Postnatal rat sympathetic neurons in culture. II. Synaptic transmission by postnatal neurons.

JOURNALOF NEUROPHYSIOLOGY Vol. 42, No. 5. September 1979. Printed

in U.S.A.

Postnatal Rat. Sympathetic Neurons in Culture. II. Synaptic Transmission by Postnatal Neurons E. WAKSHULL,

M. I. JOHNSON,

AND

H. BURTON

Depurtment oj‘ Embryology, Curnegie Institlrtion oj’ Wcrshington, Bcrltimow, Maryland 21210; crnd Departments (~‘ Anutomy-Nt~urobiolo~v ctnd P/2VSioloS?‘-BiO~~~?V,~i(..s, . Washington University School oj’ Medicine, St. Louis, Miss&i 63 I 1;

SUMMARY

AND

CONCLUSIONS

I. It was shown in the preceding paper that postnatally derived rat superior cervical ganglion neurons (SCGN) will grow in dissociated cell culture and form functional synaptic connections with each other. In this report, synaptic transmission by the postnatal SCGN is detailed. 2. Synaptic interactions between SCGN were blocked by the nicotinic cholinergic antagonist hexamathonium (C-6)) indicating that acetylcholine was the transmitter substance used by these neurons. This was found to be the case even for neurons taken from 12.5wk-old animals. 3. In a few cases, the P-adrenergic blocking agent, propranolol, was found to block synaptic potentials, suggesting that a catecholamine might be involved in the transmission process. The possible mechanisms of this involvement are discussed. 4. SCGN taken from up to lo-wk-old rats were able to form functional synaptic contacts with cocultured skeletal muscle cells. These interactions were sensitive to low external Ca”+ and to l-2 PM d-tubocurarine (d-TC). 5. It is concluded that even adult SCGN retain a certain amount of neurotransmitter “plasticity” when grown under appropriate culture conditions. From the data on the neuron-neuron and SCGN-skeletal muscle interactions, it is suggested that a matching of presynaptic transmitter with postsynaptic receptor is a sufficient condition for the formation of functional nerve-target interactions. 1426

INTRODUCTION

In the preceding paper we demonstrated that sympathetic neurons derived from postnatal rats could be grown in long-term, dissociated cell cultures. The most interesting finding was that synaptic interactions developed between the SCGN (31). Previous work on embryonically derived SCGN have demonstrated that the synaptic connections formed by these normally adrenergic neurons were cholinergic (4, 8-10, 17, 18, 30). However, recent evidence, utilizing explant cultures, showed that the levels of choline acetyltransferase declined to extremely low levels of activity when the explant cultures were prepared from animals older than 3-4 wk (24). These results suggested that the ability demonstrated by embryonic adrenergic neurons to shift neurotransmitters was age dependent. However, these primarily biochemical results were derived from explant cultures in which a more organotypic environment is provided for the principal neurons than is available in dissociated cell cultures. The purpose of the present study was to examine the pharmacology of the synapses formed by dissociated postnatal neurons on other SCGN and on skeletal muscle cells. This paper demonstrates that the synapses formed in culture by older SCGN (up to 12.5 wk of age) are also cholinergic. The results suggest that at least some fully differentiated sympathetic neurons retain the ability to express cholinergic transmitter mechanisms in culture even though these neurons have fully interacted with their peripheral

0022-3077/79/0000-OOOO$O 1.25 Copyright 0 1979 The American Physiological Society

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 16, 2019

SYMPATHETIC

NEURONS

target organs in the animal prior to explantation. Langley and Anderson (11) described experiments in adult cats that suggested that postganglionic sympathetic fibers from the superior cervical ganglion would not functionally innervate skeletal muscle (tongue) because the nerve endings were unable to affect the muscle cells. Although unknown to Langley and Anderson at the time, the fact that sympathetic neurons release norepinephrine (NE) from their terminals would make them unable to affect the acetylcholine receptors on muscle cells. However, the finding that, under certain culture conditions, dissociated SCGN from embryonic tissue become cholinergic (15, 19-21, 28, 30) and are able, in fact, to form physiologically detectable junctions with cocultured myotubes (16, 29, 30) suggests that more sensitive techniques might reveal interactions between SCGN and skeletal muscle even in the animal. Given the biochemical and morphological evidence that SCGN explant cultures from adult animals seemed to remain “adrenergic” (2, 24), we decided to test the ability of these older cells to form functional neuromuscular contacts. Preliminary results from these studies have been published (28, 29). METHODS

Pwparu

tion oj’ cdtures

POSTNATAL SCGN. The procedure for establishing postnatal SCGN in dissociated culture has been described (3 1). SKELETAL MUSCLE. Cultures of skeletal muscle were prepared from 17- to 21-day embryonic rat hindlimbs. Both muscle chunks and dissociated muscle were used. In the first case, connective tissue was removed mechanically and the muscle cut into approximately 1 mm” chunks; three to five chunks were put on each collagencoated culture dish. Within 24 h, myoblasts migrated away from the chunks and began proliferating. Myotube formation from fusion of myoblasts occurred after a couple of days and continued for at least a week, at which point either explanted or dissociated SCG were added. Superior cervical ganglia (SCG) explants were taken from 5-day-old rat pups and adult (approximately 250 g) rats. Ganglia were stripped of their connective tissue sheath and cut into l-2 mm’{ chunks. Two or three explants were added to each muscle culture dish.

IN VITRO.

II

1427

Dissociated muscle cultures were prepared by treating small muscle chunks with 0.1% trypsin (Nutritional Biologicals) for 45 min at 35°C and pH 7.6-7.8. The muscle was subsequently rinsed, triturated in a small-bore glass pipette, filtered through a IO-pm-pore Nitex (Tetko, Rolling Meadows, IL) filter, and plated onto a collagencoated culture dish. Myoblasts proliferated extensively and myotubes formed after 3-4 days in vitro. Neurons were added after 7- 10 days. The feed used to establish the muscle in vitro was either the standard medium used for neurons (30) or a feed containing 2-5s chick embryo extract, 10% horse serum, 85% Eagles minimum essential media (MEM) (Gibco, Grand Island, NY), 2 mM glutamine, and 600 mg/lOO ml glucose. The latter medium was found to promote more vigorous myotube formation than did the standard neuronal medium. Once the myotubes had formed, they could be adequately maintained for many weeks on the standard medium, which was used exclusively after the neurons were added. Muscle cultures were fed daily until the neurons were added; feedings were then 3 times a week. Cultures were maintained in a humidified atmosphere of 5% CO., at 35°C. In both chunk and dissociated muscle cultures, fusion of myoblasts began after a few days of active proliferation, the dissociated cultures taking slightly longer, and resulted in multinucleate, often branching myotubes, which varied in length from less than 100 ,um to several millimeters. No cross striations could be seen, with or without neurons present, although the muscle cells were capable of active contraction. In fact, spontaneous contractions were commonly seen after a week or so in vitro. Myotubes tended to become quiescent when put on the standard (neuronal) medium. The reason for this is unclear and was not systematically investigated. After a few weeks on the standard medium, myotubes began to deteriorate as evidenced by the accumulation of refractile (phase bright) particles in the sarcoplasm and rounding up of the normally elongate muscle cells. Myotubes that did not give an action potential (AP) or failed to contract on direct stimulation through the recording electrode were abandoned. Not only were the muscle cells capable of spontaneous contractile activity, but direct extracellular and intracellular stimulation could evoke action potentials and contractions. The mean amplitude of the APs elicited by direct intracellular stimulation was 97 _t 3.4 mV (mean 2 SE; n = 20; range = 48- 120), and the resting potential, measured by comparing the difference in electrode potential before and immediately after removal from the cell was 63 it 2.1 mV (rt


1428

WAKSHULL,

= 18: range = 38-78). determined.

Input impedance

JOHNSON,

AND

BURTON

was not

shows, C-6 was the only drug that consistently reduced or eliminated the synaptic potentials. The consecutive effects of C-6, Electrophysiology, phurmacology, and propranolol, atropine, and phenoxybenzamine horseradish peroxidase luheling on the sameneuronal pair are shown in Fig. 1. Most of the anatomical and physiological The only clear effect with drug perfusion for methods were the same as those used in the prethis pair of cells is the complete block of the ceding paper. However, no intracellular recordsynaptic potential with C-6. The potential reings were attempted from SCGN grown in excovered after 5 min in control media (trace 6 of plant cultures due to the overgrowth of supportFig. 2). Both blockade and recovery of synaping cells and a tough fibrous material found over the explants. For this reason, bipolar extracellutic potentials by C-6 (or propranolol; see lar stimulation of the ganglion explants was embelow) were graded events. A decrease in ployed. Two glass pipettes with tip diameters of amplitude was almost always apparent after 5-7 pm and filled with 4 M NaCl were separated 5 min perfusion time with the drug and a by lo-15 pm and applied to the edge of an excomplete block was achieved usually within plant closest to the myotube to be impaled. At 10 min (depending on the initial response threshold levels, stimulation evoked contracamplitude and perfusion rate, both of which tions of a variable number of myotubes on any could vary considerably). Recovery ocparticular trial. Increased stimulus strength recurred more slowly, usually over a 20- to 30cruited more myotubes until virtually all the min time course. In two of the three cases myotubes were contracting 1: 1 with stimulus frein which C-6 did not block the potential, the quency (usually 1 Hz). In order to avoid misinterpretations resulting from inadvertent direct culture was perfused for 7 and 9 min, respecstimulation of the muscle, evoked contractions tively. In the third case, C-6 was being perwere reversibly blocked with 2.5 mM of &TC or fused into the culture for 10 min before the low extracellular Ca”+ with constant-stimulation synaptic potential was even found. parameters. Pharmacological antagonists to variFive synaptic potentials (of a total of 29) ous putative neurotransmitter substances were were found to be sensitive to the P-adrenapplied via a continuous perfusion system as preergic blocker propranolol. An example of a viously described (9). propranolol block is shown in Fig. 2,+ A few minutes after the potential recovered in RESULTS the control media, the culture was perfused Pharmacology of synaptic interactions with C-6 (Fig. 2,), while maintaining the rebetween SCGN cordings from the same cell pair. It can be In order to determine the nature of the seen that the synaptic potential was also chemical transmitter being used by these blocked by this drug (Fig. 22). This was one neurons, various pharmacological blocking of two synaptic responses that was found to agents were perfused through the cultures be blocked by both propranolol and C-6. The ar-adrenergic antagonist phenoxybenwhile recording synaptic potentials. The results are summarized in Table 1. As the data zamine (PB) gave inconsistent results (see TABLE 1. Ejylcc t oj vurious putative neurotransmitter potentials in postnatal SCGN cultures

Drug

No Trials

Inhibit

No Effect

antagonists on synaptic

Incomplete

Ambiguous

Inhibit, either a reduction in amplitude or complete elimination. Incomplete, the cells were lost before in control media was achieved. Ambiguous, the effect of the drug was unclear or a second trial on synaptic potential gave a result different from the first. * An increase or facilitation of synaptic amplitude. None of the trials were definitive with phenoxybenzamine since a complete block was never A statistical analysis of mean amplitude with and without the drug was not done.

Other

recovery the same potential achieved.


SYMPATHETIC

NEURONS

PROP

IN VITRO.

II

1429 ATRP

2 x10-5 M

~xIO.~M

. f-+--2PB

2~10-~M

C-6

5

4

10-A M

I

I

IO mV 50 msec

Effect of various blocking agents on a single postnatal neuron-neuron synaptic potential. 1, synaptic FIG. 1. potential (upper trace) in control media, produced by excitation of driver cell (lower trace). 2, propranolol (PROP; 2 x 10-a; M) was perfused for 6.5 min; the increased size of synaptic potential was due to improved membrane sealing around the electrode (see text). 3, atropine (ATRP; 3 X 1O-f’ M) was perfused for 7 min. a slight decrease in amplitude due to the atropine may have occurred. 4, phenoxybenzamine (PB; 2 x 10m5 M) has no effect, after 8 min. 5, C-6 ( 10W4 M) abolishes the response in 3 min, which recovers after 5 min in control (6). Calibration: all traces, 10 mV, 50 ms. SCGN age in vivo, 28 days; age in vitro, 14 days.

A

B

i I

I prop

hdOw5M

C-6

10-4M

10

L

mV

50ms

r-Examples of a propranololand C-6-sensitive synaptic potential. Driver neuron activity not shown. FIG. 2. perfusion with propranolol stimulation of driver neuron evoked a multiple peaked synaptic response. A,: A,: response recovered on return to control medium. B,: (prop, 2 x 10 3 M) almost eliminated the potential. A,: has increased in size (possibly due to an improvesame neuronal pair, several minutes after A,; the response perfusion with C-6 blocked the synaptic poment in the penetration) and lost its multiple-peaked appearance. B,: and assumes its original appearance in control medium. Initial deflections in tential. Bt3: the response recovered all traces are the stimulus artifact. Calibration: Al-:$, B1+ 10 mV, 50 ms. SCGN age in vivo, 28 days; age in vitro, 28 days.


1430

WAKSHULL,

JOHNSON,

AND

BURTON

hctwwn Table 1). Only relatively small changes in Junc*tiond trunsmission myotubes potential amplitude (both increasing and de- postnatal SCGN and skektd creasing) were found, and an insufficient SCG EXPLANTS AND MUSCLE. Typical neusample of synaptic potentials was obtained ritic halos grew from explants derived from to differentiate statistically between normal both Sday-old and (with a slight delay) adult amplitude fluctuations (see Ref. 32) and pos- rats, and were seen to course toward and sible drug effects. In addition, improving through the muscle chunks. No preferential (increased resting potential) or deteriorating growth of fibers either toward or away from (decreased resting potential) penetrations the muscle was noted. Stimulation of the would also cause changes in response ampliganglion explants evoked vigorous synchrotudes. Chlorpromazine (dopamine antago- nous contractions in virtually every myonist) caused no change and atropine had no tube in the closest muscle chunk. Contractions evoked by stimulation of both the effect except at high concentrations.

d-TC

2.5 x w6M

20mV

L

20 50msec

1

innervation of skeletal myotubes by adult SCG explants. 1, extracellular stimulation of FIG. 3. Cholinergic ganglion explant (break in base line, lower trace) evoked an action potential in the muscle cell (upper trace). The initial biphasic deflection of the upper trace is the stimulation artifact recorded by the intracellular muscle electrode. Perfusion with tl-tubocurarine (d-TC; 2.5 x 10 ” M) completely blocked the response (not shown). 2, after a few minutes in control solution, small “end-plate” potentials were apparent and many failures. 3, continued perfusion with control medium resulted in larger amplitude responses; note the varied latency and multiplepeaked potentials. 4, responses eventually achieved suprathreshold levels. The vertical line on the left side of 2,3, and 4 is the fused stimulation artifacts from the individual traces. Calibration: 1, 20 mV, 20 ms; 2, 3,4, 20 mV, 50 ms. SCG explants age in vivo, approximately 5 wk; age in vitro, 19 days.


SYMPATHETIC

NEURONS

IN VITRO.

II

1431

FIG. 4. Camera lucida drawing of postnatal SCGN and skeletal muscle ceils. Only porttons of myotubes are shown. Intracellular recordmg from the neuron was matntained whtle nearby myotubes were Impaled. Myotubes 1 and 2 were found to be innervated by the neuron. which produced suprathreshold junctional potentials and contractions. No innervation of myotube 3 was detected. Processes ofthe neuron are more restrtcted in thetr pattern of growth than are SCGN in culture by themselves. As was the case for every neuron labeled in culture\ with muscle cells present, the processes were ortented parallel to the long axi\ of the myotubes. Processes do. however, cross myotubes perpendicular to their long axis. SCGN age in vivo, 70 days; age in vitro. 27 day\. Bar: 100 pm.

j-day and adult explants were reversibly blocked by low Ca2+ medium. The inhibition of contraction in low Ca2+ could be overcome by increasing stimulus strengths at the SCG explant or by placing the stimulating electrodes directly over the muscle. Similarly, d-TC reversibly inhibited nerve induced muscle activation by both the young and old SCG explants (Fig. 31J. The d-TC block could also be overcome by direct stimulation of the muscle. In the example shown, extracellular stimulation of an SCG explant taken from an adult rat evoked a suprathreshold junctional potential (Fig. 3,, upper trace). The initial biphasic voltage displacement is the stimulus artifact. After a complete block with d-TC perfusion, small evoked “end-plate” potentials (EPPS) of varying size (and with some failures) reappeared following several minutes in control

solution (Fig. 3.J. Continued perfusion with control media resulted in larger EPPS, some of which had multiple-peak components (Fig. 3:,; see below). Eventually, nerveevoked APs were again produced (Fig. 3,). DISSOCIATED SCGN AND MUSCLE. Because stimulation of the ganglion explants caused widespread muscle contractions, which could not be effectively distinguished by varying the stimulus strength, the complication of movement artifacts from myotubes nearby the one impaled became a problem. To obviate this problem, dissociated sympathetic neurons were cultured with muscle cells (dissociated and chunks) and intracellular recordings from nerve-muscle pairs obtained. The growth of dissociated SCGN among the myotubes contrasted to the random di-


WAKSHULL,

1432

JOHNSON,

rectionality of neuronal process growth in SCGN cultures alone. Instead, the processes tended to orient themselves parallel to the long axis of the myotubes. This was true of all SCGN grown with muscle and is demonstrated in the camera lucida drawing of an HRP-injected neuron derived from a IO-wk-old rat (Fig. 4). The myotubes (shaded areas) labeled 1 and 2 were innervated by the neuron, while myotube 3, which had a neuronal process running over it, was not. Previous reports had demonstrated that embryonic SCGN would form cholinergic connections with skeletal muscle cells in vitro (15). As a control for our culture systern, and to confirm these earlier studies, we cultured perinatal SCGN with muscle cells. Nerve-evoked muscle depolarizations were readily found. Twenty-two neurons and sixteen myotubes were penetrated; a total of six neuromuscular contacts were

d-TC

1O-6 M

\

I

fOrnV

50

msec

AND

BURTON

found, as well as one autaptic potential (see below). The EPPS recorded were sensitive to extracellular Ca*+ levels, and were also blocked by d-TC (Fig. 5), indicating that transmission was nicotinic cholinergic. Spontaneous miniature EPPS were rarely found, and the sample size obtained was insufficient for statistical analysis. We next tested the ability of postnatal SCGN to form functional connections with skeletal muscle. Dissociated SCGN from six different culture series were studied, including tissue from 5, 7-, and IO-wk-old rats cultured with skeletal muscle cells (dissociated and chunks). A total of 42 neurons and 65 myotubes were penetrated, and from this group nine neuromuscular contacts were found. This degree of interaction was comparable to that found for embryonic SCGN and muscle. Three of these EPPS were tested pharmacologically with 2 PM dTC. In all cases, the EPPS could be reversibly inhibited by &TC (Fig. 6A, B). Four neurons forming neuromuscular contacts also established synaptic connections on themselves (autapse). Figure 6A shows an example of this phenomenon. The autaptic potential is not present in Fig. 6A 1, but can be seen in Fig. 6A2 (arrow) and 6A3 (hump in falling phase of nerve AP, upper trace). Synaptic transmission between adult SCGN has been shown to be sensitive to the ganglionic blocking agent, hexamethonium (29). Perfusion with 2 PM d-TC for 11 min had little effect on the amplitude of the autaptic potential, while it reduced the size of the EPP substantially (Fig. 6A2, arrowhead). Figure 6B shows another example of a d-TC-sensitive neuromuscular junction and demonstrates the gradual blockade of the junctional potential by the continued perfusion of antagonist. DISCUSSION

FIG. 5. Cholinergic innervation of skeletal myotube by embryonically derived SCGN in dissociated cell culture. Direct stimulation of the neuron (break in base line, lower trace) evoked a suprathreshold depolarization in muscle cell (upper trace), 1. Perfusion of the culture with d-tubucurarine (d-TC; lo--” M), 2, reduced the amplitude of the response to subthreshold levels. Recovery of the response occurred following return to control medium, 3. Calibration: 1, 10 mV, 50 ms, both traces; 2,3,5 mV upper trace, 10 mV lower trace, 50 ms, both traces; SCGN age in utero, 21 days; age in vitro, 24 days.

Synaptic transmission between postnatal SCGN was found to be chemical and cholinergic. Similarly, junctional transmission between embryonic and postnatal SCGN and skeletal muscle cells was also chemical and cholinergic. Virtually every synaptic potential (neuron-neuron) for which a complete pharmacological series (block and recovery) was obtained was found to be sensitive to the


SYMPATHETIC

NEURONS

IN VITRO.

1433

II

7 d-TC

2x10-6

M

d-TC

d-TC

2 x~O-~ M

2x10-6

M

-J

1°mV 20 50 msec 3

--FIG. 6. Two examples of cholinergic skeletal muscle innervation by postnatal SCGN. A : direct stimulation of neuron (break in base line, upper trace) in control solution, 1, evoked an action potential in the innervated myotube (lower trace); 2, perfusion with d-tubocurarine (d-TC; 2 x lo-” M) blocked the muscle response, but not the autaptic potential (arrowhead), which is more easily seen with depolarizing stimulation of the neuron. The suprathreshold muscle response recovered on return to control medium, 3. The autaptic potential is visible as the hump in the falling phase of the neuronal action potential. B: direct stimulation of neuron (upper trace, AP not shown here) elicits a suprathreshold response in myotube (lower trace) in control solution, 1. After 5 min in d-TC (2 x IO+’ M) the response is diminished in amplitude, 2a. At 6 min, 2b, the response is abolished, and after 8 min in control solution, the response recovers, 3. Unfortunately, the neuronal APs did not reproduce well in these traces. 10 mV, upper trace (neuron), 20 mV lower trace (muscle), 50 ms. SCGN age in vivo, 70 days; Calibration: A, B: age in vitro, 19 days.

cholinergic antagonist C-6. More intriguing, though, are the several examples of propranolol-sensitive synaptic potentials. Propranolol, in concentrations exceeding 4 x IO+ M, has been shown to have local anesthetic activity (32). Concentrations between 5 x 10V6 and 5 x lo-” M were used in the present study, but because the culture was perfused continuously it is difficult to determine the drug concentrations at any given time. The turnover time of the fluid in the culture dish has been estimated at 20 min (9); most of the blocking action of propranolol was seen in 5-7 min. Thus, the effective concentration of the drug was much less than the maximum concentration

obtainable. In addition, the many synaptic potentials that were not blocked by propranolol (including some that were perfused for 15 min) argue against an anesthetic action of propranolol. Alternatively, propran0101 may antagonize P-receptor activation. No evidence that NE membrane receptors on or near the cell body have been found in culture (18, 3 l), thus suggesting a presynaptic location for possible P-receptors. The presence of presynaptic P-receptors on sympathetic neurons has been shown in vivo (32) and in vitro (32, 33). Activation of these receptors causes an increased release of NE, resulting in a positive feedback loop. By this model, propranolol might have de-


1434

WAKSHULL.

JOHNSON.

creased transmitter release in sympathetic terminals. Thus, it may be that both ACh and NE are released by these cells on stimulation. The NE might then act on the presynaptic p-receptors, increasing transmitter release, while ACh acts postsynaptically. Further tests may yet reveal such distinctions in pre- and postsynaptic effects of the NE antagonists. In addition to forming cholinergic junctions on the skeletal muscle cells, some of the sympathetic neurons making neuromuscular synapses established autaptic connections. Although it is possible that there is an interposed neuron, which produces the observed synaptic potential, the following arguments make this unlikely. First, the potentials were of short latency, appearing on the repolarization phase of the AP and immediately following or even obscuring the afterhyperpolarization. Second, the autaptic potential had few failures, implying that a second neuron would have to follow the first virtually 1: I with an AP. Third, the low density of neurons, particularly in the postnatal SCGN and muscle cultures (< 100 neurons per culture dish), made interneuronal distances large and the probability of interaction low. However, even if a second neuron was interposed and produced the synaptic potential in the original driver, one still has a situation where a sympathetic neuron is synaptically connected to both a muscle cell and a neuron. A similar finding has been reported by Nurse and O’Lague (15). The phenomenon of a neuron innervating a nonneuronal cell and another neuron is found normally in mammalian cymotor neurons (25) and mudpuppy cardiac ganglion (23), and following denervation in the parasympathetic ganglion innervating the heart (26) and guinea pig superior cervical ganglia (22). Attempts to innervate skeletal muscle with fibers from the SCG in vivo have been made with a notable lack of success (11, 12). One possible explanation for this negative result by Langley and Anderson (I 1) may have been their apparent failure to cut the original nerve to the target muscle. Since their experiments were done, it has been shown that a foreign nerve will not innervate muscle with intact native nerve fibers (6) unless care is taken to implant the foreign nerve over the original end plate (1). Also,

AND

BURTON

subthreshold potentials would not have been detected by gross observation of contraction (11) or extracellular recording methods (12). The cholinergic innervation of skeletal muscle cells by sympathetic neurons in vitro does not appear altogether anomalous since previous studies had established the use of ACh by these neurons (4, 8- 10, 17, 18, 27-31). It is possible, however, that special conditions in the tissue culture environment, apart from promoting cholinergic transmitter function, allows the SCGN to form these neuromuscular contacts. It would be interesting to know whether the SCGN can induce postjunctional specializations in the myotubes. Three explanations for the observation that “adult” dissociated SCGN form cholinergic synapses on each other and skeletal muscle cells are available. The small subpopulation of cholinergic sympathetic neurons thought to exist in the rat SCG (34) could be surviving selectively in our culture system. Alternatively, it is possible that both ACh and NE are produced simultaneously by sympathetic neurons even in the fully differentiated state, as first proposed by Burn and Rand (3). Finally, these neurons could retain a certain degree of flexibility in the expression of their neurotransmitter phenotype, as the embryonic neurons seem to have done (8). and acquire the ability to synthesize and release ACh under the inductive influence of as yet unknown factor(s) in our culture environment. A recent review by Bunge et al. (2) proposed a two-stage developmental model for autonomic neurons. In the first stage, neural crest cells become committed to differentiate into autonomic neurons. This occurs early in embryonic development. At this point, the presumptive neurons may be predisposed to cholinergic or adrenergic transmission. This predisposition is probably influenced by the tissues through which the cells are migrating (5, 14). However, the final commitment to one transmitter system over the other (stage two) is suggested to occur after the ganglion cells have reached their final destination, stopped dividing, and sampled their peripheral target field. This model was consistent with the findings of Ross et al. (24) who reported that when SCG explants are grown from animals older than 2 days postnatal, their ability to accrue


SYMPATHETIC

NEURONS

choline acetyltransferase (CAT) activity declines to extremely low levels when taken from rats older than 3-4 wk. It was proposed by Ross et al. (24) that a critical period existed during postnatal development, after which the ability of the SCGN to synthesize different neurotransmitters decreases. With the possible exception of the two culture series in which no synaptic interactions were found (31), the results reported here suggest that adrenergic sympathetic neurons retain the ability to respond to environmental factors and may still make detectable levels of acetylcholine. It is conceivable that those neurons that survived the profound disruption produced by the dissociation procedure had in fact been able to regress to a more dedifferentiated (and state. This extherefore, more “flexible”) planation, of course, assumes that a shift in transmitter production from adrenergic to cholinergic has occurred. The possibility has not been ruled out that the normally cholinergic sympathetic neurons thought to be present in this ganglion are selectively surviving and forming the observed synaptic potentials. The observed cholinergic connection between explanted ganglia and muscle could also be explained by the presence of extensively ramifying normal cholinergic SCGN. Assuming this subpopulation of cholinergic sympathetic neurons is in fact present within the rat SCG, it remains unclear what the mechanism of their selective survival in culture might be. In this context, one may rule out the ability to form functional synaptic connections since two culture series of otherwise healthy postnatal SCGN were found devoid of synaptic interactions (31). Also, in the absence of supporting cells or conditioned medium, O’Lague et al. (17, 18) were unable to find synaptic interactions in otherwise healthy embryonic SCGN cultures. In addition, recent evi-

IN VITRO.

1435

II

dence (unpublished observations) strongly suggest that the neurons which survive in dissociated postnatal cultures are adrenergic at early times in vitro. A developmental electron microscopic study, in which the percentage of dense-core synaptic vesicles is studied as a function of postnatal and in vitro age, is being undertaken to distinguish between a selective survival of cholinergic sympathetic neurons and a shift in transmitter production by these postnatal neurons. A similar study done on embryonic cultures indicated that only one neuron type survived and that nearly all of this population of cells acquired cholinergic properties (8). These results do not refute the developmental scenario outlined by Bunge et al. (2), as they apply to the normal in vivo situation. They do suggest, however, that a commitment to a particular transmitter may not be an immutable step in the maturation of autonomic neurons. Instead, these cells may retain the ability to choose either transmitter system indefinitely. This does not mean that both ACh and NE are used simultaneously, although this can occur (7). In this view, the cholinergic shift by SCGN does not seem paradoxical, but rather appears to be an expression of normal capabilities. ACKNOWLEDGMENTS

We appreciate the technical assistance of Mary Miller and Marc Davis. We are especially grateful to Drs. Richard Bunge and Dale Purves for critical reading of the manuscript. This study was supported by Public Health Service Grant 09809 and National Institutes of Health Grant 11888. The material presented in this paper was submitted in partial fulfillment of the requirements for the Ph.D. Degree of E. Wakshull. Received 20 December 30 April 1979.

1978; accepted

in final form

REFERENCES 1. BIXBY, J. L. AND VAN ESSEN, D. C. Suppression of original nerve inputs to a mammalian skeletal muscle by a foreign motor nerve. Nurrrosci. Ahstr. 4: 367, 1978. 2. BUNGE, R. P., JOHNSON, M., AND Ross, C. D. Nature and nurture in development of the autonomic neuron. Scicvzw 199: 1409- 1416, 1978. 3. BURN, J. H. AND RAND, M. J. A new interpretation of the adrenergic nerve fibers. Ad. P1zmww cd. 1: I-30, 1962.

4. BURTON, H., Ko, C.-P., AND BUNGE, R. Cholinergic synapses between sympathetic neurons in tissue culture. Ncwosci. Ahstr. 1: 816, 1975. 5, COHEN, A. M. Factors directing the expression of sympathetic nerve traits in cells of neural crest origin. J. ,!+p. Zool. 179: 167- 182, 1972. 6. ELSBERG, C. A. Experiments on motor nerve regeneration and the direct neurotization of paralyzed muscles by their own and by foreign nerves. Scimw 45: 3 18-320, 1917.


1436

WAKSHULL,

JOHNSON,

7. FURSHPAN, E. J. MACLEISH, P. R., O’LAGUE, P. H., AND POTTER, D. D. Chemical transmission between rat sympathetic neurons and cardiac myocytes developing in microcultures: evidence for cholinergic, adrenergic and dual-function neurons. Proc. Natl. Acad. Sci. USA 73: 4225-4229, 1976. 8. JOHNSON, M., Ross, D., MEYERS, M., REES, R., BUNGE, R. P., WAKSHULL, E., AND BURTON, H. Synaptic vesicle cytochemistry changes when cultured sympathetic neurons develop cholinergic interactions. Natuw London 262: 309-3 10, 1976. 9. Ko, C.-P. Elec~trophysiolo~icul Studies oj‘Synaptic Transmission in Nerve Tissue CUtwe oj’ Rut Spinal Cord und Dissociated Superior Cervicwl Gcrnglion Neurons (Ph.D. Thesis). St. Louis, MO: Washington University School of Medicine, 1975. 10. Ko, C.-P., BURTON, H., JOHNSON, M. I., AND BUNGE, R. P. Synaptic transmission between rat superior cervical ganglion neurons in dissociated cell cultures. Bruin Res. 117: 461-485, 1976. 11. LANGLEY, J. N. AND ANDERSON, H. K. The union of different kinds of nerve fibers. J. Physiol. London 31: 365-391, 1904. J., ARANDA, L. C., AND Luco, J. V. 12. MENDEZ, Antifibrillary effect of adrenergic fibers on denervated striated muscles. 1. Neurophysiol. 33: 882890, 1970. 13. NATHANSON, J. A. Cyclic nucleotides and nervous system function. Physiol. Rev. 57: 157-256, 1977. 14. NORR, S. C. In vitro analysis of sympathetic neuron differentiation from chick neural crest cells. De\,. Biol. 34: 16-38, 1973. C. A. AND O’LAGUE, P. H. Formation of 15. NURSE, cholinergic synapses between dissociated sympathetic neurons and skeletal myotubes of the rat in cell culture. Proc. Nutl. Acad. Sci. USA 72: 19551959, 1975. 16. OBATA, K. Transmitter sensitivities of some nerve and muscle cells in culture. Brain Rcs. 73: 7 l-88, 1974. 17. O’LAGUE, P. H., MAC LEISH, P. R., NURSE, C. A., CLAUDE, P., FURSHPAN, E. J., ANDPOTTER, D. D. Physiological and morphological studies on developing sympathetic neurons in dissociated cell culture. Cold Spring Hcrrhor Symp. Qluwt. Biol. 40: 399-407, 1976. 18. O’LAGUE, P. H., OBATA, K., CLAUDE, P., FURSHPAN, E. J., AND POTTER, D. D. Evidence for cholinergic synapses between dissociated rat sympathetic neurons in cell culture. Proc. Nut!. Acud. Sci. USA 71: 3602-3606, 1974. 19. PATTERSON, P. H. AND CHUN, L. L. Y. The influence of non-neuronal cells on catecholamine and acetylcholine synthesis and accumulation in cultures of dissociated sympathetic neurons. Proc. Ncrtl. Acwd. Sci. USA 71: 3607-3610, 1974. 20. PATTERSON, P. H. AND CHUN, L. L. Y. The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. 1. Effects of conditioned medium. De\*. Biol. 56: 263-280, 1977.

AND

BURTON

21. PATTERSON, P. H. AND CHUN, L. L. Y. The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. 11. Developmental aspects. Dev. Biol. 60: 473-48 1, 1977. 22. PURVES, D. Functional and structural changes in mammalian sympathetic neurones following colchicine application to postganglionic nerves. J. Physiol. London 259: 159- 175, 1976. 23 ROPER, S. An electrophysiological study of chemical and electrical synapses on neurons in the parasympathetic cardiac ganglion of the mudpuppy, Nectrrrxs maculosus : evidence for intrinsic ganglionic innervation. .J. Physiol. London 254: 427-454, 1976. M., AND BUNGE, R. De24 Ross, D., JOHNSON, velopment of cholinergic characteristics in adrenergic neurones is age dependent. Nc~trrw London 267: 536-539, 1977. 25 RYALL, R. W. Renshaw cell mediated inhibition of Renshaw cells: patterns of excitation and inhibition from impulses in motor axon collaterals. J. New-ophysiol. 33: 257-270, 1976. 26. SARGENT, P. B. AND DENNIS, M. J. Formation of synapses between parasympathetic neurones deprived of preganglionic innervation. Natwe London 268: 456-458, 1977. 27. WAKSHULL, E. Stlddies on the In Vitro Development oj’Rcrt Sympathetic Nwrons (Ph.D. Thesis). St. Louis, MO: Washington University, 1978. 28. WAKSHULL, E., JOHNSON, M. I.,ANDBURTON,H. Electrophysiological and morphological studies of dissociated sympathetic neuron cultures prepared from young rats. J. Ccl1 Biol. 75: 115a, 1977. 29. WAKSHULL, E., JOHNSON, M., AND BURTON, H. Physiological studies of postnatal rat sympathetic neurons and skeletal muscle cells in vitro. Nerrrosci. Ahstr. 4: 585, 1978. 30 WAKSHULL, E., JOHNSON, M.. AND BURTON, H. Persistence of an amine uptake system in cultured rat sympathetic neurons which use acetylcholine as their transmitter. J. Cc~ll Biol. 79: 12 1- 13 1, 1978. E.. JOHNSON, M. I.. AND BURTON. H. 31 WAKSHUU. Postnatal rat sympathetic neurons in culture. I. A comparison with embryonic neurons. J. Ncwophysiol. 42: 1410- 1425, 1979. M. The presynaptic effect of p32. WEINS-I-OCK, adrenoceptor antagonists on noradrenergic neurons. L(fil Sci. 19: 1453- 1466, 1976. M., THOA, N. B., AND KOPIN, I. J. 33. WEINSTOCK, P-adrenoceptors modulate noradrenaline release from axonal sprouts in cultured rat superior cervical ganglia. Eur. .J. Phcwmcrwl. 47: 297-302, 1978. A., LEVER, J. D., AND KEMP, K. W. 34. YAMAUCHI, Catecholamine loading and depletion in the superior cervical ganglion. A form01 fluorescent and enzyme histochemical study with numerical assessments. .J. Antrt. 114: 271-282, 1973.


Postnatal rat sympathetic neurons in culture. I. A comparison with embryonic neurons.

Identified postnatal mesolimbic dopamine neurons in culture: morphology and electrophysiology.

Maternal Dexamethasone Exposure Alters Synaptic Inputs to Gonadotropin-Releasing Hormone Neurons in the Early Postnatal Rat.

Postnatal development of cholinergic neurons in the rat: I. Forebrain.

Neural nicotinic acetylcholine responses in sensory neurons from postnatal rat.

Electrophysiological properties of early postnatal rat septal neurons in short-term culture.

Primary culture of postnatal rat hypothalamic neurons in astrocyte-conditioned medium.

Presynaptic GABAB receptor activation attenuates synaptic transmission to rat sympathetic preganglionic neurons in vitro.

Synaptic transmission between rat spinal cord explants and dissociated superior cervical ganglion neurons in tissue culture.

Synaptic mechanisms in sympathetic preganglionic neurons.

Blockade of ionotropic quisqualate receptor desensitization by wheat germ agglutinin in cultured postnatal rat hippocampal neurons.

Adult-born neurons modify excitatory synaptic transmission to existing neurons.

Developmental changes in transmitter sensitivity and synaptic transmission in embryonic chicken sympathetic neurons innervated in vitro.

Postnatal subventricular zone progenitors switch their fate to generate neurons with distinct synaptic input patterns.

Vasoactive intestinal polypeptide presynaptically enhances the synaptic transmission in cultured sympathetic neurons.

Postnatal maturation of neurons in the rabbit superior colliculus.

Modulation of excitatory synaptic transmission in rat hippocampal CA3 neurons by triphenyltin, an environmental pollutant.

Retrograde labeling of corticospinal motor neurons from early postnatal rodents.

Postnatal development of the hypothalamic ventromedial nucleus: neurons and synapses.

Ultrastructural studies on postnatal differentiation of neurons in the substantia gelatinosa of rat cervical spinal cord.

Postnatal development of electrogenic sodium pump activity in rat hippocampal pyramidal neurons.

Postnatal development of preproenkephalin mRNA containing neurons in the rat lower brainstem.

Postnatal exposure to tones alters the tuning characteristics of inferior collicular neurons in the rat.

Characterization of quisqualate receptor desensitization in cultured postnatal rat hippocampal neurons.