Neurotransmitter plasticity of cultured sympathetic neurones. Are the effects of muscle-conditioned medium reversible?

Transcription

1 Development 101, (1987) Printed in Great Britain The Company of Biologists Limited Neurotransmitter plasticity of cultured sympathetic neurones. Are the effects of muscle-conditioned medium reversible? SIMONE VIDAL, BRIGITTE RAYNAUD, DOMINIQUE CLAROUS and MICHEL J. WEBER Laboratoire de Pharmacologie el Toxicologie Fondamentales, 205 Route de Narbonne Toulouse Cedex, France Summary Muscle-conditioned medium (CM) induces choline acetyltransferase (CAT) activity in primary cultures of new-born rat sympathetic neurones and depresses the development of tyrosine hydroxylase (TOH). By following these two enzymes, we have determined whether (1) the effects of CM are reversible and (2) the neurones progressively lose their sensitivity to CM with time in culture. When neurones were cultured in the presence of 50 % CM (CM + medium), TOH activity developed slowly but CAT activity developed at a high rate. When the cultures were then switched to unconditioned medium (CM" medium), CAT activity remained elevated and continued to develop at higher rate than in cultures that were never exposed to CM. On the other hand, the switch to CM~ medium was accompanied by a transition from a low to a high rate of TOH development. CAT induction by CM was thus essentially irreversible, whereas the impairment of TOH development was fully reversible. Conversely, we studied the effects of altering CM" to CM + medium at progressively later culture days. CAT remained fully inducible for at least 2 to 3 weeks. On the other hand, TOH activity, which initially developed rapidly in CM" medium, first decreased to low levels after a switch to CM + medium and then increased again, but at a slower rate. Neuronal depolarization by elevated K + and exposure to CM have mirror-image, and antagonistic, effects on both CAT and TOH developments (Raynaud et al. 1987a). Walicke, Campenot & Patterson (1977) showed that a previous depolarization reduced the induction of cholinergic traits by a subsequent exposure to CM. We found that (1) such a depolarization only delayed the induction of CAT by several days and did not prevent the transition to a state of low TOH expression caused by CM and (2) an exposure of the cultures to elevated K + after exposure to CM did not cause a decline in CAT activity. These data thus suggest that a state of high TOH expression can superimpose on a previously induced state of elevated CAT expression, but that the induction of CAT caused by a delayed exposure to CM is accompanied by a transition from a high to a lower state of TOH expression. In addition, neuronal depolarization does not stabilize the noradrenergic phenotype in a permanent manner and can not reverse cholinergic expression of sympathetic neurones to a purely noradrenergic phenotype. Key words: neurotransmitter, plasticity, muscleconditioned medium (CM), choline acetyltransferase, sympathetic neurone, culture, rat. Introduction Although mature neurones only express a limited set of neurotransmitter-related genes, they can be induced to express new phenotypic traits under various experimental conditions. Mature sympathetic (noradrenergic) neurones can form cholinergic synapses with a high frequency under appropriate culture conditions (Wakshull, Johnson & Burton, 1979; Potter, Landis, Matsumoto & Furshpan, 1986). Conversely, parasympathetic (cholinergic) neurones from the ciliary ganglion can express catecholaminergic traits in vivo after perturbations of their environment (Le Douarin, Teillet, Ziller & Smith, 1978; Bjorklund etal. 1985; Coulombe & Bronner-Fraser, 1986; Iaco- \\tx\etal. 1985). These experiments raise the question as to whether these new phenotypic traits develop at the expense of the original phenotype of the neurones, or whether they are simply superimposed onto them. We have

2 618 S. Vidal, B. Raynaud, D. Clarous and M. J. Weber addressed this problem by using primary cultures of new-born rat sympathetic neurones. When cultured in the absence of conditioned medium (CM), these neurones develop high levels of catecholamine-synthesizing enzymes but only marginal choline acetyltransferase (CAT) activity. In the presence of CM, CAT is induced to high levels, whereas the development of catecholamine-synthesizing enzymes is depressed (Patterson & Chun, \911a,b; Swerts, Le Van Thai, Vigny & Weber, 1983; Wolinski & Patterson, 1983; Raynaud et al. 1987a,b). When sympathetic neurones are first allowed to acquire a CAT activity in the presence of CM (CM + medium), can they revert to a noradrenergic phenotype upon CM removal and lose their cholinergic characteristics? Conversely, if the development of tyrosine hydroxylase is first favoured in the absence of CM (CM~ medium), does CAT remain inducible by CM to the same extent? Under these conditions, will CAT superimpose on the previously acquired high TOH level? Neuronal depolarization by elevated K + fosters the development of TOH activity in sympathetic neurone cultures and suppresses CAT development (Walicke et al. 1977; Hefti, Gnahn, Schwab & Thoenen, 1982; Raynaud et al. 1987a,b). When added together to culture medium, CM and elevated K + exert antagonistic effects on both CAT and TOH developments. We thus wondered if cholinergic neurones would reverse to a purely noradrenergic phenotype upon CM removal and exposure to high K + medium. Conversely, we examined whether or not a preexposure to high K + medium, would make the neurones refractory- to CM action in a permanent manner. It has been shown previously that a 10-day depolarization decreased the development of cholinergic traits induced subsequently by CM (Walicke et al. 1977; Walicke & Patterson, 1981). Our data suggest that the phenotypic plasticity of sympathetic neurones is not fully bidirectional. A high TOH activity can superimpose on previously acquired high CAT levels, but the induction of CAT expression in cultures that already express TOH at a high level is accompanied by an actual decline in TOH activity. Moreover, high K + medium can not reverse cholinergic cultures to a purely noradrenergic status. On the other hand, a pre-exposure to high K + medium does not stabilize the noradrenergic phenotype in a permanent manner, but renders the neurones refractory to CAT induction by CM for several days. Materials and methods Neurone cultures Methods used for the culture of sympathetic neurones have been previously described (Swerts et al. 1983; Raynaud et al ). Briefly, neurones dissociated from new-born rat superior cervical ganglia were cultured in a modified L, 5 medium containing 5 % adult rat serum and 370ngml~' 7S Nerve Growth Factor (CM~ medium). The cultures were treated between days 0 and 6 with 10/^M-cytosine arabinoside to prevent the proliferation of ganglionic non-neuronal cells. Primary to tertiary cultures of skeletal muscle cells were grown in roller bottles in L 15 medium containing 10 % fetal calf serum. The medium conditioned (CM) by confluent cultures was collected after 24 h and diluted twice in L 15 medium; it was then supplemented with 5% rat serum and 370ngml~' NGF to obtain CM + medium. The final concentration of K + in both CM + and CM" was 5-15 ± 0-15 ITIM, as measured byflame photometry or with a K + -electrod (Hdpital Rangueil, Toulouse). High K + medium was supplemented to mm by additional KC1 using a 1M solution. Enzyme assays Choline acetyltransferase (CAT) activity was measured according to Fonnum (1975) in the presence of 1-6/iM- [ 3 H]AcCoA (4-8Cimmole"') and 215jiM-AcCoA (final specific activity: 46mCimmole~ 1 ). In some experiments, the sensitivity was increased by performing the assay with l-6/im-[ 3 H]AcCoA (4-8Cimmole~') without isotopic dilution. In that case, the reaction was linear with time and enzyme concentration until 30 % of the labelled substrate was exhausted; CAT activities under these conditions are expressed in ctsmin" 1 rather than pmole, as the AcCoA concentration was below its K m. Tyrosine hydroxylase (TOH) was measured by the method of Waymire, Bjur & Weiner (1971). Statistics Unless indicated, the data are means ± S.E.M. of replicate determinations made with triplicate sister cultures. The data were analysed with the Student's Mest and ANOVAR. Results CM and high K + medium do not affect neuronal survival and growth Sympathetic neurones dissociated from new-born rat superior cervical ganglia were grown for up to 30 days in the virtual absence of non-neuronal cells. In agreement with previous work (Chun & Patterson, \911a,b; Hefti et al. 1982), neither CM nor elevated K + affected the morphology of the neurones at the phase microscope level (Fig. 1). At the electron microscope level, CM decreases, and high K + medium increases, the percentage of small granular vesicles of catecholaminergic type in axonal varicosities (Landis, 1980). During the time period studied, the neuronal population was stable. As previously documented in detail, neither CM nor high K + had any effect on neuronal survival, on the total mass of protein and RNA per dish or on the specific activity of lactate

3 Neurotransmitter plasticity in culture dehydrogenase, a ubiquitous cytoplasmic marker (Patterson & Chun, 1977a; Walicke & Patterson, 1981; Hefti et al. 1982; Swerts et al. 1983; Raynaud et al. 1987b). However, the culture conditions affected the development of CAT and TOH activities. In this article, enzyme activities are given on a per dish basis, as determinations of noncollagen protein were not routinely carried out. The expression on a per protein basis would not affect the interpretation of the data. (The cultures contained neurones, the total mass of noncollagen protein in 15-day-old cultures was 2-3 ng neurone"1.) The amount of protein per dish increased with time in culture, but to a much lesser extent than neurotransmitter synthesis (Mains & Patterson, 1973; Walicke & Patterson, 1981). Reversibility of the effects of CM on CAT and TOH development A marginal CAT activity developed in cultures grown in CM" medium. When the assay was performed in the presence of 0-2ITIM-[ 3 H]ACCOA (46mCi mmole"1) (Fonnum, 1975), CAT activity was at most twice the background of the assay (ISOctsmin"1) after 10 days in culture. Even when the sensitivity of the assay was increased by using [3H]AcCoA ( Ci mmole J) without isotopic dilution, CAT activity was only significant in 8-day-old cultures. On the other hand, 2-day-old cultures already contained a detectable TOH activity (about lpmolmin" 1 dish"1 or 7-5fmolh~1 neurone"1). As expected from previous data (Patterson & Chun, 1977a; Swerts et al. 1983; Raynaud et al. 1987a), CAT activity developed rapidly after a lag of several days in cultures grown with CM (Fig. 2). When cultures were first grown in CM + medium and then switched at days 2 to 8 to CM" medium, the subsequent increase in CAT activity was greatly reduced as compared to cultures maintained in CM + medium. In the experiment of Fig. 2, the rate of CAT increase was depressed by as much as 85-93%. A similar result was obtained when the switch to CM" medium was made at days 6 or 10 and the cultures tested at day 22 (data not shown). It was thus striking that the CAT levels remained high upon removal of the inducing factor and that the residual rate of CAT development was still higher than in neurones that were never submitted to the action of CM. To study the effects of short exposures to CM on TOH development, the cultures from the experiment of Fig. 2 were also tested for TOH activity. In Fig. 1. Primary cultures of sympathetic neurones from new-born rat superior cervical ganglia. Sister cultures were grown for 8 days in (A,B) control medium (5mM-K+, CM"), (C) high K+ medium (35mM-K+, CM") or (D) 50% CM (5mM-K+, CM+). Note the virtual absence of non-neuronal cells. (A) xl50. (B-D) x370.

4 620 S. Vidal, B. Raynaud, D. Clarous and M. J. Weber c I o o. 100 O j= 10 o o. X Days Days Fig. 2. Persistence of elevated CAT levels after CM removal. 28 sister cultures were grown from day 0 onward in CM + (A A) or CM" ( ) medium. (A A) Cultures were first grown in CM + medium from day 0 to day 2, 4, 6 or 8, and then in CM" medium up to day 10. CAT assays were performed in the presence of 0-26mM-[ 3 H]AcCoA. The data points are means of triplicate determinations made on individual cultures. SEM of the assay was less than 15 %. The data are representative of two experiments. agreement with the data of Raynaud et al. (1987a), CM reduced the rate of development of TOH activity by about 75 % (Fig. 3). When cultures were first grown in CM + medium for 2-8 days and then switched to CM" medium, the rate of TOH development after the switch was similar to that observed in cultures that were never exposed to CM. From this and other experiments, we concluded that the impairment of TOH development by CM was fully reversible, at least for the first 10 to 20 culture days. Cultures expressing high levels of both CAT and TOH could thus be obtained by first inducing CAT with CM and then accelerating TOH development by removing CM. It will be shown below that the same result could not be obtained by first allowing high levels of TOH development in the absence of CM and then inducing CAT with CM. 10 Fig. 3. Reversibility of the repression of TOH development by CM. The cultures from the experiment of Fig. 1 were tested for TOH activity. The symbols are the same as those of Fig. 1. For clarity, data points at day 10 have been displaced along the x-axis. Effects of the delayed addition of CM on CAT and TOH development We then determined whether the CAT activity remained inducible in neurones grown for several weeks in CM" medium. As shown of Fig. 4, the CAT activity developed with parallel kinetics no matter whether cultures were switched from CM" to CM + medium at days 2, 9 or 18, suggesting that sympathetic neurones remain fully sensitive to CM for at least 3 to 4 weeks. To study the effects of a switch from CM" to CM + medium on TOH development, neurones were first grown for 10 days in CM" medium. As expected from earlier data (Fig. 3 and Raynaud et al. 1987a), TOH activity developed rapidly under this condition. When the neurones were then switched to CM + medium, TOH activity first dropped to low levels in 6-12 days and then developed again, but at a slower rate characteristic of cultures maintained in CM + medium for day 2 on (Fig. 5). The exposure of neurones to high K +, CM" medium increases the rate of TOH development and the level of TOH-mRNA (Hefti et al. 1982; Raynaud et al ). Even after such a treatment, a switch to

5 Neurotransmitter plasticity in culture 621 6* p I X Days Fig. 5. Effects of a delayed treatment with CM on TOH development. Neurone cultures were grown up to day 10 in low K + ( ) or high K + (O) medium without CM. At day 10, all cultures were switched to low K +, CM + medium. TOH activity is expressed in pmolmin~'dish~ l. The data are means ± S.E.M. for triplicate cultures. *, Differs at F<005 from ( ) cultures of the same age; *',P< Fig. 4. Development of CAT activity in sympathetic neurone cultures after delayed additions of CM. 39 sister cultures were grown either with (A A) or without (A A) 50 % CM. The treatment with CM was started at culture days 2, 9 or 18 as indicated. The cultures were harvested and tested for CAT activity in the presence of 1-6^M-[ 3 H]ACCOA. Data points are means ± S.E.M. for three sister cultures. low K +, CM + medium caused a rapid decline in TOH activity similar to that observed in cultures that had not been previously depolarized (Fig. 4). The reason for the difference in TOH activity between the two groups observed at day 28, but not at days 15, 21 and 32, is not clear. Nevertheless, this experiment suggested that depolarization did not induce a long-lasting state of high TOH expression. This is in sharp contrast with the consequences of such a treatment on CAT development which are considered below. Effects of sequential treatments with CM and high K + medium Previous studies have shown that CM and elevated K + (30-40 ITIM) had antagonistic effects on the development of CAT and TOH activities when added simultaneously to sympathetic neurone cultures: high 26 K + decreased by 90 % the stimulation of CAT development caused by CM, whereas CM antagonized the increase in TOH activity caused by elevated K + (Raynaud et al. 1987a). We wondered if similar phenomena occurred when CM and high K + medium were given sequentially rather than simultaneously. A first series of experiments was performed to study if high K + medium could reverse the development of CAT activity caused by a pre-exposure to CM. Cultures were first grown in CM +, low K + medium between days 2 and 12 and subsequently in CM~ medium containing either low K + (group b) or high K + (group c) between days 12 and 21. Other cultures (group a) were maintained between days 2 and 21 in CM +, low K + medium. In agreement with the data from Fig. 2, the increase in CAT activity observed between days 12 and 21 with the group b was reduced to 36 % of that observed with the group a (Table 1). Even in cultures switched to high K + medium (group c), CAT continued to increase, although at a smaller rate (24 % of the group a value). This increase was distinctly higher than that observed in cultures that were never exposed to CM. Consequently, an exposure to CM caused a stimulation in CAT development which could not be totally reversed by a switch to CM~, high K + medium, although high K + almost totally antagonized the effects of CM when the two agents were added

6 622 S. Vidal, B. Raynaud, D. Clarous and M. J. Weber Table 1. Effects of high K + medium on the development of CAT activity after CM withdrawal CAT activity Increase between (ctsmin~'dish~' xlo" 3 ) days 12 and 21 (A) Test at day 12 (B) Test at day 21 (a) CM (b) 5mM-KCl (c) 45mM-KCl 23 ± ± ±3 85 ±4 259(100%) 93 (36 %) 62 (24 %) 200 Cultures were grown between days 2 and 12 in a culture medium containing 5mM-KCl and 50% CM. Three cultures were tested at day 12 (A). (B) Between days 12 and 21, cultures were either maintained in CM (a) or in the absence of CM in a medium containing 5ITIM- (b) or 45 mm- (c) KG and tested at day 21. Data are means ± S.E.M. for three cultures, b and c differ from a, and from each other at the P< level. 100 simultaneously to culture medium (Raynaud et al. 1987a). In a second series of experiments, we tested the effects of long-term neuronal depolarization before an exposure to CM. Neurones were maintained between days 2 and 12 in either low or high K + medium without CM and then switched to low K +, CM + medium. It has been shown above that the previous depolarization did not prevent the rapid decline in TOH activity caused by a switch from CM" to CM + medium (Fig. 5). On the other hand, the previous depolarization introduced a delay of 8 to 10 days in the development of CAT activity after the switch, but CAT activity then developed with kinetics parallel to that observed in cultures not pre-exposed to high K + medium and eventually reached identical levels (Fig. 6). Thus, neuronal depolarization impaired in a long-lasting, but not permanent, manner the induction of CAT activity by CM. Depolarizations for 10 or 2 days before a switch to CM + medium caused identical delays in CAT development (data not shown), although the minimal efficient duration of pre-exposure to high K + remains to' be determined. Discussion The main objective of this study was to discover whether the expression of cholinergic or noradrenergic traits by cultured sympathetic neurones was reversible or if a superimposition of acetylcholineand catecholamine-synthesizing enzymes could be obtained by fostering first the development of CAT and then that of TOH, or vice versa. These questions have been addressed by altering the culture conditions from CM" to CM + medium or from CM + to CM" medium, at varying culture days. The data suggest that CAT remains fully inducible for at least 2 * * Days Fig. 6. Delay in the development of CAT activity caused by a previous depolarization. The cultures from the experiment of Fig. 4 were also tested for CAT activity. (O) The cultures were first maintained in low K + plus CM" medium up to day 10 and then switched to low K + plus CM + medium. ( ) The cultures were switched from high K + plus CM" medium to low K + plus CM + medium. CAT activity is expressed in ctsmin"' dish" 1 xlo" 3. to 3 weeks, and that this induction is not reversible even in the presence of elevated K +. On the other hand, the impairment to TOH development by CM is fully reversible, so that a state of high TOH expression can be superimposed on previously developed high CAT levels. However, the reverse could not be achieved, since delayed CAT induction by CM was accompanied by a marked decrease in TOH expression. Are the effects of CM reversible? CAT activity develops at a high rate in CM + medium reflecting an increase in the number of immunotitrable enzyme molecules (Raynaud et al. 1987a). Since tt has been shown that isolated neurones in microcultures do not acquire cholinergic characteristics at an uniform rate (Potter, Landis & Furshpan, 1980; Potter et al. 1986), the kinetics of CAT development in CM + cultures may result both from the progressive recruitment of neurones to a cholinergic status and from the accumulation of CAT molecules in individual neurones. This issue could be clarified 30

7 Neurotransmitter plasticity in culture 623 by immunostaining neurones expressing CAT. However, all our attempts using both poly- and monoclonal anti-cat antibodies have been unsuccessful so far. When CM was withdrawn, the subsequent rate of CAT development was decreased to 7-36 % of the value observed in cultures maintained in CM, but remained significantly higher than in cultures that were never exposed to CM. The persistence of CAT activity upon the removal of CM may result from the stability of presynthesized CAT molecules and/or CAT-mRNA. More probably, CM removal stops the further recruitment of neurones to a cholinergic status, but those neurones that were already recruited may continue to express the gene for CAT either at the same level or possibly at a lower, but still significant, level. Even elevated K +, which antagonized the effects of CM when both agents were provided simultaneously (Raynaud et al. 1987a), could not block the development of CAT activity after CM removal. It is thus plausible that CM can cause an irreversible CAT induction, although direct measurements of CAT gene transcription will be necessary to clarify this issue. In addition to inducing CAT activity, CM impairs the development of TOH activity (Swerts et al. 1983; Wolinsky & Patterson, 1983) and of TOH-mRNA levels (Raynaud et al ). The cholinergic factor purified to homogeneity from CM also decreases [ 3 H]catecholamine synthesis in similar cultures (Fukada, 1985). When CM was withdrawn at progressively later times in culture, the development of TOH activity was stimulated as compared to cultures maintained in CM + medium. Thus a switch from CM + to CM~ medium led to significant increases in both CAT and TOH activities as if cholinergic and noradrenergic traits were simply superimposed. This situation may occur in vivo in the adult rat iris, where parasympathetic nerve fibres can express TOH immunoreactivity following sympathectomy (Bjorklund et al. 1985). Similarly, cholinergic ciliary neurones can express adrenergic traits when transplanted into a permissive environment (Coulombe & Bronner- Fraser, 1986). However, in these two cases, it has not been determined whether the new phenotypic traits developed at the expense of CAT activity. On the other hand, it has been reported that CAT activity and TOH immunoreactivity develop concomitantly in chick ciliary neurone cultures (Iacovitti et al. 1985). Do neurones lose their sensitivity to CM in culture? The addition of CM to 9- or 18-day-old cultures caused increases in CAT activity at a rate similar to that observed when CM was first given at day 2. As far as the induction of CAT is concerned, sympathetic neurones remained fully sensitive to CM for at least 2 to 3 weeks. During the first culture days in CM~ medium, TOH activity developed rapidly. After a switch to CM + medium, TOH activity first declined to low levels in 6-12 days and then increased again, but at a slower rate, similar to that observed in cultures maintained in CM + medium from day 2 onward. The transient decline in TOH activity during the transition period occurs at a rate compatible with the estimated half-life of TOH molecules (about 40 h, Max, Rohrer, Otten & Thoenen, 1978) and may result from the normal turnover of TOH-mRNA. However, we can not exclude the possibility that CM may destabilize TOH and/or TOH-mRNA molecules. High K + medium increases TOH activity and TOH-mRNA level in these cultures (Hefti et al. 1982; Raynaud et al. 1987a,6). However, the transition from a high to a lower state of TOH expression caused by CM was not affected by a previous depolarization. This suggests that K + do not stabilize the noradrenergic phenotype in a permanent manner. Even after initially fostering TOH development with high K + medium, the delayed addition of CM could not lead to cultures expressing high levels of both CAT and TOH. This contrasts with the effects of a switch from CM + to CM" medium, after which the cultures could express high levels of both enzymes because CAT activity did not decline after the switch. As a working hypothesis, we thus suggest that in neural crest derivatives, noradrenergic traits can simply superimpose to previously acquired cholinergic traits, but that the induction of CAT in noradrenergic neurones reduces the expression of catecholamine synthesizing enzymes. This is reminiscent of the development of cholinergic sympathetic neurones in vivo, as CAT development in sympathetic nerve terminals in rat sweat glands is accompanied by a marked decline in TOH and dopamine-j3-hydroxylase immunoreactivities (Siegel, Schwab & Landis, 1982; Landis & Keefe, 1983; Leblanc & Landis, 1986). In addition, our experiments suggest that sympathetic neurones remain fully sensitive to CM for at least 2 to 3 weeks in culture. However, other experiments suggest that these neurones progressively lose their plasticity, both in vivo, (Johnson, Ross & Bunge, 1980) and in culture (Patterson & Chun, 19776). Nevertheless, it is well established that mature sympathetic neurones can acquire cholinergic traits when placed into culture, although at a reduced frequency compared to perinatal neurones (Wakshull et al. 1979; Potter etal. 1986). A pre-exposure to elevated K + delays CAT induction by CM Previous experiments by Walicke et al. (1977) and Walicke & Patterson (1981) shdwed that a 7- to 20-

8 Vidal, B. Raynaud, D. Clarous and M. J. Weber day depolarization of cultured sympathetic neurones depressed two- to threefold the increase in [ 3 H]- acetylcholine synthesis caused by a subsequent 10-day exposure to CM. We now show that such a predepolarization did not impair permanently CAT induction by CM, but rather delayed the induction by about 10 days. The molecular mechanisms of this memory effect are presently unknown. An entry of Ca 2+ plays a major role in the developmental effects of depolarization (Walicke & Patterson, 1981; Heftier al. 1982). Specifically, the inhibition of CAT development by K + is mediated by a Ca 2+ entry through dihydropyridine-sensitive channels (S. Vidal, unpublished data). The delay in CAT development observed after a prolonged depolarization may thus result from a continued Ca 2+ entry or a persistent activation of Ca 2+ -mediated intracellular processes. We thank Dr A. C. Kato, Dr C. Bader and Dr Y. A. Barde for the critical reading of the manuscript and Dr J. P. Swerts for TOH assays. This work was supported by grants from the Centre National de la Recherche Scientifique and the Direction de la Recherche, Etudes et Techniques ( ). References BJORKLUND, H., HOKFELT, T., GOLDSTEIN, M., TERENIUS, L. & OLSON, L. (1985). Appearance of the noradrenergic markers tyrosine hydroxylase and neuropeptide Y in cholinergic nerves of the iris following sympathectomy. J. Neurosci. 5, CHUN, L. L. Y. & PATTERSON, P. H. (1977a). Role of nerve growth factor in the development of rat sympathetic neurons in vitro. I. Survival, growth and differentiation of catecholamine production. J. Cell Biol. 75, COULOMBE, J. & BRONNER-FRASER, M. (1986). Cholinergic neurones acquire adrenergic neurotransmitters when transplanted into an embryo. Nature, Lond. 2>2A, FONNUM, F. (1975). A rapid radiochemical method for the determination of choline acetyltransferase. J. Neurochem. 24, FUKADA, K. (1985). Purification and partial characterization of a cholinergic neuronal differentiation factor. Proc. natn. Acad. Sci. U.S.A. 82, HEFTI, F., GNAHN, H., SCHWAB, M. E. & THOENEN, H. (1982). Induction of tyrosine hydroxylase by Nerve Growth Factor and by elevated K + concentrations in cultures of dissociated sympathetic neurons. J. Neurosci. 2, iacovrrn, L., JOH, T. H., ALBERT, V. R., PARK, D. H., REIS, D. J. & TEITELMAN, G. (1985). Partial expression of catecholaminergic traits in cholinergic chick ciliary ganglia: studies in vivo and in vitro. Devi Biol. 110, JOHNSON, M. I., Ross, C. D. & BUNGE, R. P. (1980). Morphological and biochemical studies on the development of cholinergic properties in cultured sympathetic neurons. II. Dependence on postnatal age. J. Cell Biol. 84, LANDIS, S. C. (1980). Developmental changes in the neurotransmitter properties of dissociated sympathetic neurons: a cytochemical study of the effects of medium. Devi Biol. 77, LANDIS, S. C. & KEEFE, D. (1983). Evidence for neurotransmitter plasticity in vivo. Developmental changes in properties of cholinergic sympathetic neurons. Devi Biol. 98, LEBLANC, G. & LANDIS, S. (1986). Development of choline acetyltransferase (CAT) in the sympathetic innervation of rat sweat glands. /. Neurosci. 6, LE DOUARIN, N. M., TEILLET, M. A., ZILLER, C. & SMITH, J. (1978). Adrenergic differentiation of cells of the cholinergic ciliary and Remak ganglia in avian embryo after in vivo transplantation. Proc. natn. Acad. Sci. U.S.A. 75, MAINS, R. E. & PATTERSON, P. H. (1973). Primary cultures of dissociated sympathetic neurons. I. Establishment of long-term growth in culture and studies of differentiated properties. J. Cell Biol. 59, MAX, S. R., ROHRER, H., OTTEN, U. & THOENEN, H. (1978). Nerve growth factor-mediated induction of tyrosine hydroxylase in rat superior cervical ganglion in vitro. J. biol. Chem. 253, PATTERSON, P. H. & CHUN, L. L. Y. (1977a). The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. I. Effects of conditioned medium. Devi Biol. 56, PATTERSON, P. H. & CHUN, L. L. Y. (19776). The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. II. Developmental aspects. Devi Biol. 60, POTTER, D. D., LANDIS, S. C. & FURSHPAN, E. J. (1980). Dual function during development of rat sympathetic neurons in culture. /. exp. Biol. 89, POTTER, D. D., LANDIS, S. C, MATSUMOTO, S. G. & FURSHPAN, E. J. (1986). Synaptic functions in rat sympathetic neurons in microcultures. II. Adrenergic/ cholinergic dual status and plasticity. J. Neurosci. 6, RAYNAUD, B., CLAROUS, D., VIDAL, S., FERRAND, C. & WEBER, M. J. (1987a). Comparison of the effects of elevated K + ions and muscle-conditioned medium on the neurotransmitter phenotype of cultured sympathetic neurons. Devi Biol. 121, RAYNAUD, B., FAUCON-BIGUET, N., VIDAL, S., MALLET, J. & WEBER, M. J. (19876). The use of a tyrosine hydroxylase cdna probe to study the neurotransmitter plasticity of rat sympathetic neurons in culture. Devi Biol. 119, SIEGEL, R., SCHWAB, M. & LANDIS, S. C. (1982). Developmental changes in the neurotransmitter properties of cholinergic sympathetic neurons in vivo. Soc. Neurosci. Abstr. 8, 7.

9 Neurotransmitter plasticity in culture 625 SWERTS, J. P., LE VAN THAI, A., VIGNY, A. & WEBER, M. J. (1983). Regulation of enzymes responsible for neurotransmitter synthesis and degradation in cultured rat sympathetic neurons. I. Effects of muscle conditioned medium. Devi Biol. 100, WAKSHULL, E., JOHNSON, M. I. & BURTON, H. (1979). Postnatal rat sympathetic neurons in culture. II. Synaptic transmission by post-natal neurons. J. Neurophysiol. 42, WALICKE, P. A., CAMPENOT, R. B. & PATTERSON, P. H. (1977). Determination of transmitter function by neuronal activity. Proc. natn. Acad. Sci. U.S.A. 74, WALICKE, P. A. & PATTERSON, P. H. (1981). On the role of cyclic nucleotides in the transmitter choice made by cultured sympathetic neurons. J. Neurosci. 1, WAYMIRE, J. C, BJUR, R. & WEINER, N. (1971). Assay for tyrosine hydroxylase by coupled decarboxylation of dopa formed from [l- l4 C]tyrosine. Analyt. Biochem WOLINSKY, E. & PATTERSON, P. H. (1983). Tyrosine hydroxylase activity decreases with induction of cholinergic properties in cultured sympathetic neurons. J. Neurosci. 3, (Accepted 1 July 1987)

10