- Email: [email protected]
Synapse formation and spontaneous activity in rat brainstem neurons in primary culture
Developmental Brain Research 117 Ž1999. 31–38 www.elsevier.comrlocaterbres
Research report
Synapse formation and spontaneous activity in rat brainstem neurons in primary culture Romain Guinamard 1, Eric Delpy, Jean-Pierre Denizot, Thierry Didier Jacquin
)
Laboratoire de biologie fonctionnelle du neurone, Institut Alfred Fessard, CNRS, 1 AÕ de la Terrasse, 91198 Gif sur YÕette Cx, France Accepted 29 June 1999
Abstract The correlation between synaptogenesis and onset of spontaneous action potentials was assessed in rat brainstem cells up to 29 days in primary culture. Cells exhibited different stages of maturation followed by electron microscopy and patch clamp recordings. Terminal boutons with no preferential orientation of presynaptic vesicles appeared after 2 days in culture. After 5 days, preferential orientation of presynaptic vesicles and thickening of postsynaptic membranes were observed. The spontaneous discharge of action potentials, single or bursting, was observed after 7 days in vitro. This was followed by the expression of a 128-pS Kq channel starting at 13 days in vitro. A 69-pS Kq channel was also present throughout the duration of the cultures. These results suggest that spontaneous discharge of action potentials does not occur before synapses are formed and Kq channel types develop differentially in brainstem neurons in vitro. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Primary culture; Neuronal differentiation; Synapse; Potassium channel expression
1. Introduction Electrical activity influences several aspects of cellular differentiation in spinal cord, retina and muscle, including neurite extension w8,23,31x and refinement of neuronal connections w6x. However, neuronal activity is not necessary at the first stages of development for orchestrating the formation of neuronal connections in these structures w6x. In the rat brainstem network, the correlation of synaptogenesis and spontaneous electrical activity has been partially studied, using primary cultures. Neuromodulators are expressed at early stages w1,11,21x, which correlates to that observed in vivo from embryonic day 14 to birth w10,12,13,15,25x, and neuromodulators can regulate the discharge of action potentials in vitro w20,29x. Evoked action potentials can be induced by current injection after 3 days in culture w1x. Also, spontaneous action potentials occur independently of synaptic connections w18x. However, no information is available on synaptic differentiation in relation to the onset of the spontaneous discharge of
)
Corresponding author. Fax: q33-1-69-07-05-38; E-mail: [email protected] 1 Present address: UMR 6558 LBSC, Universite´ de Poitiers, 40 Av du recteur Pineau, 86022 Poitiers CX, France.
action potentials. The aim of the present study was to assess if spontaneous action potentials induce synaptic connections in rat brainstem neurons in primary culture by following distinct morphological and physiological stages of maturation.
2. Materials and methods 2.1. Cell culture Culture technique was as previously described w11x, but without the addition of antibiotics. All experiments were performed according to national ethical guidelines ŽFrench Ministry of Agriculture.. Fourteen- or fifteen-day embryos were obtained from time mated pregnant rats Žalbino Sprague–Dawley, Iffa Credo.. The females were anesthetized with carbon dioxide Žissue from dry ice. and underwent cervical dislocation before the embryos were removed under aseptic conditions. Embryo brainstems were isolated in 0.1 M phosphate-buffered saline ŽPBS. at 48C and the dura was removed. The brainstems were finely minced with iridectomy scissors and gently triturated using a fire-polished pasteur pipette. An aliquot of the suspension was placed on a Malasez cell. The viable cells were
0165-3806r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 3 8 0 6 Ž 9 9 . 0 0 0 9 4 - 2
32
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
estimated using Trypan Blue exclusion and plated at 10 6 per dish. All cell types were plated to allow the formation of synaptic connections as they occur in vivo. They were cultivated in poly-L-ornithine Ž15 mgrml; MW s 40,000.coated tissue culture plastic dishes Ž35 mm, Falcon, Oxnard, CA., prefilled with a 1:1 mixture of Dulbecco’s modified Eagle’s medium ŽGibco, Grand Island, NY. and Ham F-12 nutrient Ž6 grl., glutamine Ž10 mM., sodium bicarbonate Ž13 mM., glucose Ž30 mM. and HEPES buffer Ž5 mM.. The mixture was supplemented with 10% horse serum and 10% fetal calf serum ŽGibco. during the first 2 days of plating and changed and supplemented with 10% horse serum twice weekly. Brief treatment with 5-fluorodeoxyuridine Ž0.1%; third and fourth day after plating. retarded the growth of non-neuronal cells. Dishes were kept at 378C with a constant moist 95% air–5% CO 2 environment. 2.2. Electronic microscopy of cell cultures After 2 to 20 days in culture, cells were fixed with 2% glutaraldehyde in 0.1 M phosphate buffer ŽPB., pH 7.4, for 3 h, rinsed in 0.1 M PB, post-fixed with 2% osmium tetroxide in 0.1 M PB for 1 h, dehydrated in an ethanol series and embedded in a thin layer of Spurr w28x. Fields were randomly chosen in culture sets for electron microscopic ŽPhilips CM10. analysis. Serial ultrathin sections were cut using an LKB ultramicrotome. They were mounted on coper grids Ž3.05 mm ø, 200 Mesh HT, Pelanne Instruments, France., although bars of the grids could prevent systematic observation of the successive sections, and stained with uranyl acetate and lead citrate w22x. Ž1. Growth cones, Ž2. terminal boutons with no preferential orientation of presynaptic vesicles and Ž3. synapses with preferential orientation of presynaptic vesicles and thickening of postsynaptic membrane were counted at different ages in culture at 15,500 = magnification obtained in three different culture sets Ž200 mm2 total surface at each developmental age.. 2.3. Neuronal recording For patch-clamp recordings, cells were switched to a standard artificial cerebro-spinal fluid ŽACSF. of the following composition Žin mM.: KCl, 3.5; NaCl, 137; KH 2 PO4 , 0.4; MgSO4 , 2; Na 2 HPO4 , 1.25; D-glucose, 10; CaCl 2 2H 2 O, 2.2 and HEPES–NaOH hemisodium, 10 ŽpH 7.4. at room temperature and superfused at a rate of 1
ml miny1 . Electrodes Ž4–7 M V . were prepared from borosilicate glass and coated with N, N-dimethyltrimethyl-silylamine ŽFluka.. They were filled with highKCl saline containing Žin mM.: KCl, 154; NaCl, 6; MgCl 2 , 2.2; glucose, 10; HEPES–NaOH 10 ŽpH 7.4.; EGTA, 10 and CaCl 2 , 6.9 to give a free Ca2q concentration of 100 nM. The use of high-KCl saline in the electrodes allowed to keep near-physiological gradients for Kq concentrations in the outside-out configuration and to establish ionic selectivity of the channels by differentiating reversal potentials for Kq, Naq and Cly ions. The amplifier RK 300 ŽBiologic. was used. However, from 0 to 2 days in culture, no gigaOhm seal could be performed successfully. Kq channel types exhibiting low inactivation were studied using successive long-duration holding potentials Žup to 5 min. combined with careful visual, manual and computer evaluation of the data. The sampling rate was 10 kHz. Recorded data ŽDAT system, Biologic. were filtered at 1–3 kHz Ž3 db cutoff; eight-pole Bessel filter; Frequency Devices. and played back for analysis onto a polygraph recorder or a computer using the 1401plus interface ŽCED, United Kingdom.. Computer analysis generated Ži. amplitude histograms for construction of I–V curves, Žii. estimates of the open probability ŽPo. and Žiii. mean open durations Žtau values. derived from records with single active channels. All presented values are mean " S.E.M. 3. Results Following the dissociation procedure of brainstem cells at embryonic day 14, cells were devoid of dendritic processes and were spherical in shape. Within 2 days of plating, cells attached to the substrate and filopodia emerged from the cell bodies. Then, the cultures could be rinsed and processed for electron microscopic analysis. Also, gigaOhm seal formation could be performed successfully for patch-clamp recordings. Neurons were identified spread out on top of a layer of glial cells and exhibited cell bodies with variable shape: fusiform, bipolar or multipolar. 3.1. Synaptogenesis in culture The fine structures of synaptic connections were observed during the course of synaptogenesis by electron microscopy. After 2 days in culture, processes and growth cones could be readily identified. A terminal bouton with no preferential orientation of presynaptic vesicles and no
Fig. 1. Electron micrographs from brainstem cells. ŽA. Contact between terminal bouton and neuronal process after 2 days in culture. Note that synaptic vesicles exhibit no preferential orientation. Scale bar s 0.5 mm. ŽB. After 5 days in culture, a terminal bouton exhibits two types of contacts on neuronal processes. One type is a synapse, enclosed within a box at higher magnification Žscale bar s 0.1 mm. in the upper right panel. It has small round or elliptical electron-luscent vesicles Žopen arrow. on the presynaptic terminal as well as thickening of postsynaptic membrane Žblack arrowhead.. The other type does not have thickening of synaptic membrane Žopen arrowhead.. Also, a growth cone makes a contact on one of the processes. Scale bar s 0.5 mm. Abbreviations: gc, growth cone; p, process; tb, terminal bouton.
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
33
34
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
thickening of synaptic membrane was also observed in one case ŽFig. 1A.. After 5 days, thickening of postsynaptic membranes and vesicles exhibiting preferential orientation on the presynaptic site could be observed in rare occasions ŽFig. 1B.. After 20 days, randomly chosen fields exhibited more compact tissue with cells organised in a well-defined network with numerous possible synaptical targets on both soma and processes. Dense core vesicles could also be observed on presynaptic sites and glycogen particles were present in both processes and soma. Contacts with no synaptic differentiation were still present Žnot shown.. Throughout the duration of the cultures, a quantitative estimation of the number of synapses at different stages of development was performed in randomly chosen fields
ŽFig. 2A.. The number of terminal boutons with no preferential orientation of presynaptic vesicles and the number of synapses with preferential orientation of presynaptic vesicles and thickening of postsynaptic membrane increased significantly with time in culture Žunpaired t-test, two-tailed p-value- 0.05.. The number of growth cones was found also to increase with time in culture although not significant. 3.2. Spontaneous discharge of action potentials in culture We investigated the occurrence of spontaneous discharge of action potentials in neurons during their develop-
Fig. 2. ŽA. Number Žmean " S.E.M.. of growth cones Ž(., terminal boutons with no preferential orientation of presynaptic vesicles Ž^. and synapses with preferential orientation of presynaptic vesicles and thickening of postsynaptic membrane ŽI. in randomly chosen fields at different stages of maturation in culture. ŽB. Percentage Ž%. of cells exhibiting spontaneous discharge of action potentials at different ages in brainstem neurons in primary culture. ŽC. Spontaneous rhythmic discharge of bursts of action potentials recorded in whole-cell configuration after 9 days in culture Žresting membrane potential; y50 mV.. Action potentials are truncated.
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
ment from 2 days in vitro on, using cell-attached or whole-cell configuration. The first spontaneous activity was observed at 7 days in culture and the percentage of brainstem neurons exhibiting action potentials increased to reach a maximum rapidly with almost all cells being active
35
at 10–14 days in culture ŽFig. 2B.. Most of the active cells exhibited single action potentials at irregular frequencies of 0.05–1 Hz, but a few cells exhibited spontaneous and regular rhythmic bursts of action potentials at 0.004–0.016 Hz ŽFig. 2C; n s 5..
Fig. 3. Activity of the 69 pS channel type recorded in an outside-out patch from a brainstem neuron at 6 days in culture. ŽA. Current traces illustrating the activity of the 69 pS channel type Ž'. at different membrane potentials indicated on the left. See openings of the 16 pS channel type Žcircle. in some traces. ŽB. Current–voltage relationship of the channels shown in A. The calculated unit conductance is 66 pS for the large channel and 16 pS for the small channel with an extrapolated Erev at y50 mV.
36
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
3.3. Single channel actiÕity in culture Unitary conductances were recorded from brainstem neurons between 2 and 29 days in culture. The most common channel type ŽFig. 3; n s 17. had a linear current–voltage relationship in the voltage range tested Žy30 mV to q30 mV. in outside-out configuration, with a conductance of 69.3 " 1.5 pS. The reversal potential extrapolated by linear regression was y54.4 " 2.1 mV, indicating a major selectivity for Kq with a permeability ratio PKqrPNaqs 16.5 according to the Goldman–Hodgkin– Katz equation. This channel type exhibited no voltage
sensitivity. It was present at all ages: the percentage of patches expressing the channel was 20% Ž4 out of 20 patches. before the period of onset of spontaneous action potentials at 7 days in vitro and 37% Ž13 out of 35 patches. at latter stages. A large channel type ŽFig. 4; n s 3. had a conductance of 128 " 15 pS in the outside-out configuration. The reversal potential extrapolated by linear regression was y70 " 10 mV Ž n s 3., indicating a major selectivity for Kq with a permeability ratio P KqrPNaqs 43.4. This channel type showed voltage sensitivity with increased activity at depolarised voltages. It was not observed before 13 days in
Fig. 4. Activity of the 128 pS channel type recorded in an outside-out patch from a brainstem neuron at 27 days in culture. ŽA. Current traces illustrating the activity of the channel at different membrane potentials indicated on the left. ŽB. Current–voltage relationship of the channel shown in A. ŽC, D, E. Effect of membrane potential on the open probability ŽPo; C., mean open times Žtau values; D. and frequency of openings ŽE.. Note the increase of activity at depolarized voltages.
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
culture out of 45 patches, while it was observed three times out of 10 patches at latter stages. A 16 " 1 pS channel type ŽFig. 3A; n s 7. showed an extrapolated reversal potential of y50 " 2 mV in the outside-out configuration, indicating a major selectivity for Kq with a permeability ratio PKqrPNaqs 13.7. This channel type was observed at all ages in culture in 14% of patches. The presence of multiple channel types in most patches prevented a detailed analysis of kinetic properties and voltage sensitivity of this channel.
4. Discussion The different stages of synaptic development and spontaneous activity were correlated in rat brainstem neurons maintained in primary culture after plating at E14 – 15 . 4.1. Synaptogenesis The sequence of morphological events during synaptogenesis began with the appearance of growth cones followed by terminal boutons with non-oriented clear vesicles 2 days after plating. Complete synaptic contacts on both soma and processes, with preferential orientation of vesicles on the presynaptic membrane and thickening of the postsynaptic membrane, were observed within 5 days after plating. The number of synapses at the different stages of development increased with time in culture suggesting that synaptogenesis continued throughout the duration of the cultures. Synaptogenesis was faster than in other culture conditions of brainstem cells without Ham F-12 nutrient and 5-fluoro-deoxyuridine w17x. In the latter experiment, elongation processes and the onset of synaptogenesis were delayed up to 1 and 2 weeks after plating, respectively. In our experiments, the chronology for synaptogenesis in brainstem neurons was similar to that observed in vivo histologically w2x and physiologically with a respiratory-related activity first appearing in 18-day-old rat embryos w5x. However, numerous synaptic connections may form randomly in primary culture between neurons arising from numerous brainstem structures and constituting a well-diversified network. 4.2. Spontaneous discharge of action potentials The discharge of action potentials appeared spontaneously in brainstem cells 7 days after plating, as previously described under the same culture conditions w1x, and continued throughout the duration of the cultures. Thus, synapse formation preceded spontaneous discharge of action potentials, suggesting that spontaneous activity is not necessary for establishment of early neuronal connections. This is similar to Purkinje neurons of the cerebellum whose main period of synapse formation was found to be concurrent with the main period of onset of spontaneous
37
activity, between 4 and 6 days in vitro w7x. This is also similar to explants of 17 days fetal mouse cerebral cortex grown in culture w4,19x. Although already started, synapses differentiation pursued in culture in medium containing xylocaine, which prevents the discharge of spontaneous and evoked action potentials, similarly to that in normal medium. After transfer of xylocaine-treated explants to normal physiological saline, electrophysiological activity was observed similar to that of non-treated explants at the same age. This confirms that spontaneous activity is not required for synapses differentiation. Although structurally mature synapses were observed after 5 days in culture, synapses may not be functional until 7 days in vitro when spontaneous discharge of action potentials was observed. However, inhibitory synapses between medullary neurons are found functional after 5 days in culture w4,14x and could delay the onset of action potentials discharge. We observed that some neurons exhibited regular bursting pacemaker-like activity after 7 days in culture spontaneously. Thus, the onset of spontaneous bursting activity also followed the development of synaptic contacts in culture. Spontaneous pacemaker activity may occur in cells of the nucleus ambiguus or the nucleus tractus solitarius as shown previously in primary cultures w24x. These neurons may activate like conditional bursters with specific intrinsic membrane properties that rely on synaptic connections as suggested for the brainstem respiratory network in vivo w3x. However, these neurons may exhibit pacemaker activity independently of synaptic connections, with intrinsic membrane properties not fully developed before 7 days in culture, the time required to observe rhythmic activity in the network. Spontaneous bursting activity, eventually occurring simultaneously in multiple axons, may increase the probability of successful transmission through synapses more reliably than single spikes w16,27x. This may reinforce neural signaling as suggested by Hebb w9x with consolidation of synapses between neurons that fire together, thus participating in synaptic plasticity during the development of neurons in culture. It is known that electrical activity influences several aspects of cellular differentiation including neurite extension w8,23,31x, and refinement of neuronal connections w6x, as may occur in our preparation with on going synaptogenesis still observed after 7 days in culture. Our data are consistent, however, with the general view that neuronal activity is not necessary, at the first stages of development, for orchestrating the formation of neural connections w6x. 4.3. ActiÕity of potassium channels We observed Kq channel types of 16, 69 and 128 pS conductance exhibiting steady-state activity during long voltage commands in brainstem neurons in primary culture. The absence of run down suggests that cytoplasmic messengers are not required for channel activity. Open
38
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
probability of the 69 pS Kq channel type did not change with variations in voltage, properties attributed to background currents w26x. The channel was present throughout the duration of the cultures. The 128 pS Kq channel type exhibited voltage sensitivity with increased activity at depolarized voltages. This channel was detected in patches from cells cultured in vitro after the onset of spontaneous discharge of action potentials, suggesting a refinement of initially crude electrical properties as previously shown on evoked action potentials in rat brainstem slice preparation w30x or primary culture w1x. Both channel types may act as negative feedback regulators of the spontaneous discharge of action potentials, and reduce excessive activity which can lead to cellular toxicity and death w23x.
w12x
w13x
w14x
w15x
w16x w17x
Acknowledgements We wish to thank H. Mc Lean and J. Teulon for helpful comments on this manuscript.
References w1x Z. Ahmed, P.S. Walker, R.E. Fellows, Properties of neurons from dissociated fetal rat brain in serum-free culture, J. Neurosci. 3 Ž12. Ž1983. 2448–2462. w2x J. Altman, S.A. Bayer, Development of the brain stem in the rat: I. Thymidine-radiographic study of the time of origin of neurons of the lower medulla, J. Comp. Neurol. 194 Ž1980. 1–35. w3x A.L. Bianchi, M. Denavit-Saubie, ´ J. Champagnat, Central control of breathing in mammals: neuronal circuitry, membrane properties and neurotransmitters, Phys. Rev. 75 Ž1995. 1–45. w4x S.M. Crain, Neurophysiologic Studies in Tissue Culture, Raven Press, New York, 1976. w5x E. Di Pasquale, F. Tell, R. Monteau, G. Hilaire, Perinatal developmental changes in respiratory activity of medullary and spinal neurons: an in Õitro study on fetal and newborn rats, Dev. Brain Res. 91 Ž1996. 121–130. w6x C.S. Goodman, C.J. Shatz, Developmental mechanisms that generate precise patterns of neuronal connectivity, Cell 72 Ž1993. 77–98, Neuron 10 Suppl. w7x D.L. Gruol, C.L. Franklin, Morphological and physiological differentiation of purkinje neurons in cultures of rat cerebellum, J. Neurosci. 7 Ž5. Ž1987. 1271–1293. w8x X. Gu, N.C. Spitzer, Breaking the code: regulation of neuronal differentiation by spontaneous calcium transients, Dev. Neurosci. 19 Ž1997. 133–141. w9x D.O. Hebb, The Organization of Behavior: a Neurophysiological Theory, Wiley, New York, 1949. w10x S. Inagaki, M. Sakanaka, S. Shiosaka, E. Senba, K. Takatsuki, H. Takagi, Y. Kawai, H. Minagawa, M. Tohyama, Ontogeny of substance P-containing neuron system of the rat: immunohistochemical analysis: I. Forebrain and upper brain stem, Neuroscience 7 Ž1. Ž1982. 251–277. w11x T.D. Jacquin, J.P. Denizot, M. Denavit-Saubie, ´ Substance P
w18x
w19x
w20x w21x
w22x w23x w24x
w25x
w26x
w27x w28x w29x
w30x
w31x
immunoreactivity of rat brain stem neurons in primary culture, J. Neurosci. Res. 31 Ž1992. 131–135. G.M. Jonakait, L. Ni, P.D. Walker, R.P. Hart, Development of substance P ŽSP.-containing cells in the central nervous system: consequences of neurotransmitter co-localization, Prog. Neurobiol. 36 Ž1991. 1–21. J.M. Lauder, J.A. Wallace, H. Krebs, P. Petrusz, K. McCarthy, In ÕiÕo and in Õitro development of serotonergic neurons, Brain Res. Bull. 9 Ž1982. 605–625. C.A. Lewis, D.S. Faber, Giant, TTX-insensitive, inhibitory postsynaptic currents in cultured rat spinal cord and medullary neurons, J. Neurophys. 76 Ž5. Ž1996. 3341–3350. H.G.W. Lidov, M.E. Molliver, Immunohistochemical study of the development of serotonergic neurons in the rat CNS, Brain Res. Bull. 9 Ž1982. 559–604. J.E. Lisman, Bursts as a unit of neural information: making unreliable synapses reliable, Trends Neurosci. 20 Ž1. Ž1997. 38–43. E.M. Lopez, S. Peressini, M.F. Kubke, J.A. Pecci Saavedra, Ultrastructural study of reaggregated culture of brainstem: synaptogenesis, Microsc. Electron. Biol. Celular 14 Ž2. Ž1990. 89–99. S. Masuko, Y. Nakajima, S. Nakajima, K. Yamaguchi, Noradrenergic neurons from the locus ceruleus in dissociated cell culture: culture methods, morphology, and electrophysiology, J. Neurosci. 6 Ž11. Ž1986. 3229–3241. P.G. Model, M.B. Bornstein, S.M. Crain, G.D. Pappas, An electron microscopic study of the development of synapses in cultured fetal mouse cerebrum continuously exposed to xylocaine, J. Cell Biol. 49 Ž1971. 362–371. J. Okada, M. Miura, Transmitter sensitivities in oval neurons from rat brainstem cultures, Brain Res. 268 Ž1983. 377–381. I. Reisert, V. Han, S. Hartwig, G. Ahnert-Hilger, C. Pilgrim, Rapid maturation of synaptic functions of prenatal serotoninergic neurons in short-term cultures: absence of sex differences and hormone effects, Neuroscience 32 Ž1. Ž1989. 133–139. E.S. Reynolds, The use of lead citrate at high pH as an electron opaque stain in electron microscopy, J. Cell Biol. 17 Ž1963. 208–212. A.B. Ribera, Ion channel activity drives ion channel expression, J. Physiol. ŽLondon. 511 Ž3. Ž1998. 645. H. Rigatto, S.C. Fitzgerald, M.A. Willis, C. Yu, In search of the real respiratory neurons: culture of medullary fetal cells inherently sensitive to CO 2 and low pH, Biol. Neonate 65 Ž1994. 149–155. M. Sakanaka, S. Inagaki, S. Shiosaka, E. Senba, H. Takagi, K. Takatsuki, Y. Kawai, H. Iida, Y. Hara, M. Tohyama, Ontogeny of substance P-containing neuron system of the rat: immunohistochemical analysis: II. Lower brain stem, Neuroscience 7 Ž5. Ž1982. 1097–1126. P. Schweitzer, S.G. Madamba, G.R. Siggins, Somatostatin increases a voltage-insensitive Kq conductance in rat CA1 hippocampal neurons, J. Neurophysiol. 79 Ž1998. 1230–1238. N.C. Spitzer, T.J. Sejnowski, Biological information processing: bits of progress, Science 277 Ž1997. 1060–1061. A.R. Spurr, A low viscosity epoxy resin embedding medium for electron microscopy, J. Ultrastruct. Res. 26 Ž1969. 31–43. W.E. Thomas, Studies of neurotransmitter chemistry of central nervous system neurons in primary tissue culture, Life Sci. 38 Ž1986. 297–308. A. Vincent, F. Tell, Postnatal changes in electrophysiological properties of rat nucleus tractus solitarii neurons, Eur. J. Neurosci. 9 Ž1997. 1612–1624. W.T. Wong, J.R. Sanes, R.O.L. Wong, Developmentally regulated spontaneous activity in the embryonic chick retina, J. Neurosci. 18 Ž21. Ž1998. 8839–8852.
Research report
Synapse formation and spontaneous activity in rat brainstem neurons in primary culture Romain Guinamard 1, Eric Delpy, Jean-Pierre Denizot, Thierry Didier Jacquin
)
Laboratoire de biologie fonctionnelle du neurone, Institut Alfred Fessard, CNRS, 1 AÕ de la Terrasse, 91198 Gif sur YÕette Cx, France Accepted 29 June 1999
Abstract The correlation between synaptogenesis and onset of spontaneous action potentials was assessed in rat brainstem cells up to 29 days in primary culture. Cells exhibited different stages of maturation followed by electron microscopy and patch clamp recordings. Terminal boutons with no preferential orientation of presynaptic vesicles appeared after 2 days in culture. After 5 days, preferential orientation of presynaptic vesicles and thickening of postsynaptic membranes were observed. The spontaneous discharge of action potentials, single or bursting, was observed after 7 days in vitro. This was followed by the expression of a 128-pS Kq channel starting at 13 days in vitro. A 69-pS Kq channel was also present throughout the duration of the cultures. These results suggest that spontaneous discharge of action potentials does not occur before synapses are formed and Kq channel types develop differentially in brainstem neurons in vitro. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Primary culture; Neuronal differentiation; Synapse; Potassium channel expression
1. Introduction Electrical activity influences several aspects of cellular differentiation in spinal cord, retina and muscle, including neurite extension w8,23,31x and refinement of neuronal connections w6x. However, neuronal activity is not necessary at the first stages of development for orchestrating the formation of neuronal connections in these structures w6x. In the rat brainstem network, the correlation of synaptogenesis and spontaneous electrical activity has been partially studied, using primary cultures. Neuromodulators are expressed at early stages w1,11,21x, which correlates to that observed in vivo from embryonic day 14 to birth w10,12,13,15,25x, and neuromodulators can regulate the discharge of action potentials in vitro w20,29x. Evoked action potentials can be induced by current injection after 3 days in culture w1x. Also, spontaneous action potentials occur independently of synaptic connections w18x. However, no information is available on synaptic differentiation in relation to the onset of the spontaneous discharge of
)
Corresponding author. Fax: q33-1-69-07-05-38; E-mail: [email protected] 1 Present address: UMR 6558 LBSC, Universite´ de Poitiers, 40 Av du recteur Pineau, 86022 Poitiers CX, France.
action potentials. The aim of the present study was to assess if spontaneous action potentials induce synaptic connections in rat brainstem neurons in primary culture by following distinct morphological and physiological stages of maturation.
2. Materials and methods 2.1. Cell culture Culture technique was as previously described w11x, but without the addition of antibiotics. All experiments were performed according to national ethical guidelines ŽFrench Ministry of Agriculture.. Fourteen- or fifteen-day embryos were obtained from time mated pregnant rats Žalbino Sprague–Dawley, Iffa Credo.. The females were anesthetized with carbon dioxide Žissue from dry ice. and underwent cervical dislocation before the embryos were removed under aseptic conditions. Embryo brainstems were isolated in 0.1 M phosphate-buffered saline ŽPBS. at 48C and the dura was removed. The brainstems were finely minced with iridectomy scissors and gently triturated using a fire-polished pasteur pipette. An aliquot of the suspension was placed on a Malasez cell. The viable cells were
0165-3806r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 3 8 0 6 Ž 9 9 . 0 0 0 9 4 - 2
32
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
estimated using Trypan Blue exclusion and plated at 10 6 per dish. All cell types were plated to allow the formation of synaptic connections as they occur in vivo. They were cultivated in poly-L-ornithine Ž15 mgrml; MW s 40,000.coated tissue culture plastic dishes Ž35 mm, Falcon, Oxnard, CA., prefilled with a 1:1 mixture of Dulbecco’s modified Eagle’s medium ŽGibco, Grand Island, NY. and Ham F-12 nutrient Ž6 grl., glutamine Ž10 mM., sodium bicarbonate Ž13 mM., glucose Ž30 mM. and HEPES buffer Ž5 mM.. The mixture was supplemented with 10% horse serum and 10% fetal calf serum ŽGibco. during the first 2 days of plating and changed and supplemented with 10% horse serum twice weekly. Brief treatment with 5-fluorodeoxyuridine Ž0.1%; third and fourth day after plating. retarded the growth of non-neuronal cells. Dishes were kept at 378C with a constant moist 95% air–5% CO 2 environment. 2.2. Electronic microscopy of cell cultures After 2 to 20 days in culture, cells were fixed with 2% glutaraldehyde in 0.1 M phosphate buffer ŽPB., pH 7.4, for 3 h, rinsed in 0.1 M PB, post-fixed with 2% osmium tetroxide in 0.1 M PB for 1 h, dehydrated in an ethanol series and embedded in a thin layer of Spurr w28x. Fields were randomly chosen in culture sets for electron microscopic ŽPhilips CM10. analysis. Serial ultrathin sections were cut using an LKB ultramicrotome. They were mounted on coper grids Ž3.05 mm ø, 200 Mesh HT, Pelanne Instruments, France., although bars of the grids could prevent systematic observation of the successive sections, and stained with uranyl acetate and lead citrate w22x. Ž1. Growth cones, Ž2. terminal boutons with no preferential orientation of presynaptic vesicles and Ž3. synapses with preferential orientation of presynaptic vesicles and thickening of postsynaptic membrane were counted at different ages in culture at 15,500 = magnification obtained in three different culture sets Ž200 mm2 total surface at each developmental age.. 2.3. Neuronal recording For patch-clamp recordings, cells were switched to a standard artificial cerebro-spinal fluid ŽACSF. of the following composition Žin mM.: KCl, 3.5; NaCl, 137; KH 2 PO4 , 0.4; MgSO4 , 2; Na 2 HPO4 , 1.25; D-glucose, 10; CaCl 2 2H 2 O, 2.2 and HEPES–NaOH hemisodium, 10 ŽpH 7.4. at room temperature and superfused at a rate of 1
ml miny1 . Electrodes Ž4–7 M V . were prepared from borosilicate glass and coated with N, N-dimethyltrimethyl-silylamine ŽFluka.. They were filled with highKCl saline containing Žin mM.: KCl, 154; NaCl, 6; MgCl 2 , 2.2; glucose, 10; HEPES–NaOH 10 ŽpH 7.4.; EGTA, 10 and CaCl 2 , 6.9 to give a free Ca2q concentration of 100 nM. The use of high-KCl saline in the electrodes allowed to keep near-physiological gradients for Kq concentrations in the outside-out configuration and to establish ionic selectivity of the channels by differentiating reversal potentials for Kq, Naq and Cly ions. The amplifier RK 300 ŽBiologic. was used. However, from 0 to 2 days in culture, no gigaOhm seal could be performed successfully. Kq channel types exhibiting low inactivation were studied using successive long-duration holding potentials Žup to 5 min. combined with careful visual, manual and computer evaluation of the data. The sampling rate was 10 kHz. Recorded data ŽDAT system, Biologic. were filtered at 1–3 kHz Ž3 db cutoff; eight-pole Bessel filter; Frequency Devices. and played back for analysis onto a polygraph recorder or a computer using the 1401plus interface ŽCED, United Kingdom.. Computer analysis generated Ži. amplitude histograms for construction of I–V curves, Žii. estimates of the open probability ŽPo. and Žiii. mean open durations Žtau values. derived from records with single active channels. All presented values are mean " S.E.M. 3. Results Following the dissociation procedure of brainstem cells at embryonic day 14, cells were devoid of dendritic processes and were spherical in shape. Within 2 days of plating, cells attached to the substrate and filopodia emerged from the cell bodies. Then, the cultures could be rinsed and processed for electron microscopic analysis. Also, gigaOhm seal formation could be performed successfully for patch-clamp recordings. Neurons were identified spread out on top of a layer of glial cells and exhibited cell bodies with variable shape: fusiform, bipolar or multipolar. 3.1. Synaptogenesis in culture The fine structures of synaptic connections were observed during the course of synaptogenesis by electron microscopy. After 2 days in culture, processes and growth cones could be readily identified. A terminal bouton with no preferential orientation of presynaptic vesicles and no
Fig. 1. Electron micrographs from brainstem cells. ŽA. Contact between terminal bouton and neuronal process after 2 days in culture. Note that synaptic vesicles exhibit no preferential orientation. Scale bar s 0.5 mm. ŽB. After 5 days in culture, a terminal bouton exhibits two types of contacts on neuronal processes. One type is a synapse, enclosed within a box at higher magnification Žscale bar s 0.1 mm. in the upper right panel. It has small round or elliptical electron-luscent vesicles Žopen arrow. on the presynaptic terminal as well as thickening of postsynaptic membrane Žblack arrowhead.. The other type does not have thickening of synaptic membrane Žopen arrowhead.. Also, a growth cone makes a contact on one of the processes. Scale bar s 0.5 mm. Abbreviations: gc, growth cone; p, process; tb, terminal bouton.
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
33
34
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
thickening of synaptic membrane was also observed in one case ŽFig. 1A.. After 5 days, thickening of postsynaptic membranes and vesicles exhibiting preferential orientation on the presynaptic site could be observed in rare occasions ŽFig. 1B.. After 20 days, randomly chosen fields exhibited more compact tissue with cells organised in a well-defined network with numerous possible synaptical targets on both soma and processes. Dense core vesicles could also be observed on presynaptic sites and glycogen particles were present in both processes and soma. Contacts with no synaptic differentiation were still present Žnot shown.. Throughout the duration of the cultures, a quantitative estimation of the number of synapses at different stages of development was performed in randomly chosen fields
ŽFig. 2A.. The number of terminal boutons with no preferential orientation of presynaptic vesicles and the number of synapses with preferential orientation of presynaptic vesicles and thickening of postsynaptic membrane increased significantly with time in culture Žunpaired t-test, two-tailed p-value- 0.05.. The number of growth cones was found also to increase with time in culture although not significant. 3.2. Spontaneous discharge of action potentials in culture We investigated the occurrence of spontaneous discharge of action potentials in neurons during their develop-
Fig. 2. ŽA. Number Žmean " S.E.M.. of growth cones Ž(., terminal boutons with no preferential orientation of presynaptic vesicles Ž^. and synapses with preferential orientation of presynaptic vesicles and thickening of postsynaptic membrane ŽI. in randomly chosen fields at different stages of maturation in culture. ŽB. Percentage Ž%. of cells exhibiting spontaneous discharge of action potentials at different ages in brainstem neurons in primary culture. ŽC. Spontaneous rhythmic discharge of bursts of action potentials recorded in whole-cell configuration after 9 days in culture Žresting membrane potential; y50 mV.. Action potentials are truncated.
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
ment from 2 days in vitro on, using cell-attached or whole-cell configuration. The first spontaneous activity was observed at 7 days in culture and the percentage of brainstem neurons exhibiting action potentials increased to reach a maximum rapidly with almost all cells being active
35
at 10–14 days in culture ŽFig. 2B.. Most of the active cells exhibited single action potentials at irregular frequencies of 0.05–1 Hz, but a few cells exhibited spontaneous and regular rhythmic bursts of action potentials at 0.004–0.016 Hz ŽFig. 2C; n s 5..
Fig. 3. Activity of the 69 pS channel type recorded in an outside-out patch from a brainstem neuron at 6 days in culture. ŽA. Current traces illustrating the activity of the 69 pS channel type Ž'. at different membrane potentials indicated on the left. See openings of the 16 pS channel type Žcircle. in some traces. ŽB. Current–voltage relationship of the channels shown in A. The calculated unit conductance is 66 pS for the large channel and 16 pS for the small channel with an extrapolated Erev at y50 mV.
36
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
3.3. Single channel actiÕity in culture Unitary conductances were recorded from brainstem neurons between 2 and 29 days in culture. The most common channel type ŽFig. 3; n s 17. had a linear current–voltage relationship in the voltage range tested Žy30 mV to q30 mV. in outside-out configuration, with a conductance of 69.3 " 1.5 pS. The reversal potential extrapolated by linear regression was y54.4 " 2.1 mV, indicating a major selectivity for Kq with a permeability ratio PKqrPNaqs 16.5 according to the Goldman–Hodgkin– Katz equation. This channel type exhibited no voltage
sensitivity. It was present at all ages: the percentage of patches expressing the channel was 20% Ž4 out of 20 patches. before the period of onset of spontaneous action potentials at 7 days in vitro and 37% Ž13 out of 35 patches. at latter stages. A large channel type ŽFig. 4; n s 3. had a conductance of 128 " 15 pS in the outside-out configuration. The reversal potential extrapolated by linear regression was y70 " 10 mV Ž n s 3., indicating a major selectivity for Kq with a permeability ratio P KqrPNaqs 43.4. This channel type showed voltage sensitivity with increased activity at depolarised voltages. It was not observed before 13 days in
Fig. 4. Activity of the 128 pS channel type recorded in an outside-out patch from a brainstem neuron at 27 days in culture. ŽA. Current traces illustrating the activity of the channel at different membrane potentials indicated on the left. ŽB. Current–voltage relationship of the channel shown in A. ŽC, D, E. Effect of membrane potential on the open probability ŽPo; C., mean open times Žtau values; D. and frequency of openings ŽE.. Note the increase of activity at depolarized voltages.
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
culture out of 45 patches, while it was observed three times out of 10 patches at latter stages. A 16 " 1 pS channel type ŽFig. 3A; n s 7. showed an extrapolated reversal potential of y50 " 2 mV in the outside-out configuration, indicating a major selectivity for Kq with a permeability ratio PKqrPNaqs 13.7. This channel type was observed at all ages in culture in 14% of patches. The presence of multiple channel types in most patches prevented a detailed analysis of kinetic properties and voltage sensitivity of this channel.
4. Discussion The different stages of synaptic development and spontaneous activity were correlated in rat brainstem neurons maintained in primary culture after plating at E14 – 15 . 4.1. Synaptogenesis The sequence of morphological events during synaptogenesis began with the appearance of growth cones followed by terminal boutons with non-oriented clear vesicles 2 days after plating. Complete synaptic contacts on both soma and processes, with preferential orientation of vesicles on the presynaptic membrane and thickening of the postsynaptic membrane, were observed within 5 days after plating. The number of synapses at the different stages of development increased with time in culture suggesting that synaptogenesis continued throughout the duration of the cultures. Synaptogenesis was faster than in other culture conditions of brainstem cells without Ham F-12 nutrient and 5-fluoro-deoxyuridine w17x. In the latter experiment, elongation processes and the onset of synaptogenesis were delayed up to 1 and 2 weeks after plating, respectively. In our experiments, the chronology for synaptogenesis in brainstem neurons was similar to that observed in vivo histologically w2x and physiologically with a respiratory-related activity first appearing in 18-day-old rat embryos w5x. However, numerous synaptic connections may form randomly in primary culture between neurons arising from numerous brainstem structures and constituting a well-diversified network. 4.2. Spontaneous discharge of action potentials The discharge of action potentials appeared spontaneously in brainstem cells 7 days after plating, as previously described under the same culture conditions w1x, and continued throughout the duration of the cultures. Thus, synapse formation preceded spontaneous discharge of action potentials, suggesting that spontaneous activity is not necessary for establishment of early neuronal connections. This is similar to Purkinje neurons of the cerebellum whose main period of synapse formation was found to be concurrent with the main period of onset of spontaneous
37
activity, between 4 and 6 days in vitro w7x. This is also similar to explants of 17 days fetal mouse cerebral cortex grown in culture w4,19x. Although already started, synapses differentiation pursued in culture in medium containing xylocaine, which prevents the discharge of spontaneous and evoked action potentials, similarly to that in normal medium. After transfer of xylocaine-treated explants to normal physiological saline, electrophysiological activity was observed similar to that of non-treated explants at the same age. This confirms that spontaneous activity is not required for synapses differentiation. Although structurally mature synapses were observed after 5 days in culture, synapses may not be functional until 7 days in vitro when spontaneous discharge of action potentials was observed. However, inhibitory synapses between medullary neurons are found functional after 5 days in culture w4,14x and could delay the onset of action potentials discharge. We observed that some neurons exhibited regular bursting pacemaker-like activity after 7 days in culture spontaneously. Thus, the onset of spontaneous bursting activity also followed the development of synaptic contacts in culture. Spontaneous pacemaker activity may occur in cells of the nucleus ambiguus or the nucleus tractus solitarius as shown previously in primary cultures w24x. These neurons may activate like conditional bursters with specific intrinsic membrane properties that rely on synaptic connections as suggested for the brainstem respiratory network in vivo w3x. However, these neurons may exhibit pacemaker activity independently of synaptic connections, with intrinsic membrane properties not fully developed before 7 days in culture, the time required to observe rhythmic activity in the network. Spontaneous bursting activity, eventually occurring simultaneously in multiple axons, may increase the probability of successful transmission through synapses more reliably than single spikes w16,27x. This may reinforce neural signaling as suggested by Hebb w9x with consolidation of synapses between neurons that fire together, thus participating in synaptic plasticity during the development of neurons in culture. It is known that electrical activity influences several aspects of cellular differentiation including neurite extension w8,23,31x, and refinement of neuronal connections w6x, as may occur in our preparation with on going synaptogenesis still observed after 7 days in culture. Our data are consistent, however, with the general view that neuronal activity is not necessary, at the first stages of development, for orchestrating the formation of neural connections w6x. 4.3. ActiÕity of potassium channels We observed Kq channel types of 16, 69 and 128 pS conductance exhibiting steady-state activity during long voltage commands in brainstem neurons in primary culture. The absence of run down suggests that cytoplasmic messengers are not required for channel activity. Open
38
R. Guinamard et al.r DeÕelopmental Brain Research 117 (1999) 31–38
probability of the 69 pS Kq channel type did not change with variations in voltage, properties attributed to background currents w26x. The channel was present throughout the duration of the cultures. The 128 pS Kq channel type exhibited voltage sensitivity with increased activity at depolarized voltages. This channel was detected in patches from cells cultured in vitro after the onset of spontaneous discharge of action potentials, suggesting a refinement of initially crude electrical properties as previously shown on evoked action potentials in rat brainstem slice preparation w30x or primary culture w1x. Both channel types may act as negative feedback regulators of the spontaneous discharge of action potentials, and reduce excessive activity which can lead to cellular toxicity and death w23x.
w12x
w13x
w14x
w15x
w16x w17x
Acknowledgements We wish to thank H. Mc Lean and J. Teulon for helpful comments on this manuscript.
References w1x Z. Ahmed, P.S. Walker, R.E. Fellows, Properties of neurons from dissociated fetal rat brain in serum-free culture, J. Neurosci. 3 Ž12. Ž1983. 2448–2462. w2x J. Altman, S.A. Bayer, Development of the brain stem in the rat: I. Thymidine-radiographic study of the time of origin of neurons of the lower medulla, J. Comp. Neurol. 194 Ž1980. 1–35. w3x A.L. Bianchi, M. Denavit-Saubie, ´ J. Champagnat, Central control of breathing in mammals: neuronal circuitry, membrane properties and neurotransmitters, Phys. Rev. 75 Ž1995. 1–45. w4x S.M. Crain, Neurophysiologic Studies in Tissue Culture, Raven Press, New York, 1976. w5x E. Di Pasquale, F. Tell, R. Monteau, G. Hilaire, Perinatal developmental changes in respiratory activity of medullary and spinal neurons: an in Õitro study on fetal and newborn rats, Dev. Brain Res. 91 Ž1996. 121–130. w6x C.S. Goodman, C.J. Shatz, Developmental mechanisms that generate precise patterns of neuronal connectivity, Cell 72 Ž1993. 77–98, Neuron 10 Suppl. w7x D.L. Gruol, C.L. Franklin, Morphological and physiological differentiation of purkinje neurons in cultures of rat cerebellum, J. Neurosci. 7 Ž5. Ž1987. 1271–1293. w8x X. Gu, N.C. Spitzer, Breaking the code: regulation of neuronal differentiation by spontaneous calcium transients, Dev. Neurosci. 19 Ž1997. 133–141. w9x D.O. Hebb, The Organization of Behavior: a Neurophysiological Theory, Wiley, New York, 1949. w10x S. Inagaki, M. Sakanaka, S. Shiosaka, E. Senba, K. Takatsuki, H. Takagi, Y. Kawai, H. Minagawa, M. Tohyama, Ontogeny of substance P-containing neuron system of the rat: immunohistochemical analysis: I. Forebrain and upper brain stem, Neuroscience 7 Ž1. Ž1982. 251–277. w11x T.D. Jacquin, J.P. Denizot, M. Denavit-Saubie, ´ Substance P
w18x
w19x
w20x w21x
w22x w23x w24x
w25x
w26x
w27x w28x w29x
w30x
w31x
immunoreactivity of rat brain stem neurons in primary culture, J. Neurosci. Res. 31 Ž1992. 131–135. G.M. Jonakait, L. Ni, P.D. Walker, R.P. Hart, Development of substance P ŽSP.-containing cells in the central nervous system: consequences of neurotransmitter co-localization, Prog. Neurobiol. 36 Ž1991. 1–21. J.M. Lauder, J.A. Wallace, H. Krebs, P. Petrusz, K. McCarthy, In ÕiÕo and in Õitro development of serotonergic neurons, Brain Res. Bull. 9 Ž1982. 605–625. C.A. Lewis, D.S. Faber, Giant, TTX-insensitive, inhibitory postsynaptic currents in cultured rat spinal cord and medullary neurons, J. Neurophys. 76 Ž5. Ž1996. 3341–3350. H.G.W. Lidov, M.E. Molliver, Immunohistochemical study of the development of serotonergic neurons in the rat CNS, Brain Res. Bull. 9 Ž1982. 559–604. J.E. Lisman, Bursts as a unit of neural information: making unreliable synapses reliable, Trends Neurosci. 20 Ž1. Ž1997. 38–43. E.M. Lopez, S. Peressini, M.F. Kubke, J.A. Pecci Saavedra, Ultrastructural study of reaggregated culture of brainstem: synaptogenesis, Microsc. Electron. Biol. Celular 14 Ž2. Ž1990. 89–99. S. Masuko, Y. Nakajima, S. Nakajima, K. Yamaguchi, Noradrenergic neurons from the locus ceruleus in dissociated cell culture: culture methods, morphology, and electrophysiology, J. Neurosci. 6 Ž11. Ž1986. 3229–3241. P.G. Model, M.B. Bornstein, S.M. Crain, G.D. Pappas, An electron microscopic study of the development of synapses in cultured fetal mouse cerebrum continuously exposed to xylocaine, J. Cell Biol. 49 Ž1971. 362–371. J. Okada, M. Miura, Transmitter sensitivities in oval neurons from rat brainstem cultures, Brain Res. 268 Ž1983. 377–381. I. Reisert, V. Han, S. Hartwig, G. Ahnert-Hilger, C. Pilgrim, Rapid maturation of synaptic functions of prenatal serotoninergic neurons in short-term cultures: absence of sex differences and hormone effects, Neuroscience 32 Ž1. Ž1989. 133–139. E.S. Reynolds, The use of lead citrate at high pH as an electron opaque stain in electron microscopy, J. Cell Biol. 17 Ž1963. 208–212. A.B. Ribera, Ion channel activity drives ion channel expression, J. Physiol. ŽLondon. 511 Ž3. Ž1998. 645. H. Rigatto, S.C. Fitzgerald, M.A. Willis, C. Yu, In search of the real respiratory neurons: culture of medullary fetal cells inherently sensitive to CO 2 and low pH, Biol. Neonate 65 Ž1994. 149–155. M. Sakanaka, S. Inagaki, S. Shiosaka, E. Senba, H. Takagi, K. Takatsuki, Y. Kawai, H. Iida, Y. Hara, M. Tohyama, Ontogeny of substance P-containing neuron system of the rat: immunohistochemical analysis: II. Lower brain stem, Neuroscience 7 Ž5. Ž1982. 1097–1126. P. Schweitzer, S.G. Madamba, G.R. Siggins, Somatostatin increases a voltage-insensitive Kq conductance in rat CA1 hippocampal neurons, J. Neurophysiol. 79 Ž1998. 1230–1238. N.C. Spitzer, T.J. Sejnowski, Biological information processing: bits of progress, Science 277 Ž1997. 1060–1061. A.R. Spurr, A low viscosity epoxy resin embedding medium for electron microscopy, J. Ultrastruct. Res. 26 Ž1969. 31–43. W.E. Thomas, Studies of neurotransmitter chemistry of central nervous system neurons in primary tissue culture, Life Sci. 38 Ž1986. 297–308. A. Vincent, F. Tell, Postnatal changes in electrophysiological properties of rat nucleus tractus solitarii neurons, Eur. J. Neurosci. 9 Ž1997. 1612–1624. W.T. Wong, J.R. Sanes, R.O.L. Wong, Developmentally regulated spontaneous activity in the embryonic chick retina, J. Neurosci. 18 Ž21. Ž1998. 8839–8852.