
events
t
FIGURE 3. (A) retrograde labelling of motoneurons, 25 pn cryostat section of a 4% paraformaldehyde fixed spinal cord from a 2-dayold mouse injected 1 day before with RlTC into the left brachial muscles. Fluorescence microscopy. Calibration bar, 65 pm. (6-D) Dual parameter histograms of forward light scatter and RlTC fluorescence intensity, increasing fluorescence from left to right, increasing light scatter from right to left. Number of events increasing along vertical axis in logarithmic units. Frequency histograms of lo4 events for spinal cord cells obtained from injected animals, (C) spinal cord cells from control animals, (D) sorted motoneurones each sample. (6) reanalyzed by FACS; single arrowhead, debris, and dead cells; double arrowhead, live cells.
FIGURE 4. Dissociated rnotoneurons after isolation by FACS sorting. (A) phase; (B) fluorescence photomicrographs of sorted cells after 24 hours in culture; (C) NSE irnrnunocytochemistry,cell 3 days in culture; (D)AChE cytochernistry, cell 2 days in culture, arrow indicates neurite; (E) phase; (F) irnrnunofluorescence photomicrographs of a sorted cell labelled with NF after 7 days in culture. Calibration bars, 20 pm. Cells in A, B, C, and D obtained from 2-day-old mice; E and F obtained from 3-day-old mice.
Cultured Spinal Neurons
MUSCLE & NERVE
January 1991
19
exclusion placed the number of viable sorted MN at between 75% to 90% of the total number of sorted cells (typically about 2 to 5 X lo5 from 6 animals). When plated, cells adhered to the substratum in about 6 hours, and fluorescence was detectable in virtually all the plated cells within the first 24 hours (Figs. 4A and B). After this time, the number of fluorescent M N gradually decreased, which may in part be related to a fading of the fluorochrome or MN death.3 T h e seeding efficiency was 50%, and virtually all of these M N exhibited a phase bright cell body of about 15 to 20 pm, 9% to 12% had a smooth cell surface and many possessed vacuoles. About 19%to 2% of the total cell population produced neurites, usually within 24 hours of plating when these cells were cultured on poly-L-lysine or laminin substrata. The number of viable cells having processes decreased with time but occasional viable MN were still detectable after 3 days in culture. These cells demonstrated NSE (Fig. 4C) and AChE (Fig. 4D). When cultured on feeder layers composed of rat brain astrocytes, quantification of MN survival was difficult because of the dense background. However MN were identified u p to 1 week in culture and demonstrated NF immunoreactivity (Figs. 4E and F). Neurofilament immunoreactivity was never seen on feeder layers which had not been seeded with labeled MN. DISCUSSION
The present study demonstrates that neonatal mammalian SCN and MN can survive, produce neuron specific proteins, and extend processes in monolayer culture. In the case of SC cultures, neurons can survive for at least 2 weeks and produce extensive processes. Cultures enriched in their content of MN do not appear capable of survival for more than a week when cultured under similar conditions. Keasons for this difference in survival time may relate to physical injury of M N during sorting or to the additional time needed for this process as well as requirements for trophic factors in the media. Previous investigators have observed a dramatic decline in the survival of fetal MN after only a few days in culture."""'" This decline can be re-
','
tarded by muscle-conditioned media. Improvement in the duration of MN survival is also apparent when cells are cultured on monolayers of rat brain astrocytes as was noted here and as has been noted previously for fetal MN.'* Astrocyte feeder layers may facilitate the initial adhesion of MN to the substrate as well as process extension. These results also suggest that trophic factors may be essential for the maintenance of MN in culture. Although M N trophic factors have not been completely characterized, l 7 recent work has indicated that probably a number of distinct molecules have trophic activity and may differentially regulate process outgrowth, MN survival, and cholinergic differentiation in vitro. l 4 Use of cultures enriched in MN may aid in the understanding of trophic factor action. Many unsuccessful attempts have been made to prepare cultures from the spinal cord of postnatal animals.5 T h e single published study of cultured adult mouse MN obtained by fluorescenceactivated cell sorting did not indicate any M N survival after 24 hours in culture and no mention of process formation was made.' T h e successful cultivation of postnatal SCN and MN appears to be dependent on 3 features: (1) the use of trypsin rather than mechanical trituration as a means of dissociation. (2) A short period of exposure of minced spinal cord to high concentrations of trypsin (1%) aids in the trituration process. (3) Poorer cell survival occurs when spinal cords are obtained from older animals and we have favored using animals immediately after birth. Although these techniques have permitted the growth of SCN, neonatal M N have not survived for more than one week in culture. Improvements in substrate and media composition may, however, increase the survival of these motoneurons for longer periods. T h e present observations indicate that viable neurons can be obtained from postnatal mammalian spinal cord tissue for culture purposes. Further evaluation of the potential of postnatal tissue for culturing is warranted for studies relevant to motoneuron diseases. Consideration should also be given to the possibility of culturing adult human central neurons for experimental or therapeutic purposes.
REFERENCES 1 . Calof AL, Reichardt LF: Motoneurons purified by cell sorting respond to two distinct activities in myotube-conditioned medium. Dev Bzol 1984;106:194-210.
20
Cultured Spinal Neurons
2. Eagleson KL, Bennett MR: Survival of purified motor neurones in v i m : effects of skeletal muscle-conditioned medium. Neurosci Lett 1983;38:187- 192.
MUSCLE & NERVE
January 1991
3. Fruns M, Krieger C, Sears T A : Identification and electrophysiological investigations of embryonic mammalian motoneurones in culture. Neurosci Lett 1987;83:82-88. 4. Graham RC, Karnovsky MJ: The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J HOtochem Cytochem 1966;14:291-304. 5. Kim SU: Neuronal aging in tissue and cell cultures: a review. I n Vitro 1983;19:73-82. 6. Kim SU, Stern J, Kim MW, Pleasure DE: Culture of purified rat astrocytes in serum-free medium supplemented with mitogen. Brain Res 1983;274:79-86. 7. Lance-Jones C: Motoneuron cell death in the developing lumbar spinal cord of the mouse. Den Brain Res 1982;4:473-479. 8. Lee V, Wu L, Schlaepfer WW: Monoclonal antibodies recognize individual neurofilament triplet proteins. Proc Nat Acad SCZ USA 1982;79:6089-6092. 9. O'Brien RJ, Fischbach GD: Isolation of embryonic chick motoneurons and their survival in vitro. J Neuroscz 1986;6:3265-3274. 10. Oppenheim RW: T h e absence of significant postnatal motoneuron death in the brachial and lumbar spinal cord of the rat. J Comp Neurol 1986;246:281-286. 11. St. John PA, Kell WM, Mazzetta GD, Lange (;D, Barker
Cultured Spinal Neurons
JL: Analysis and isolation of embryonic mammalian neurons by fluorescence-activated cell sorting. J Neurosti 1986;6:1492- 1512. 12. Schaffner AE, St. John PA, Barker JL: Fluorescence-activated cell sorting of embryonic mouse and rat motoneurons and their long-term survival in vitro. J Neuroscz 1987;7: 3088- 3 104. 13. Schmechel DE, Brightman MW, Barker JL: Localization of neuron-specific enolase in mouse spinal neurons grown in tissue culture. Brain Res 1980;181:391-400. 14. Smith RG, Vaca K, McManaman J, Appel SH: Selective effects of skeletal muscle extract fractions on motoneuron development in vitro. J Neuroscz 1986;6:439-447. 15. Smyrnis E, Kim SU, Kim MW, Oger J , Sylvester C, Paty DW: Fluorescence-activated cell sorter analysis of bulk-isolated porcine oligodendrocytes. J Neurozmmunol 1986;13: 47-60. 16. Tago H, Kimura H, hfaeda T: Visualization of detailed acetylcholinesterase fiber and neuron staining in rat brain by a sensitive histochemical procedure. J Histoehem Cytochem 1986;34:1431- 1438. 17. Varon S, Manthorpe M, Longo FM: Growth factors and motoneurons, in Rowland LAP (ed): Advances in Nuurol00. New York, Raven Press, 1982, vol 36, pp 453471.
MUSCLE & NERVE
January 1991
21
Previously, embryonic tissues have been used to produce monolayer cultures containing mammalian spinal cord neurons (SCN) and motoneurons (MN) for studies of the pathophysiology of motoneuron diseases. We demonstrate here that viable SCN and MN were observed in dissociated cultures from neonatal rat and mouse. These SCN and MN produced neurites and expressed acetylcholinesterase,neuron-specific enolase and neurofilament protein. These results indicate that cultured postnatal SCN and MN are capable of survival, neurite extension, and phenotypic expression in culture. Key words: spinal cord neuron motoneuron cell culture fluorescence activated cell sorting MUSCLE & NERVE 14~14-21 1991
NEONATAL MAMMALIAN SPINAL CORD NEURONS AND MOTONEURONS IN MONOLAYER CULTURE CHARLES KRIEGER, MD, PhD, FRCP(C) and SEUNG U. KIM, MD, PhD
T h e pathophysiology of diseases affecting motoneurons (MN) are poorly understood. The technique of neural tissue culture permits the study of M N in a controlled environment and has been used for studies relevant to motoneuron diseases. 14,17 Neural tissue culture involves the mechanical or enzymatic dissociation of neural tissue into cells which are then maintained in vitro. Spinal cord cultures (SC cultures) are often monolayer cultures containing neurons and non-neuronal cellular elements of the SC which include ciand y-motoneurons, interneurons, and dorsal horn neurons. l 7 Recently, cultures enriched in their content of motoneurons (MN cultures) have been developed using the retrograde labeling of M N by fluorescent markers followed by cell sorting and these cultures consist largely of ci- and y~~.2,9,12 Generally, mammalian SC cultures or MN cultures have been prepared from embryonic material. Only a few studies have examined the viabil-
From the Division of Neurology, Department of Medicine, Health Sciences Centre Hospital, University of British Columbia, Vancouver, British Columbia, Canada. Supported by MRC (Canada) and the ALS Society of Canada. We express our thanks to M. Kim, D. Osborne, and D Zecchini. Address reprint requests to C. Krieger, Division of Neurology, Department of Medicine, Health Sciences Centre Hospital, University of British Columbia, 221 1 Westbrook Mall, Vancouver, B.C , V6T 1W5, Canada Accepted for publication November 30, 1989. CCC 0148-639X/91/01014-08 $04 00 0 1991 John Wiley & Sons, Inc
14
Cultured Spinal Neurons
ity or phenotypic expression of monolayer cultures of postnatal spinal neurons"" and these have met with little success. An evaluation of the viability of postnatal spinal cord neurons (SCN) and MN in culture is of importance as embryonic neurons differ from postnatal neurons with the most notable difference being a period of naturally occurring cell death during embryogenesi~~; in the mouse this be ins at embryonic day 13 and is completd at birth! Since there is no significant postnatal motoneuron cell death in the rat" or mouse7 investigations of cell death in neurons from neonatal animals may be more relevant to an understanding of the mechanism of cell death in adult human neurons than comparable studies using embryonic cells. In this report we demonstrate that cultures of spinal cord cells from neonatal rodents can be prepared which contain viable SCN and MN. To confirm the presence of M N in these cultures we have retrograde-labeled these cells in situ followed by dissociation and fluorescence-activated cell sorting. MATERIALS AND METHODS
Neonatal rats (Sprague-Dawley) or mice (CU-1), aged between 1 to 6 days were cold anesthetized, decapitated, and the spinal cords were removed. Cords were minced, incubated in 1.0% trypsin (Gibco, Grand Island, NY) and 0.01% DNAse (Sigma, St. Louis, MO) in Dulbecco's phosphate buffered saline with 10 mg % D-glucose (PBS-D) for 10 min at 37°C followed Preparation of S c Cultures.
MUSCLE & NERVE
January 1991
Cultured Spinal Neurons
MUSCLE & NERVE
January 1991
15
FIGURE 1. (Opposite page and above) Murine spinal cord neurons from newborn animals in culture. (A) phase, (B) NF immunofluorescence, (C) NSE immunocytochernistry, (D)AchE histochemistry. Calibration bars, 20 wm.
by incubation in 0.1% trypsin for 10 min. Dissociated cells were obtained by repeated gentle trituration. T h e supernatant was collected in a tube containing 0.6 mL of heat-inactivated horse serum (Gibco), passed through a 50 pm nylon mesh filter, and concentrated by centifugation at 800 rpm for 10 min. Cells were then washed in PBS-D, suspended in feeding medium and plated onto poly-l-lysinecoated coverslips in concentrations of 2 to 8 x lo6 cells/mL of which over 90% were viable as determined by trypan blue exclusion. The feeding medium consisted of 90% Eagle’s minimum essential medium with 10% fetal calf serum (Gibco), 5 mgi mL D-glucose, 50 pg/mL streptomycin, and 50 U/ mL penicillin. Coverslips were placed in dishes maintained at 36°C in a 5% co2 incubator with media changes twice per week. Preparation of Cultures Containing Identified MN (MN Cultures). For preparation of cultures en-
riched in their content of M N or containing identified MN all 4 extremities of cold anesthetized neonatal or 2 day-old animals were injected with 3.0 to 4.0 pL of 0.4% rhodamine isothiocyanate (RITC, Sigma) or 10% horseradish peroxidase (HRP, Toyobo, Osaka, Japan). Injections were made into the muscle masses of the proximal portions of each limb using a Hamilton microsyringe (Reno, NV).
16
Cultured Spinal Neurons
After 24 to 48 hours had elapsed, spinal cords were removed and dissociated as described above. In some experiments the cell suspension was then sorted using a Becton-Dickinson FAGS IV flow cytometer operating at 488 nm for light scatter and 568 nm for fluorescence measurements with a power output of 300 mW.I5 Cells were analyzed at rates of 500 to 1000 eventsisec, where events constituted live and dead cells as well as cell fragments. Sorted cells were collected in sterile glass tubes containing feeding medium and were plated onto glass coverslips in Petri dishes coated with poly-L-lysine alone, poly-L-lysine and EHS laminin (BKL, Bethesda, MD), or feeder layers of rat brain astrocytes6 After 2 to 14 days in vitro, cultures were examined for the presence of neurofilament protein (NF), neuronspecific enolase (NSE), and acetylcholinesterase (AChE). NF8 immunoreaction was performed by fixation in acid alcohol and incubation with a primary antibody and rhodamine or fluorescein-conjugated second antibodies (Cappel, West Chester, PA) used at 1 : 100 dilution. NSEI3 was detected after fixation in 4% paraformaldehyde and incubation in primary antibody using a biotin-avidin method (Vectastain rabbit IgG ABC kit, Vector Labs, Burlingame, CA). AChE was determined as described previously16 but with the addition of 24 Cytochemistry and Immunocytochemistry.
MUSCLE & NERVE
January 1991
FIGURE 2. (A) and (6):rnotoneurons labelled by retrograde transport of HRP from muscle. Cells from 2-day-old mice plated 3 days prior to fixation. Calibration bars 20 p.
p M tetraisopropylpyrophosphoramide (Sigma) to block pseudocholinesterase activity. The presence of intracellular HRP was detected using the diaminobenzidine method after paraformaldehyde f i ~ a t i o n .Monoclonal ~ antibody specific for 1501 200 kda neurofilament triplet protein was provided by Dr. V. Lee' and rabbit antiserum monospecific for NSE was obtained from Dako (Santa Barbara, CA). RESULTS
Spinal cord cultures contained numerous large, phase bright, processbearing cells with a multipolar appearance (Fig. 1A). Generally about 1% of the cells in culture had this morphology and were identified as neurons using neuronal markers such as NF (Figs. 1A and B), NSE (Fig. lC), or AChE (Fig. ID). Neurons could be identified in cultures for at least 2 weeks. To detect the presence of MN within some of these cultures HRP was injected intramuscularly prior to spinal cord dissociation and plating. After 3 to 4 days in culture HRP was detected within M N as a black reaction product within the cytoplasm (Figs. 2A and B). Labeled cells demonstrated short processes which also contained black reaction product. Less than 0.1% of cultured cells had such labeling and all of these cells had multipolar cell bodies and processes. There was no apparent difference between the number of surviving neurons when mice o r rats were used for culturing. Older animals (day 5 and 6) produced slightly fewer surviving neurons and only newSpinal Cord Cultures.
Cultured Spinal Neurons
borns and 2 day-old animals were used for preparation of motoneuron cultures. Cultures enriched in their content of M N were prepared by cell sorting after the retrograde labeling. T o detect viable M N , as opposed to identification after fixation, M N were labeled specifically by the retrograde transport of RITC from muscle as shown in Figure 3A. Labeling was confined to the MN pool on the injected side and in the appropriate segments of the cord. T h e dissociation protocol was optimized to obtain cells having both high levels of fluorescence as well as large light scatter which corresponds to a viable labeled cell population.' 1x12215Fluorescence and light scatter ditributions were highly reproducible for experimental (Fig. 3B) and control (Fig. 3C) spinal cord tissue. Control experiments were done using segments of cord that did not contain M N innervating the injected muscles. Dissociations performed without trypsin and using only mechanical trituration produced fewer live cells and higher amounts of debris (not shown). Sorting windows were chosen to maximize the number of fluorescent cells with large light scatter and these sorted cells comprised about 10% of the total number of events analysed by the cell sorter. The sorted cells included both live MN as well as some dead cells; these two populations were distinguishable as relatively distinct peaks in the light scatter distributions when the sorted cells were reanalyzed by the FACS (Fig. 3D). Estimates from the Light scatter distributions of sorted cells as well as trypan blue Yotoneuron Cultures.
MUSCLE & NERVE
January 1991
17
D
events
t
FIGURE 3. (A) retrograde labelling of motoneurons, 25 pn cryostat section of a 4% paraformaldehyde fixed spinal cord from a 2-dayold mouse injected 1 day before with RlTC into the left brachial muscles. Fluorescence microscopy. Calibration bar, 65 pm. (6-D) Dual parameter histograms of forward light scatter and RlTC fluorescence intensity, increasing fluorescence from left to right, increasing light scatter from right to left. Number of events increasing along vertical axis in logarithmic units. Frequency histograms of lo4 events for spinal cord cells obtained from injected animals, (C) spinal cord cells from control animals, (D) sorted motoneurones each sample. (6) reanalyzed by FACS; single arrowhead, debris, and dead cells; double arrowhead, live cells.
FIGURE 4. Dissociated rnotoneurons after isolation by FACS sorting. (A) phase; (B) fluorescence photomicrographs of sorted cells after 24 hours in culture; (C) NSE irnrnunocytochemistry,cell 3 days in culture; (D)AChE cytochernistry, cell 2 days in culture, arrow indicates neurite; (E) phase; (F) irnrnunofluorescence photomicrographs of a sorted cell labelled with NF after 7 days in culture. Calibration bars, 20 pm. Cells in A, B, C, and D obtained from 2-day-old mice; E and F obtained from 3-day-old mice.
Cultured Spinal Neurons
MUSCLE & NERVE
January 1991
19
exclusion placed the number of viable sorted MN at between 75% to 90% of the total number of sorted cells (typically about 2 to 5 X lo5 from 6 animals). When plated, cells adhered to the substratum in about 6 hours, and fluorescence was detectable in virtually all the plated cells within the first 24 hours (Figs. 4A and B). After this time, the number of fluorescent M N gradually decreased, which may in part be related to a fading of the fluorochrome or MN death.3 T h e seeding efficiency was 50%, and virtually all of these M N exhibited a phase bright cell body of about 15 to 20 pm, 9% to 12% had a smooth cell surface and many possessed vacuoles. About 19%to 2% of the total cell population produced neurites, usually within 24 hours of plating when these cells were cultured on poly-L-lysine or laminin substrata. The number of viable cells having processes decreased with time but occasional viable MN were still detectable after 3 days in culture. These cells demonstrated NSE (Fig. 4C) and AChE (Fig. 4D). When cultured on feeder layers composed of rat brain astrocytes, quantification of MN survival was difficult because of the dense background. However MN were identified u p to 1 week in culture and demonstrated NF immunoreactivity (Figs. 4E and F). Neurofilament immunoreactivity was never seen on feeder layers which had not been seeded with labeled MN. DISCUSSION
The present study demonstrates that neonatal mammalian SCN and MN can survive, produce neuron specific proteins, and extend processes in monolayer culture. In the case of SC cultures, neurons can survive for at least 2 weeks and produce extensive processes. Cultures enriched in their content of MN do not appear capable of survival for more than a week when cultured under similar conditions. Keasons for this difference in survival time may relate to physical injury of M N during sorting or to the additional time needed for this process as well as requirements for trophic factors in the media. Previous investigators have observed a dramatic decline in the survival of fetal MN after only a few days in culture."""'" This decline can be re-
','
tarded by muscle-conditioned media. Improvement in the duration of MN survival is also apparent when cells are cultured on monolayers of rat brain astrocytes as was noted here and as has been noted previously for fetal MN.'* Astrocyte feeder layers may facilitate the initial adhesion of MN to the substrate as well as process extension. These results also suggest that trophic factors may be essential for the maintenance of MN in culture. Although M N trophic factors have not been completely characterized, l 7 recent work has indicated that probably a number of distinct molecules have trophic activity and may differentially regulate process outgrowth, MN survival, and cholinergic differentiation in vitro. l 4 Use of cultures enriched in MN may aid in the understanding of trophic factor action. Many unsuccessful attempts have been made to prepare cultures from the spinal cord of postnatal animals.5 T h e single published study of cultured adult mouse MN obtained by fluorescenceactivated cell sorting did not indicate any M N survival after 24 hours in culture and no mention of process formation was made.' T h e successful cultivation of postnatal SCN and MN appears to be dependent on 3 features: (1) the use of trypsin rather than mechanical trituration as a means of dissociation. (2) A short period of exposure of minced spinal cord to high concentrations of trypsin (1%) aids in the trituration process. (3) Poorer cell survival occurs when spinal cords are obtained from older animals and we have favored using animals immediately after birth. Although these techniques have permitted the growth of SCN, neonatal M N have not survived for more than one week in culture. Improvements in substrate and media composition may, however, increase the survival of these motoneurons for longer periods. T h e present observations indicate that viable neurons can be obtained from postnatal mammalian spinal cord tissue for culture purposes. Further evaluation of the potential of postnatal tissue for culturing is warranted for studies relevant to motoneuron diseases. Consideration should also be given to the possibility of culturing adult human central neurons for experimental or therapeutic purposes.
REFERENCES 1 . Calof AL, Reichardt LF: Motoneurons purified by cell sorting respond to two distinct activities in myotube-conditioned medium. Dev Bzol 1984;106:194-210.
20
Cultured Spinal Neurons
2. Eagleson KL, Bennett MR: Survival of purified motor neurones in v i m : effects of skeletal muscle-conditioned medium. Neurosci Lett 1983;38:187- 192.
MUSCLE & NERVE
January 1991
3. Fruns M, Krieger C, Sears T A : Identification and electrophysiological investigations of embryonic mammalian motoneurones in culture. Neurosci Lett 1987;83:82-88. 4. Graham RC, Karnovsky MJ: The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J HOtochem Cytochem 1966;14:291-304. 5. Kim SU: Neuronal aging in tissue and cell cultures: a review. I n Vitro 1983;19:73-82. 6. Kim SU, Stern J, Kim MW, Pleasure DE: Culture of purified rat astrocytes in serum-free medium supplemented with mitogen. Brain Res 1983;274:79-86. 7. Lance-Jones C: Motoneuron cell death in the developing lumbar spinal cord of the mouse. Den Brain Res 1982;4:473-479. 8. Lee V, Wu L, Schlaepfer WW: Monoclonal antibodies recognize individual neurofilament triplet proteins. Proc Nat Acad SCZ USA 1982;79:6089-6092. 9. O'Brien RJ, Fischbach GD: Isolation of embryonic chick motoneurons and their survival in vitro. J Neuroscz 1986;6:3265-3274. 10. Oppenheim RW: T h e absence of significant postnatal motoneuron death in the brachial and lumbar spinal cord of the rat. J Comp Neurol 1986;246:281-286. 11. St. John PA, Kell WM, Mazzetta GD, Lange (;D, Barker
Cultured Spinal Neurons
JL: Analysis and isolation of embryonic mammalian neurons by fluorescence-activated cell sorting. J Neurosti 1986;6:1492- 1512. 12. Schaffner AE, St. John PA, Barker JL: Fluorescence-activated cell sorting of embryonic mouse and rat motoneurons and their long-term survival in vitro. J Neuroscz 1987;7: 3088- 3 104. 13. Schmechel DE, Brightman MW, Barker JL: Localization of neuron-specific enolase in mouse spinal neurons grown in tissue culture. Brain Res 1980;181:391-400. 14. Smith RG, Vaca K, McManaman J, Appel SH: Selective effects of skeletal muscle extract fractions on motoneuron development in vitro. J Neuroscz 1986;6:439-447. 15. Smyrnis E, Kim SU, Kim MW, Oger J , Sylvester C, Paty DW: Fluorescence-activated cell sorter analysis of bulk-isolated porcine oligodendrocytes. J Neurozmmunol 1986;13: 47-60. 16. Tago H, Kimura H, hfaeda T: Visualization of detailed acetylcholinesterase fiber and neuron staining in rat brain by a sensitive histochemical procedure. J Histoehem Cytochem 1986;34:1431- 1438. 17. Varon S, Manthorpe M, Longo FM: Growth factors and motoneurons, in Rowland LAP (ed): Advances in Nuurol00. New York, Raven Press, 1982, vol 36, pp 453471.
MUSCLE & NERVE
January 1991
21
