Synaptic Functions in Rat Sympathetic Neurons in Microcultures. IV

The Journal of Neuroscience, February 1987, 7(2): 380-390

Synaptic Functions in Rat Sympathetic Neurons in Microcultures. IV. Nonadrenergic Excitation of Cardiac Myocytes and the Variety of Multiple-Transmitter States

S. G. Matsumoto, D. Sah, D. D. Potter, and E. J. Furshpan

Department of Neurobiology, Harvard Medical School, Boston, Massachusetts 02115

In the first 3 papers of this series (Furshpan et al., 1988a, b; Potter et al., 1988), a sensitive microculture procedure was used to show that sympathetic principal neurons, dissociated from newborn or adult superior cervical ganglia and grown singly on cardiac myocytes, display adrenergic, cholinergic, and purinergic functions, sometimes in isolation but more often in combination. In this paper we describe additional effects on cardiac myocytes evoked by these neurons; the effects were excitatory and insensitive to adrenergic blocking agents (and to agents that block the inhibitory effects of acetylcholine and purines). In some of these microcultures, evidence consistent with secretion of serotonin was obtained; the nonadrenergic excitatory effect was diminished or abolished by serotonin blockers or reserpine. Further evidence for serotonergic transmission is presented in the accompanying paper by Sah and Matsu- moto (1987). In other cases, an as-yet-unidentified agent “X” also produced a nonadrenergic excitation. The X effect characteristically required a prolonged train of neuronal impulses, had a time course of 50-200 set, and was insensitive to agents that affected the other transmitters, including serotonin.

In addition, we discuss 2 remarkable features of the transmitter repertoire of the microcultured sympathetic neurons: expression of the several transmitters in a variety of combinations, including at-least-quadruple function, and expression of the transmitters within a particular combination in varying relative strengths. The result is a diversity of transmitter release greater than that previously reported for vertebrate or invertebrate neurons.

The microculture procedure we have previously used to investigate adrenergic, cholinergic, and purinergic functions in sympathetic principal neurons (Furshpan et al., 1986a, b; Potter et al., 1986) was used in the work described here to investigate

Received Jan. 7, 1986; revised May 27, 1986; accepted June 11, 1986. Expert assistance was provided by Robert Bosler, Wendy Brooks, Delores Cox,

Karen Fischer, Joseph Gagliardi, Nona Haynes, James LaFratta, Michael La- Fratta, Doreen McDowell, Claudia Miles, Geraldine Spencer, Shirley Wilson, and Vivienne Yee. Geraldine Spencer and Drs. Linda Chun, Allison Doupe, and Eve Wolinsky provided nerve growth factor, and Dr. Keiko Fukada provided partially purified CM factor. We are indebted to many colleagues for helpful discussions, especially Drs. Partick Hogan, Story Landis, Rae Nishi, Paul Patterson, and Alan Willard. This work was supported by NIH Grants NS02253, NSl1576, NS18316, NS03273, and Training Grants NS07112 and MH18012. Some experiments were performed at the Marine Biological Laboratory at Woods Hole, MA.

Correspondence should be addressed to Dr. S. G. Matsumoto, Department of Neurobiology, Harvard Medical School, 25 Shattuck Street, Boston, MA 02 115. Copyright 0 1987 Society for Neuroscience 0270-6474/87/020380-l 1$02.00/O

additional transmitter states. In most cases, the microculture contained only 1 neuron, growing on a small island of cardiac myocytes; in the few remaining cases, 2 neurons were present. Electrical recordings from the neuron and the myocytes provided a sensitive assay for synaptic function(s) in the neuron, as described in Furshpan et al. (1986a).

The starting point of the work described here was the dis- covery of excitatory effects on the myocytes that were not sensitive to agents that readily block adrenergic excitation or cholinergic and purinergic inhibitions of the myocytes. The first concern of this paper is to illustrate these nonadrenergic excitatory effects; we present evidence consistent with the idea that these effects were sometimes caused by secretion of serotonin and sometimes by secretion of another transmitter (or transmitters), “X,” that we have not yet succeeded in blocking or mimicking with conventional agents. A second concern is to illustrate the impressive diversity of transmitter states expressed by the microcultured neurons.

Materials and Methods Culture procedures. The methods for making and recording from microcultures were in most respects identical to those previously described (Furshpan et al., 1986a). In brief, principal neurons were dissociated from the superior cervical ganglia (SCG) of newborn rats by a mechan- ical procedure (Bray, 1970), or from the SCG of adult rats by an en- zymatic procedure based on the use of collagenase and dispase (Fursh- pan et al., 1986a). Adult rats were at least 8 weeks old and weighed 200-450 gm. The neurons were grown on small islands (0.3-0.5 mm diameter) of heart cells previously dissociated from the atria and ven- tricles of newborn rats. Twenty-five such islands were plated in the central well of each 35 mm plastic petri dish. The wells were made by fastening a plastic coverslip over a hole (ca. 1 cm diameter) cut in the bottom of each dish. The neurons were plated at low density such that some of the heart cell islands received a single neuron.

The several media in which tissues were dissected, and in which cells were plated and grown, were all based on Leibowitz’s L-15 medium. The modifications for each medium were as described in Furshpan et al. (1986a). The growth medium contained NGF (7S, 1 &ml). To delay or reduce expression of cholinergic properties, some cultures were fed a medium that contained 20 mM K+ (see Walicke et al.. 1977): to make this medium, NaCl was replaced by KCl.

,_

Assay procedures. Almost all the experiments described here were performed with single-neuron microcultures. The presence of only a single neuron was determined by careful through-focusing with the phase- contrast microscope; the accuracy of this procedure had previously been verified with the electron microscope (Landis, 1976). A few of the data points in Figure 4 were from 2-neuron microcultures.

Tests of transmitter status were made by stimulating the neuron while recording with intracellular microelectrodes any synaptic effects on the neuron itself and on the myocytes, either in normal solution (Furshpan et al., 1986a) or in the presence of receptor blocking agents. Drugs were applied in the constant perfusion stream (temperature about 36°C; flow rate usually 0.3-0.4 ml/mitt; volume of fluid to be exchanged, ca. 0.1

The Journal of Neuroscience, February 1997, 7(2) 381

a .

cant

d

atrop+ 8-PT

b a

atrop

r- , e

C

l atrop + 8-PT

Figure 1. Nonadrenergic excitation exerted by a solitary, neonate-derived neuron grown for 42 d in a microculture. As in most other figures, the upper trace in each record is an intracellular recording from cardiac myocytes and the lower truce is an intracellular recording from the neuron at low gain, so that impulses or trains of impulses appear as small deflections. a-c, Myocyte trace starts with a 10 mV calibration step; d and e, 1 mV steps. a-c, Neuron was first stimulated once, then 5 x at 1.5 Hz, and then at 16 Hz. In control solution (cant), the quiescent myocytes were sharply hyperpolarized by each of the stimulations (a). A substantial block of the hyperpolarization was produced by 1 PM atropine sulfate (atrop) in b and a complete block by addition of 8-PT (nominally 1 PM) to the atropine in c, demonstrating that the neuron secreted both ACh and purine(s). The neuronal activity in c now produced a small, slow depolarization shown at higher gain in d (neuronal stimulation at 16 Hz); an initial negative-going deflection, presumably an incompletely blocked effect of 1 or both inhibitory transmitters, was followed by a depolarization of about 1.6 mV that persisted onto the next sweep; after about 3 min, a train of cardiac impulses appeared (dejkctiom at the right end of the truce). Addition of the (Y- and @-adrenergic blockers atenolol (20 PM) and phentolamine (0.1 PM) to the atropine and 8-PT (4 blockers = 4B, e), hardly altered the slow depolarization and established it as an NAE effect. u-c, 20 mV, d and e, 8 mV, 20 set for u-c, 60 set for d and e.

ml) or “puffed” locally from a micropipette (method of Choi and Fisch- bath, 198 l), as described in Furshpan et al. (1986a); puffed peptides were applied in a solution containing 0.1% BSA (Sigma). Chart recorders were often used to investigate the nonadrenergic excitation, which had a time course of many tens of seconds.

Drugs. Adrenergicand cholinergic agents were obtained from the same sources as in Furshpan et al. (1986a). The sources of other reagents were serotonin-creatinine sulfate and gramine: Sigma; methysergide: Sandoz; somatostatin; substance P, neurotensin, peptide YY (PYY), and vasoactive intestinal DolvDeDtide (VIP): Peninsula: I-Dhenvltheophvlline (8- PT): Calbiochem; ;6-&-sulfophenyl)-theophylline (8-SPT): kesearch Biochemicals. The peptides were stored frozen in 0.1 N HCl.

Results The growth, appearance, and general properties of the microcultured myocytes (neonate-derived) and neurons (neonate- and adult-derived) were similar to those described in the previous papers of this series (Furshpan et al., 1986a, b; Potter et al., 1986). In considering the synaptic effects reported here, the unusual relationship of the neuron and its target myocytes is relevant. Under the influence of NGF in the medium, a microcultured neuron formed an increasing number of varicosities on a restricted number of myocytes and often achieved a density of innervation higher than that reported for adult sympathetic target tissues in vivo (Furshpan et al., 1986a). When the neuron was stimulated, the many varicosities acted in parallel on the relatively small population of electrically coupled myocytes. This feature enhanced the sensitivity of the assay of synaptic effects. When only a single neuron was present, as in all the

figures in this paper except Figure 4, and in the great majority of assays summarized in that figure, the synaptic effects, no matter how complex, could be assigned to that solitary neuron.

Nonadrenergic excitation of the cardiac myocytes

Figure 1 illustrates a nonadrenergic excitatory (NAE) effect exerted by a neonate-derived neuron. The pronounced rapid hy- pet-polarizations of the myocytes (a) produced by a single neuronal impulse or by trains of impulses (at 1.5 and 16 Hz) were sharply reduced by atropine (b), indicating a substantial cholinergic effect (Furshpan et al., 1986a). Although the concentration of atropine (1 PM) was 10 x that usually needed to block cholinergic effects, hyperpolarizations with a slower time course were still produced, most effectively by trains of neuronal impulses (b). These were blocked, in turn, by 8-phenyltheophylline (8-PT; c), an agent effective against adenosine receptors, indicating that the atropine-resistant inhibition was purinergic in origin and, plausibly, mediated by adenosine. [As discussed in Furshpan et al. (1986b) it is not known whether the secreted agent was adenosine, one or more of its phosphorylated deriv- atives, or a combination of these substances.] Blocking of the cholinergic and purinergic effects of the neuron unmasked a slow, small depolarization seen clearly in Figure Id at higher amplification and slower sweep speed. As the depolarization was also insensitive to CY- and ,&adrenergic blockers (Fig. le) at concentrations sufficient to block substantial adrenergic effects (Furshpan et al., 1986a), it was characterized as an NAE effect.

382 Matsumoto et al. l Nonadrenergic Excitation: Diversity of Transmitter States

Figure 2. NAE effects in 2 neonate- derived neurons grown singly in separate microcultures: N, for 35 d and N, for 49 d. Recordings as for Figure 1. N,. Stimulation (20 Hz for 6.5 set) in control solution (cant) produced an excitatory effect on the myocytes (a,). In the presence of 4 conventional blockers (0: atropine sulfate, 1 PM; atenolol, 50 PM;

phentolamine, 5 PM; and &SPT, 10 PM),

stimulation at 20 Hz for 18.3 set produced a small, slow depolarization, an NAE effect (b,; note change of gain). Addition of 10 KM methysergide (MS) and 10 fly gramine (CR), 2 serotonin- receptor blockers, to 4B eliminated the NAE effect (c,); when the serotonin blockers were removed, the NAE effect reappeared (d,; 4B still present). Per- fusion (wash) with 10 PM reserpine phosphate for 20 min (res; e,) irrever- sibly blocked both the adrenergic (not shown) and NAE effects. This evidence is consistent with the idea that N, is at least adrenergic/serotonergic in function. NZ, As in the previous case., stimulation of N2 (20 Hz for about 6.5 set) produced an excitatory effect on the myocytes (a,) that was reduced by 4B to an NAE effect (b,; stimulation at 20 Hz for 16 set). However, in the presence of 4B plus methysergide and gramine (MS/GR, c,), the NAE effect was still present, nor was it eliminated by perfusion (wash) of the neuron for 120 min in 10 PM reserpine phosphate (res; dJ. Because the NAE effect of N2 was resistant to the serotonin blockers and reserpine, NZ was classified as at least adrenergic/X-ergic. Cardiac myocyte traces in a, and a,, 80 mV, in all other myocyte traces, 8 mV.

al cant

--t-----c 46

“‘-

48 +Ms/Gr

Cl...s,.lrp- C22

l--m--- m

wash

d1 - 80 mV

8 mV L 20 s

Consequently, this neuron was classified as cholinergic/purinergic/NAE. Because the NAE effect was not blocked, it is not known whether the effect was produced by a single agent. Adult- derived neurons could also exert the NAE effect (13 of 8 1 adult- derived neurons tested; see Fig. 4).

As in the case illustrated in Figure 1, NAE effects were generally small (l-4 mV in amplitude) and required rather prolonged trains (15-30 set in duration) of neuronal impulses at frequencies of 1 O-20 Hz to obtain a clear response. The duration of NAE effects was dependent on the frequency and duration of these trains, but was usually 50-200 sec. It was usually nec- essary to wait lo-30 min between trials to regain the full response. In some cases the response declined progressively so that the number of effective trials was small. In these respects, NAE effects were less robust and more difficult to investigate than the adrenergic, choline& and purinergic effects described in the previous papers of this series (Furshpan et al., 1986a, b, Potter et al., 1986).

Observations of the type illustrated in Figure 1 established that the NAE effects were distinct from the adrenergic, cholin-

ergic, and purinergic effects described previously (Furshpan et al., 1986a, b). Could NAE effects be adrenergic, but exerted via an unconventional class of adrenergic receptors [cf. the “y-receptors” proposed by Hirst and Neild, 1980 (for discussion see Neild and Zelcer, 1982)]? Evidence against this possibility is provided by an experiment on a 2-neuron microculture illustrated in figure 8 of Potter et al. (1986); 1 of the 2 neurons produced an apparently purely adrenergic effect on the myocytes that was completely blocked by (Y- and P-blockers, while the other neuron exerted an NAE effect resistant to these blockers. As the neurites of the 2 neurons were presumably intermingled on the coupled myocytes, if the NAE effect was exerted via unconventional adrenergic receptors, the adrenergic effect of the first neuron should not have been eliminated by the conventional blockers. Further evidence that the NAE effect is not an adrenergic response is provided in the next section.

Pharmacology of the NAE efect

Serotonin is known to have an excitatory action on cardiac myocytes and vascular smooth muscle cells (see Discussion).


cant

a4P

3B

C

30 mV I

atrop

3B + 8-PT

d

I 30 s

Figure 3. A solitary, neonate-derived neuron, at least quadruple in function after 44 d in microculture. Recordings as for Figure 1. a, In control solution (cant), a single neuronal impulse produced a rapid, pronounced hyperpolarization of the myocytes followed by a depolarization; both effects on the myocytes were larger and more prolonged in response to a train of neuronal impulses. b, Elimination of the rapid hyperpolarization by addition of 1 PM atropine sulfate (strop) to the perfusion fluid showed that this neuron was relatively strongly choline+; the effect of a single neuronal impulse or a train was now strongly excitatory. c, Addition of 10 PM atenolol, 0.1 /LM phentolamine to the atropine (3B) unmasked a substantial atropine-resistant hyperpolarization; the net excitation in b, in spite of this inhibitory effect, was attributable to a relatively strong adrenergic effect. d, The addition of 1 PM 8-PT to 3B sharply reduced the atropine-resistant hyperpolarization, which was thus characterized as a moderate purinergic effect; an NAE effect was unmasked. Trains of stimuli in a-d were at 20 Hz.

Figure 2 shows pharmacological tests for serotonergic function in 2 single-neuron microcultures; for neuron N, (a+,), the pharmacological evidence is consistent with a serotonergic NAE effect, while for neuron N, (a&) the pharmacological evidence suggests that some other, unidentified agent(s) produced the NAE effect. In both cases, stimulation of the neuron in control solution produced an excitatory effect on the myocytes that was reduced to an NAE effect by perfusion with “4B,” a mixture of the 4 blockers, atropine, atenolol, phentolamine, and 8-SPT (Fig. 2, b, and b,). The NAE effect of N, was eliminated by the further addition of 2 serotonin antagonists, methysergide and gramine (both at 10 pm), to the 4B (Fig. 2, c,). The NAE effect of N, was insensitive to these agents (c,). When the serotonin blockers were removed (Fig. 2, d,; 4B still present), the NAE effect of N, was restored. The distinction between the two NAE effects was seen again after perfusion of the 2 microcultures with 10 I.LM reserpine phosphate for about 20 min; the effect of N, (still in 4B) was eliminated (Fig. 2, e,), but the effect of N, was still present (L&).

The sensitivity of the NAE effect of N, to serotonin blockers and to reserpine suggests strongly that secretion of 5-HT was responsible for the effect (reserpine has long been known to reduce neuronal storage of 5-HT, e.g., Brodie and Shore, 1957). Substantial or complete blocking of an NAE effect by these blockers (l-20 pm) was seen in 7 of the 13 microcultures in which blockers were used. Elimination of an NAE effect by reserpine (10 PM) occurred in 1 of 6 cases in which reset-pine was tried.

Puffs of authentic 5-HT (lo-100 PM) were applied to 8 microcultures. In all cases, small depolarizations (2-10 mV) were

evoked in the myocytes with time courses similar to that of NAE effects.

These lines of evidence for serotonergic transmission by cultured sympathetic neurons led to further biochemical and im- munocytochemical experiments, described in the accompanying paper by Sah and Matsumoto (1987). It was found that mass cultures (several thousand neurons per dish) contain a substance identified as 5-HT by high-performance liquid chromatography (HPLC) and that this substance is released into the medium in a Ca2+-dependent manner; some neurons are immunoreactive for 5-HT. Evidence was also obtained for the synthesis of 5-HT by at least some of the neurons grown in certain culture con- ditions. The combination of this biochemical and immunocy- tochemical evidence with pharmacological evidence like that of Figure 2 leads us to conclude that secretion of 5-HT is responsible for 1 class of NAE effects.

Resistance of the NAE effect to the serotonin blockers (up to 20 PM) and to reserpine, as in the case of N, in Figure 2, is consistent with secretion of an additional excitatory agent or agents, which we designate “X.” Attempts to mimic the X effect with several peptides whose immunoreactivities have been reported in adult sympathetic neurons in vivo have so far been unsuccessful. The peptides, applied by pressure ejection from a micropipette, were somatostatin (10 trials at 0.01-l PM), substance P ( 10 trials at 0.00 l-l PM), neurotensin ( 10 trials at 0.0 l- 1 PM), PYY (10 trials at 0.001-l MM), and vasoactive intestinal polypeptide (VIP) (40 trials at 0.001-l PM). As we cannot yet block or mimic the X effect, we cannot state whether more than 1 agent is responsible. Consequently, when an unresolved NAE effect or an X effect was exerted by an apparently dual-, triple-,

384 Matsumoto et al. l Nonadrenergic Excitation: Diversity of Transmitter States

or quadruple-function neuron, the phrase “at least” is tacitly included in the description of the multiplicity of the function.

Although we have not identified a candidate for mediating the X effect, it seems likely that it is produced by a transmitter- like agent. The duration of the effect after cessation of the stim- ulus (up to 3 min) is much greater than the time constant of the myocyte membrane (see Fig. 2 of Furshpan et al., 1986a). This is inconsistent with production by electrical coupling between neuronal terminals and myocytes. In 3 instances an X effect was blocked by perfusion with a high Mgz+/-low Caz+ solution (2 at 20 mM Mg*+/0.75 mM Ca2+; 1 at 3 mM Mg2+/“O” Ca’+), a result

transmitter” by adrenergic neurons in vivo (see Discussion), and as the adult rat serum fed to the microcultured neurons contained serotonin (Sah and Matsumoto, 1987), it is likely that “false” serotonergic function was present in at least some of the neurons classified in Figure 4 as A/C/P/NAE, A/P/NAE, A/P/ S, A/C/NAE, A/NAE, or A/S. However, it is also possible that endogenous 5-HT was responsible for some NAE or S effects since some principal neurons can synthesize and store 5-HT when grown with heart-cell-conditioned medium (Sah and Ma- tsumoto, 1987).

Did the synaptic effects listed in Figure 4 differ because of consistent with chemical neurotransmission.

Most of our observations on the NAE effect were made before it was found that serotonin blockers are sometimes effective. Therefore, the relative frequencies of the serotonergic and X effects are not yet known with confidence. Neither quintuple function nor the occurrence of serotonergic or X effects in the absence of other transmitter functions has yet been seen (see next section).

The variety of combinations of transmitter functions

variations in transmitter repertoire, as is assumed in Figure 4, or because the target myocytes varied in transmitter sensitivity, or both? This question was raised in the previous papers of this series with regard to adrenergic, cholinergic, and purinergic states and their combinations; it was tentatively concluded that at least the major source of the variation between assays arose from differences in transmitter repertoire. This conclusion depended on the fact that the myocytes always responded to these agents upon testing, and also on observations on 2-neuron microcultures in which the response(s) to 1 neuron established the sen-

In the preceding papers of this series we have illustrated 8 of sitivity of the common pool of myocytes to the function(s) ap- the transmitter combinations observed so far in the microcultured sympathetic neurons: neurons that were purely adrenergic, cholinergic, or purinergic within the sensitivity of the physio- logical assays (Furshpan et al., 1986a, b); the 3 dual-function combinations of these transmitters (Furshpan et al., 1986b, Pot- ter et al., 1986); a case of adrenergic/cholinergic/purinergic triple function, and a possible case of at least quadruple function (including an NAE effect; Potter et al., 1986). Figure 1 of this paper illustrates a second category of at least triple function.

Figure 3 shows a clear case of at least quadruple function. In

parently lacking in the other neuron (e.g., Fig. 8 in Potter et al., 1986). Until X has been identified and applied, it will not be possible to determine whether the myocytes are reliably sensitive to it. Pending further information, a conclusion about the source(s) of variation with respect to X effects cannot be reached. Several other questions about the incidence and significance of the states listed in Figure 4 are considered in the Discussion.

Variations in the relative strengths of transmitter eflects In addition to the diversity of transmitter combinations illus-

control solution (a), neuronal impulses evoked pronounced hy- trated in Figure 4, there appeared to be another level of variation perpolarization of the myocytes, followed by depolarization. in transmitter repertoire arising from differences in the relative The striking alterations in these effects as atropine (b), then strengths of the expressed transmitter functions. Although our adrenergic blockers (3B; c), and finally 8-PT (d) were sequen- assay was not quantitative, there were clear indications for such tially added show that the neuron exerted cholinergic, adrener- variations. These are illustrated (Figs. 5 and 6) by a comparison gic, and purinergic effects. Each of these responses was robust. between 2 triple-function (adrenergic/cholinergic/purinergic) Even single neuronal impulses produced substantial choline@ neurons. (a) and adrenergic (b) effects; a train of impulses produced a In the neuron shown in Figure 5, cholinergic function was purinergic effect (c) that was also substantial and rose rapidly indicated by the hexamethonium-sensitive autaptic EPSPs (a- (cf. Fig. 5g). The residual depolarization in din the presence of c) and the prominent myocyte hyperpolarizations (d) that were the 4 blockers is an NAE effect. Eleven other quadruple-function blocked by atropine (e). The subsequent addition of adrenergic neurons were observed in this series. cf) and purinergic (h) blockers revealed the presence of these

Figure 4 lists all of the apparently different states observed so transmitter functions. When all 4 blockers were present(h), only far and the culture ages at which each assay was made. The a small complex response remained, due probably to an incom- character of the assays in Figure 4 requires comment. An assay plete block of the 3 transmitter actions. The reversibility of the was included if it was clear that the indicated functions were drug effects is partly illustrated in j and k. present, without any substantial effect of another kind. It was The neurons shown in Figure 6 displayed the same 3 trans- difficult to characterize very weak responses. The assays were mitter functions. The choline& effect on the myocytes is seen not all equally sensitive, and it is likely that a more sensitive in isolation in c (in the presence of adrenergic and purinergic assay would have placed some of the neurons in a higher cat- blockers); the adrenergic effect is seen in f (in the presence of egory. Some neurons were grown in high-K+ medium. Figure 4 muscarinic and purinergic blockers); the purinergic effect is seen is intended simply to emphasize the occurrence of neurons whose in e (in the presence of muscarinic and adrenergic blockers). physiologically significant effects fell into each of the indicated The small autaptic EPSP seen in h provided confirmation of categories. The designation “NAE” is used when no attempt choline@ function. was made to resolve the effect into S (serotonergic) or X. If it The patterns of responses evoked by the 2 neurons were, is assumed that all cases of NAE status are resolvable into S or however, quite different. In Figure 5, a single neuronal impulse X, that X is a single agent and that S occurs only with A (ad- in control solution (arrow in d) evoked a large, rapid cholinergic renergic), the 20 states of Figure 4 reduce to a maximum of 18. hyperpolarization of the myocytes. The purinergic response (Fig. As serotonin is known to be taken up and secreted as a “false 5, g and i) was much smaller and rose gradually during the

AICIPINAE

AICIPIX

Cl PINAE

AIPNAE

Al P/S

A/P/X

AICINAE

A/C/X

A/C/P

PINAE

Status C/NAE

C/P

AINAE

AIS

A/X

A/P

A/C

P

C

A


. . .

.

. . .

. . . .

01 2345676 9 10 11 12 13 14 15 16 17 16 19

Week

Figure 4. Census of transmitter states in microcultures and the culture age at which each state was observed. States are indicated along the vertical uxis: A, adrenergic; C, cholinergic; P, purinergic; NAE, a nonadrenergic excitatory effect not characterized as S (serotonergic) or X (an NAE effect insensitive to serotonin blockers and/or to reserpine). Each symbol represents an assay; its position along the horizontal axis corresponds to the age of the culture in days, where 0 represents the day on which the neuron was placed in culture. Circles, Neonate-derived neurons; stars, adult- derived neurons. An assay was included (303 in all) if it was clear that the indicated transmitter or transmitter combination was secreted and that no other transmitter contributed substantially. In a few cases, 2 neurons were present in the microcultures; in each of these cases, choline& interaction between the neurons was completely blocked with hexamethonium, and there was no remaining sign of synaptic interaction.

prolonged (10 set) train of neuronal impulses. In contrast, the cholinergic response in Figure 6c was much smaller than the purinergic response (e) and rose more slowly. The net effect on the myocytes of trains of impulses in the 2 neurons, in control solution, was also quite different. In Figure 5d the adrenergic excitation was delayed until the end of the train by the intense choline& hyperpolarization. In Figure 6a only depolarization was present throughout the response.

It seems likely that the differences in relative strength between

cholinergic and purinergic myocyte responses reflected differences in the relative amounts of the 2 transmitters released. Bellardinelli and Isenberg (1983) have provided evidence that ACh and adenosine open similar K-channels in isolated atria1 myocytes. This suggests that a large difference in amplitude between the cholinergic and purinergic responses, evoked by a neuron in the same island of myocytes, is not due to a difference in the electrochemical driving force on the ions generating the responses. The slow rise of the smaller responses (purinergic in

366 Matsumoto et al. - Nonadrenergic Excitation: Diversity of Transmitter States

g 3B 4

j atrop +

a 8PT

e 5pM atrop

l -

h 3B+8PT e

k cant

f 3B

i 3B 6

I 40mV,d-f 20mV, g-k

20s

Figure 5. A solitary, neonate-derived neuron, adrenergic/cholinergic/purinergic in function after 33 d in microculture. a-c, Intracellular recordings from the neuron; d-k, recordings as in Figure 1. a, A neuronal impulse was followed by an autaptic EPSP; this was blocked by addition of 1 mM hexamethonium (C,) to the perfusion fluid (b) and recovered when the C, was removed (c; scales at upper right apply to these 3 traces). d, In control (cant) solution, a single neuronal impulse (arrow) produced a large, rapid hyperpolarization followed by depolarization of the myocytes; a train at 16 Hz produced larger effects of these kinds with an apparent relative enhancement of the excitatory effect. e, In the presence of 5 PM

atropine sulfate (utrop), 50 x the concentration usually required to block a cholinergic effect, a train of neuronal impulses at 16 Hz produced a substantial depolarization (ca. 7 mv); the marked change from d is attributable to a relatively strong choline& effect. J When atenolol(20 CM) and phentolamine (0.1 PM) were added to the atropine (3B), the myocyte response was reduced to a small hyperpolarization; this change established an adrenergic effect. g, The residual hyperpolarization is shown at higher gain; it was blocked by addition of 8-PT (nominally 1 PM) to 3B (h), indicating that the effect in g was (rather weakly) purinergic. i, Washout of the 8-PT with 3B solution restored the purinergic effect. j, The moderate adrenergic effect is seen in isolation in the presence of 5 PM atropine (atrop) and 8-PT, nominally 1 PM. k, Removal of all the blockers restored the pattern of control responses; the responses in j and k were somewhat smaller than those recorded earlier in d and e. Voltage scales at the lower right refer to the myocyte traces only (d-k).

Fig. 5; cholinergic in Fig. 6) suggests a slow increase in receptor ations in the relative effectiveness of the 2 postsynaptic mech- occupancy, consistent with the release of small amounts oftrans- anisms is an alternative possibility. This would require that the mitter per impulse. Conversely, the small number of impulses cholinergic mechanisms be much more effective than the pur- needed to generate appreciable effects during the larger responses inergic in some myocyte islands, and that the converse occurred (cholinergic in Fig. 5; purinergic in Fig. 6) is consistent with the in other islands. While this alternative cannot be ruled out by release of appreciable amounts of transmitter per impulse. Vari- our evidence, it seems less likely; as indicated above, we have

cant a

3B/8-PT d

cant

aten/phentol

e 3B

4OmV(a,g,h) 8 mV(b-f) I

The Journal of Neuroscience, February 1987. 7(2) 387

aten/phentol/8-PT

C

strop/8-PT

h ’ I I!

Figure 6. A solitary, adult-derived neuron, adrenergic/cholinergic/purinergic in function after 55 d in microculture. The apparent relative strengths of the cholinergic and purinergic effects were different from those shown in Figure 5. a, In control solution (cant), a train of neuronal impulses (20 Hz) depolarized the myocytes. b, In the presence of 10 PM atenolol (aten) and 0.1 PM phentolamine @hen&), a somewhat longer train of neuronal impulses (20 Hz) produced a pronounced hyperpolarization; the change in myocyte response demonstrates an adrenergic effect. The hyperpolarization had a dual origin in a rather weak cholinergic effect, seen in isolation in c in the presence of atenolol (aten), phentolamine @hentoZ), and 10 PM

8-PT (and blocked by the addition of atropine in d), and a relatively strong purinergic effect, seen in isolation in e in the presence of atenolol, phentolamine and 0.1 PM atropine sulfate (3B). f; The relatively strong adrenergic effect is seen in isolation in the presence of atropine (drop) and 8-PT. d, A combination of all 4 blockers (3B/&PT) effectively eliminated the response of the myocytes to a long train of neuronal impulses at 20 Hz and established that no other transmitter slaved a substantial role. g, Washout of the four blockers restored the control (cant) response. h, A single neuronal impulse was followed by a s-mail autaptic EPSP.

not observed large variations from 1 microculture to another in the responsiveness of the myocytes to puffed agonists. Ob- servations on 2-neuron microcultures have demonstrated differences in transmitter release between neurons (e.g., Potter et al., 1986). In addition, the adrenergic-to-cholinergic transition appeared to occur gradually in our microcultures, with neurons remaining in the dual-function state for 2-3 weeks (Potter et al., 1986). Thus, variations in the levels of expression of the cholinergic and adrenergic states from neuron to neuron are to be expected. Evidence consistent with plasticity of the purinergic state was also found (Furshpan et al., 1986b), but we have no information about the time course of such a change.

Discussion NAE effects These effects differed from the adrenergic, cholinergic, and purinergic effects described in previous papers of this series (Fursh- pan et al., 1986a, b; Potter et al., 1986) both physiologically and pharmacologically: They were generally smaller and slower in time course, required more prolonged trains of neuronal stimulation at higher frequencies and longer intervals, and were resistant to concentrations of blockers that eliminated even intense effects of the other 3 kinds (Figs. l-3). Thus, there is strong reason to believe that these effects were not caused by secretion of any of the other 3 agents.

NAE effects were detected by using cardiac myocytes as target cells in the microcultures. We know of no previous report in which stimulation of the sympathetic innervation of the heart in vivo produced a similar slow depolarization resistant to adrenergic blockers, although a slow, nonadrenergic (e.g., peptid- ergic) effect is plausible in certain blood vessels and glands (see below). As discussed previously (Potter et al., 1986), the microcultured neurons were dissociated from the SCG, which in- nervates diverse target cells (several glands; smooth muscles of several types, including vascular; probably cardiac myocytes; brown fat; parasympathetic neurons in several ganglia) and then were placed at random, for assay, on cardiac myocytes. Pre- sumably, the majority of the microcultured neurons innervated other target cells in vivo. Thus, the NAE effects described here may be found in vivo in target tissues other than cardiac myocytes. The same may be true of other unconventional transmitter states described in this paper (see below).

The assays illustrated in Figure 2 strongly suggest that more than 1 transmitter was responsible for the NAE effect, and pro- vide pharmacological evidence that in some neurons the transmitter was serotonin. It is no surprise that serotonin excited the microcultured myocytes. While it is clear that the positive chronotopic effect of serotonin on the intact heart of some species is mediated via release of norepinephrine (NE) from sympathetic axons (e.g., Fozard and Mwaluko, 1976; Gothert and

388 Matsumoto et al. - Nonadrenergic Excitation: Diversity of Transmitter States

Duhrsen, 1979), it has been shown that serotonin acts directly on rat cardiac myocytes in culture (Higgins et al., 1981), and there is strong evidence for a direct effect in intact cardiac tissues of the rat, cat, and guinea pig (Trendelenburg, 1960; Sakai and Akima, 1979).

The accompanying paper (Sah and Matsumoto, 1987) reports evidence for the presence of serotonin in some, but not all, neurons in mass cultures, for uptake of serotonin from the rat serum in the growth medium, for release of serotonin by a calcium-dependent mechanism, and for synthesis of serotonin by some principal neurons cultured in conditioned medium. Recently, HappGlI et al. (1986) reported serotonin immunoreactivity in pre- and postnatal superior cervical ganglia of the rat; this immunoreactivity became rarer in putative principal neurons with age and was no longer detectable 90 d after birth. Collectively, these findings establish serotonin as an interesting and potentially important tranmitter in mammalian sympathetic principal neurons, as has been postulated for some sympathetic small, intensely fluorescent (SIF), interneurons (e.g., Verhofstad et al., 198 1) and demonstrated for myenteric neurons (see Gershon, 198 1, for discussion and references). Evi- dence that serotonin can act as a “false transmitter” (taken up and released, but not synthesized) in the sympathetic system is considered by Sah and Matsumoto (1987, the following paper).

The identity of the agent(s) responsible for the X effect re- mains a puzzle. So far, we have not succeeded in mimicking this effect with several peptides whose immunoreactivities have been reported in autonomic neurons, but further attempts will be made. The roster of peptides currently of interest in this connection is large and may well grow larger.

Variety of transmitter states

The existence of neurons that contain or release more than 1 transmitter is no longer a novelty (for reviews of coexistence and corelease, see HSkfelt et al., 1980, 1984; Potter et al., 198 1, 1986; Cuello, 1982; Furshpan et al., 1982; Lundberg and Hok- felt, 1983), and several types of peripheral and central neurons are now known to contain, as a class, many transmitter can- didates (e.g., myenteric, primary sensory, autonomic preganglionic, and amacrine neurons; for reviews, see the papers just cited and Furness and Costa, 1980). However, we know of no precedent for the diversity of synaptic functions (combinations of released transmitters in a single type of neuron) reported here (Fig. 4). Nor do we know of a prior report of a neuron that secretes 4 transmitters (Figs. 3 and 4), although several types of neurons that coexpress 3 or more transmitter immunoreactivities have been reported (e.g., Johansson et al., 198 1; Hokfelt et al., 1983; Sawchenko and Swanson, 1985).

It is noteworthy that all the states listed in Figure 4 were seen in adult-derived neurons (stars) as were seen in neonate-derived neurons (circles), except for UNAE, one of the rare states in this sample of neurons. At this early stage in the investigation, we see no reason to assume that control of the transmitter repertoire in adult-derived neurons in microculture is qualitatively different from that in neonate-derived neurons.

Is the diversity of states shown in Figure 4 a culture artifact?

We previously reviewed the apparent normality of the microcultured neurons with respect to adrenergic and cholinergic states, the adrenergic-to-cholinergic transition, and several other im-

al., 1986), and discussed evidence consistent with the expression in vivo of purinergic function (Furshpan et al., 1986b). The observed adrenergic-to-cholinergic plasticity (Potter et al., 1986) may have added choline@ function to neonate-derived neurons that would have lacked this function in vivo and deleted adrenergic function from some neurons that would have pos- sessed it in vivo. However, this transition is part of the normal developmental repertoire (Landis and Keefe, 1983). We also provided evidence for this plasticity in adult-derived neurons and urged caution in accepting the conventional view that adrenergic and choline@ functions are never simultaneously expressed by adult sympathetic neurons in vivo (Potter et al., 1986).

We know little about the plasticity of the nonadrenergic, non- cholinergic transmitter states, or about factors that control expression ofthese states. It may be that many neurons in culture express transmitters, or combinations of transmitters, that they would not individually have expressed in vivo, as the concentrations of control factors and the timing of their appearance are likely to be different in vivo and in vitro. However, it is interesting that in microculture the neurons did not lapse into 1 or a few degenerate states or into all possible combinations of the 5 agents (25 combinations; 3 1 positive states plus a null state); neither did all expressed states occur with equal frequency. For these reasons, and because of the growing evidence for multiple-transmitter states in mammalian sympathetic principal neurons in vivo, it is possible that each of the states in Figure 4 (bearing in mind the qualifications described above) occurs in some neurons in vivo, either in development or during the various contingencies of adult life.

With regard to the diversity summarized in Figure 4, several limitations of the microculture assays should be emphasized. First, it is obviously possible that more stringent tests (e.g., longer trains of neuronal stimuli at higher frequency and at longer intervals between tests, combined with records from the myocytes at higher gain) would detect each state at a younger culture age than that in Figure 4, and we see no reason to doubt that in further work each state will be found at a greater culture age than is shown in Figure 4. It is notable that the complex state (A/C/P/X) appeared as early as 20 d and as late as 127 d (an adult-derived neuron; the oldest microculture included in Fig. 4). A second limitation is that the apparent transmitter repertoire of a microcultured neuron is dependent on the nature of the target cell and its receptors. For example, vas deferens myocytes are relatively insensitive to adenosine and more sensitive to serotonin than are cardiac myocytes in microculture (D. Sah, unpublished observations). A neuron whose true repertoire included adenosinergic and serotonergic function might be categorized differently in the 2 assay systems. This consid- eration affects the interpretation of the transmitter categories listed in Figure 4. It is obviously possible that many of the neurons actually expressed an additional transmitter or transmitters to which the cardiac myocytes were insensitive. Perhaps the full repertoire of released transmitters in the microcultured neurons will not be known with confidence until the neurons have been assayed against every type of cell sympathetically innervated in vivo. Given that the transmitter repertoire of mammalian sympathetic principal neurons in several different ganglia may now include over a dozen agents [with varying degrees of confidence we may list NE, ACh, dopamine, epi- nephrine, 5-HT, ATP, adenosine, VIP, neuropeptide HI (PHI),

portant neuronal properties (Furshpan et al., 1986a; Potter et somatostatin, substance P, the enkephalins, neuropeptide Y

The Journal of Neuroscience, February 1987, T(2) 389

(NPY), calcitonin-gene related polypeptide (CGRP), and a va- sopressin-like substance], the potential diversity of transmitter states and control of target cells is much greater than that seen so far in microcultures or reported in vivo. A third limitation of the microculture assays is that however sensitive they may be to electrogenic synaptic effects, the assays would not detect electrically silent effects.

Given the second and third limitations and the rapidly growing evidence of multiple transmitters in a variety of neurons, the neurons categorized as A, C, or P in Figure 4 seem especially suspect: Were these neurons really monofunctional? Evidence for monofunctional neurons in the peripheral nervous system in vivo is similarly suspect. Even the a-motoneurons of vertebrates probably secrete ATP as well as ACh, as there is evidence that the 2 agents are co-stored (e.g., Carlson et al., 1978) and released (Silinsky, 1975) and that at least some ar-motoneurons are immunoreactive for CGRP (Rosenfeld et al., 1983). In fact, we know of no compelling evidence for a monofunctional neuron in the peripheral nervous system of vertebrates.

Graded relative strengths of adrenergic, cholinergic, and purinergic functions Large variations in the relative strengths of transmitter effects, illustrated by a comparison of Figures 5 and 6, were routinely observed. They were seen in both neonate- and adult-derived neurons. As indicated above, it seems likely that much of this variation is accounted for by differences in the level of expression of transmitter functions in the neurons. According to the traditional view, transmitters are expressed in the adult nervous system in vivo either full-OFF or full-ON, with activity-dependent modulation during the full-ON state. However, there is evidence for transmitter plasticity in adult neurons in vivo (see Black et al., 1984; Bjijrkhmd et al., 1985; Sawchenko and Swan- son, 1985; see also Gesser and Larsson, 1985) and broad variation in the expression of transmitter functions is expected during a transition.

An aspect of synaptic action by microcultured neurons that individually exert both relatively strong and weak effects may be of interest. In the neurons of Figures 1 and 5, single impulses or the first few impulses in a train produced primarily a cholinergic effect on the myocytes (Figs. la and 5d), but later in the train a purinergic effect was also exerted (Figs. lb and 5g). The neuron of Figure 6 showed the reverse behavior; early in the train the inhibitory effect was mainly purinergic, while later a choline& effect was also present (Fig. 6c vs 6e). Thus, the relative effectiveness of the cosecreted transmitters depended on the duration ofthe train. A change in the relative effectiveness of cosecreted transmitters with altered frequency of a train has been reported, for example, by Lundberg et al. (1982; see also fig. 5 of Furshpan et al., 1986b).

Is there a plausible role for triple-, quadruple-, or even higher-order function in neurons? There is a growing awareness of the potential complexity of the control of target tissues by active neurons. The simultaneous control of a gland or muscle with its local blood supply was of concern to Dale and Gaddum (1930). More recently, evidence for control of presynaptic endings and postsynaptic receptor sensitivity (or effectiveness) has been obtained in several systems; the possible roles of multiple transmitters in control of these functions and of vessels have been reviewed by Hijkfelt

et al. (1980) and Lundberg and Hiikfelt (1983). There is also rising interest in a more “silent” control of target cell metabolism-for example, of tyrosine hydroxylase by several transmitters in sympathetic ganglia, with time scales of hours or days (e.g., Zigmond, 1980; Ip et al., 1982, 1983) or of glycogen metabolism by several transmitters in mouse cortex (e.g., Magis- tretti et al., 198 1, 1983). There is evidence that endocrine and exocrine cells sometimes display high orders of secretory function, presumably related to the complexity of their target do- mains. For example, adrenal medullary cells are reported to collectively contain at least a dozen agents (see, for example, Lundberg et al., 1979; Role et al., 1979; Livett et al., 1981; Daniels et al., 1982; Day et al., 1982; Wilson et al., 1982; Eiden et al., 1983; Verhofstad and Jonsson, 1983; Black et al., 1984; Kataoka et al., 1984; Rokaeus et al., 1984; Vamdell et al., 1984); each hepatocyte is thought to secrete a large number of proteins and enzymes into the blood (e.g., albumin, most alpha- and beta-globulins, fibrinogen, thrombin, and other clotting factors), as well as bile into the bile canaliculi; each pancreatic acinar cell is thought to secrete the full repertoire of digestive enzymes (proteolytic, lipolytic, and amylolytic).

There is no known upper limit to the number of agents secreted by a single endocrine or exocrine cell; it appears that each cell secretes as many agents as there are properties of the ac- cessible cellular and extracellular environments that require linked control. It is reasonable to assume that the same will prove true of neurons.

References Bellardinelli, L., and G. Isenberg (1983) Isolated atria1 myocytes:

Adenosine and acetylcholine increase potassium conductance. Am. J. Physiol. 244: H734-H737.

Bjorklund, H., T. Hokfelt, M. Goldstein, L. Terenius, and L. Olson (1985) Appearance of the noradrenergic markers tyrosine hydroxylase and neuropeptide Y in cholinergic nerves of the iris following sympathectomy: J: Neurosci. 5: 1633-l 643.

Black. I. B.. J. E. Adler. C. F. Drevms. G. M. Jonakait. D. M. Katz. E. F. LaGamma and K. M. Markey (1984) Neurotransmitter plasticity at the molecular level. Science 225: 1266-l 270.

Bray, D. (1970) Surface movements during the growth of single ex- planted neurons. Proc. Natl. Acad. Sci. USA 65: 905-910.

Brodie, B. B., and P. Shore (1957) A concept for a role of serotonin and norepinephrine as chemical mediators in the brain. Ann. NY Acad. Sci. 66: 63 l-642.

Carlson, S. S., J. A. Wagner, and R. B. Kelly (1978) Purification of synaptic vesicles from elasmobranch electric organ and the use of biophysical criteria to demonstrate purity. Biochemistry 17: 1188- 1206.

Choi, D., and G. D. Fischbach (1981) GABA conductance of chick spinal cord and dorsal root ganglion neurons in cell culture. J. Neu- rophysiol. 45: 605-620.

Cuello, A. C. (ed.) (1982) Co-Transmission, Macmillan, London. Dale, H. H.. and J. H. Gaddum (1930) Reactions of denervated vol-

untary muscle and their bearing on the mode of action of parasympathetic and related nerves. J. Physiol. (Lond.) 70: 109-144.

Daniels. A. J.. G. Dean. 0. H. Viveros. and E. J. Diliberto. Jr. (1982) Secretion of newly taken-up ascorbic’ acid by adrenomedullaj chro: maffin cells. Science 216: 737-739.

Day, R., D. Denis, J. Barabe, S. St. Pierre, and S. Lemaire (1982) Dynorphin in bovine adrenal medulla. I. Detection in glandular and cellular extracts and secretion from isolated chromaffin cells. Int. J. Pept. Prot. Res. 19: 10-17.

Eiden, L. E., R. L. Eskay, J. Scott, H. Pollard, and A. J. Hot&kiss (1983) Primary cultures of bovine chromaffin cells synthesize and secrete vasoactive intestinal polypeptide (VIP). Life Sci. 33: 687-694.

Fozard, J. R., and G. M. Mwaluko (1976) Mechanism of the indirect sympathomimetic effect of 5-hydrotryptamine on the isolated heart of the rabbit. Br. J. Pharmacol. 57: 115-125.

390 Matsumoto et al. * Nonadrenergic Excitation: Diversity of Transmitter States

Fumess, J. B., and M. Costa (1980) Types of nerves in the enteric nervous system. Neuroscience 5: l-20.. -

Furshnan. E. J.. D. D. Potter. and S. C. Landis (1982) On the trans- mitier repertoire of sympathetic neurons in culture. Harvey Lect. 76: 149-191.

Furshpan, E. J., S. C. Landis, S. G. Matsumoto, and D. D. Potter (1986a) Synaptic functions in rat sympathetic neurons in microcultures. I. Secretion of norepinephrine and acetylcholine. J. Neurosci. 6: 1061-1079.

Furshpan, E. J., D. D. Potter, and S. G. Matsumoto (1986b) Synaptic functions in rat sympathetic neurons in microcultures. III. A purinergic effect on cardiac myocytes. J. Neurosci. 6: 1099-l 107.

Gershon, M. D. (198 1) The enteric nervous system. Annu. Rev. Neu- rosci. 4: 227-272.

Gesser, B. P., and L.-I. Larsson (1985) Changes from enkephalin-like to gastrin/cholecystokinin-like immunoreactivity in snail neurons. J. Neurosci. 5: 1412-1417.

Gothert, M., and U. Duhrsen (1979) Effects of 5-hydroxytryptamine and related compounds on the sympathetic nerves of the rabbit heart. Naunyn Schmiedebergs Arch. Pharmacol. 308: 9-18.

Hapippala, O., H. P%iv%rinta, S. Soinila, and H. Steinbusch (1986) Pre- and postnatal development of 5-hydroxytryptamine-immunoreactive cells in the superior cervical ganglion of the rat. J. Autonom. Nerv. Syst. 15: 21-31.

- -

Higgins, T. J. C., P. J. Bailey, and D. Allsopp (1981) Mechanism of stimulation of cardiac myocyte beating rate by 5-hydroxytryptamine. Life Sci. 28: 999-1005. _ - - - - - --

Hirst, G. D. S., and T. 0. Neild (1980) Reply to J. C. McGrath. Nature 288: 302.

H&felt, T., J. M. Lundberg, M. Schultzberg, 0. Johansson, A. Ljung- dahl, and J. Rehfeld (1980) Coexistence of peptides and putative transmitters in neurons. In Neural Peptides and Neuronal Commu- nication, E. Costa and M. Trabucchi, eds., pp. l-23, Raven, New York.

Hiikfelt, T., J. Fahrenkrug, K. Tatemoto, V. Mutt, S. Werner, A.-L. Hulting, L. Terenius, and K. J. Chang (1983) The PHI (PHI27)/ corticotropin-releasing factor/enkephalin immunoreactive hypotha- lamic neurons: Possible morphological basis for integrated control of prolactin, corticotropin and growth hormone secretion. Proc. Natl. Acad. Sci. USA 80: 895-898.

Hiikfelt, T., 0. Johansson, and M. Goldstein (1984) Chemical anat- omy of the brain. Science 225: 1326-1334.

Ip, N. Y., C. K. Ho, and R. E. Zigmond (1982) Secretin and vasoactive intestinal peptide acutely increase tyrosine 3-monooxygenase activity in the rat superior cervical ganglion. Proc. Natl. Acad. Sci. USA 79: 7566-7569.

Ip, N. Y., R. L. Perlman, and R. E. Zigmond (1983) Acute transsyn- aptic regulation of tyrosine 3-monooxygenase activity in the rat su- p&or c&&al ganglion: Evidence for both cholinergic and nonchol- ineraic mechanisms. Proc. Natl. Acad. Sci. USA 80: 2081-2085.

Joha&son, O., T. H&felt, B. Pemow, S. L. Jeffcoate, N. White, H. W. M. Steinbusch, A. A. J. Verhofstad, P. C. Emson, and E. Spindel (198 1) Immunohistochemical support for three putative transmitters in one neuron: Coexistence of 5-hydroxytryptamine, substance P- and thyrotropin releasing hormone-like immunoreactivity in medullary neurons projecting to the spinal cord. Neuroscience 6: 18 57-l 88 1.

Kataoka, Y., Y. Gutman, A. Guidotti, P. Panula, J. Wroblewski, D. Cosenza-Murphy, J. Y. Wu, and E. Costa (1984) Intrinsic GABA- ergic system of adrenal chromaffin cells. Proc. Natl. Acad. Sci. USA 81: 32 18-3222.

Landis, S. C. (1976) Rat sympathetic neurons and cardiac myocytes developing in microcultures: Correlation of the fine structure of endings with nemotransmitter function in single neurons. Proc. Natl. Acad. Sci. USA 73: 4220-4224.

Landis, S. C., and D. Keefe (1983) Evidence for nemotransmitter plasticity in vivo: developmental changes in properties of cholinergic sympathetic neurons. Dev. Biol. 98: 349-372.

Livett, B. G., D. M. Dean, L. G. Whelan, S. Udenfriend, and J. Rossier (198 1) Co-release of enkephalin and catecholamines from adrenal chromaffin cells. Nature 289: 3 17-3 19.

Lundberg, J. M., and T. H&felt (1983) Coexistence of peptides and classical neurotransmitters. Trends Neurosci. 6: 325-333.

Lundberg, J. M., B. Hamberger, M. Schultzberg, T. Hokfelt, P.-O. Gran- berg, S. Efendic, L. Terenius, M. Goldstein, and R. Luft (1979) Enkephalin- and somatostatin-like immunoreactivities in human adrenal medulla and pheochromocytoma. Proc. Natl. Acad. Sci. USA 76: 4079-4083.

Lundberg, J. M., A. Anggard, and J. Fahrenkrug (1982) VIP as a mediator of hexamethonium-sensitive, atropine-resistant vasodila- tion in the cat tongue. Acta Physiol. Scand. 116: 387-392.

Magistretti, P. J., J. H. Morrison, W. J. Shoemaker, V. Sapin, and F. E. Bloom (198 1) Vasoactive intestinal polypeptide induces glyco- genolysis in mouse cortical slices: A possible regulatory mechanism for the local control of enerav metabolism. Proc. Natl. Acad. Sci. USA -_ 78: 6535-6539.

Magistretti, P. J., J. H. Morrison, W. J. Shoemaker, and F. E. Bloom (1983) Effect of 6-hydroxydopamine lesions on norepinephrine-in- duced PH] glycogen hydrolysis in mouse cortical slices. Brain Res. 261: 159-162.

Neild, T. O., and E. Zelcer (1982) Noradrenergic neuromuscular transmission with special reference to arterial smooth muscle. Prog. Neu- robiol. 19: 141-158.

Potter. D. D.. E. J. Furshnan. and S. C. Landis (1981) Multinle- transmitter ‘status and %&‘s Principle.” Neurosci. Cornmen{ I: l-9.

Potter, D. D., S. C. Landis, S. G. Matsumoto, and E. J. Furshpan (1986) Synaptic functions in rat sympathetic neurons in microcultures. II. Adrenergic/cholinergic dual status and plasticity. J. Neurosci. 6: 1080- 1098.

Rokaeus, A., G. Fried, and J. M. Lundberg (1984) Occurrence, storage and release of neurotensin-like immunoreactivity from the adrenal gland. Acta Physiol. Stand. 120: 373-380.

Role. L. W.. R. L. Perlman. and S. E. Leeman (1979) Somatostatin and substance P inhibit ‘catecholamine secretion f;om guinea-pig chromaffin cells. Sot. Neurosci. Abstr. 5: 203.1.

Sah, D. W. Y., and S. G. Matsumoto (1987) Evidence for serotonin synthesis, uptake, and release in dissociated rat sympathetic neurons in culture. J. Neurosci. 7: 39 l-400.

Sakai, K., and M. Akima (1979) An analysis of the stimulant effects of 5-hydroxytryptamine on isolated, blood-perfused rat heart. Eur. J. Pharmacol. 55: 42 l-424.

Sawchenko, P. E., and L. W. Swanson (1985) Localization, colocali- zation and plasticity of corticotropin-releasing factor immunoreactivity in rat brain. Fed. Proc. 44: 221-227.

Silinsky, E. M. (1975) On the association between transmission, secretion and the release of adenine nucleotides from mammalian motor nerve terminals. J. Physiol. (Lond.) 247: 145-162.

Trendelenburg, U. (1960) The action of histamine and 5-hydroxytryptamine on isolated mammalian atria. J. Phannacol. Exp. Ther. 130: 450-460.

Vamdell, I. M., J. M. Polak, J. M. Allen, G. Terenghi, and S. R. Bloom (1984) Neuropeptide tyrosine (NPY) immunoreactivity in norepinephrine-containing cells and nerves of the mammalian adrenal gland. Endocrinology 114: 1460-1462.

Verhofstad, A. A. J., and G. Jonsson (1983) Immunohistochemical and neurochemical evidence for the presence of serotonin in the adrenal medulla of the rat. Neuroscience 10: 1443-1453.

Verhofstad, A. A. J., H. W. M. Steinbusch, B. Penke, J. Varga, and H. W. J. Joosten (198 1) Serotonin-immunoreactive cells in the superior cervical ganglion of the rat: Evidence for the existence of separate serotonin- and catecholamine-containing small ganglionic cells. Brain Res. 212: 39-49.

Walicke, P. A., R. B. Campenot, and P. H. Patterson (1977) Deter- mination of transmitter function by neuronal activity. Proc. Natl. Acad. Sci. USA 74: 5767-5771.

Wilson. S. P.. K.-J. Chane. and 0. H. Viveros (1982) Prooortional IT \ , .

secretion of opioid peptides and catecholamines from adrenal chromaffin cells in culture. J. Neurosci. 2: 1150-l 156.

Zigmond, R. E. (1980) The long-term regulation of ganglionic tyrosine hydroxylase by preganglionic nerve activity. Fed. Proc. 39: 3003- 3008.

Documents

Synaptic Functions in Rat Sympathetic Neurons in Microcultures. IV