News Physiol Sci 19: 85-91, 2004;
doi:10.1152/nips.01464.2003
1548-9213/04 $5.00
News in Physiological Sciences, Vol. 19, No. 3, 85-91,
June 2004
© 2004 Int. Union Physiol. Sci./Am. Physiol. Soc.
Activity-Dependent Synaptic Competition at Mammalian Neuromuscular Junctions
Mario Buffelli,
Giuseppe Busetto,
Carlo Bidoia,
Morgana Favero and
Alberto Cangiano
Dipartimento di Scienze Neurologiche e della Visione, Sezione di Fisiologia, Universita’ di Verona, 37134 Verona, Italy
 |
Abstract
|
|---|
Synapse elimination is a widespread developmental process in the peripheral and central nervous system that brings about refinement of neural connections through epigenetic mechanisms. Here we describe recent advances concerning the role of the pattern of motoneuronal firing, synchronous or asynchronous, in neuromuscular synapse elimination.
|
Introduction
|
|---|
Action potential activity in neural networks can profoundly influence the structure and function of neuronal connections, a fact that is hardly surprising given the effects of experience, and thus of activity, on our brains. Yet knowledge of the precise mechanisms involved is still in its infancy, and various models of activity-dependent plasticity of synapses are being actively investigated in many laboratories. It is presently recognized that the long-lasting changes affecting the nervous system during development and learning share a number of common features. One of these is the Hebbian paradigm (14), according to which repeated coincident spike activity of pre- and postsynaptic neurons leads to strengthening of the intervening synapses, whereas noncoincident activity tends to weaken them (Stent’s extension of Hebb’s postulate; see Ref. 27). A Hebbian rule underlies, for example, long-term potentiation, by which short bursts of high-frequency action potential activity across hippocampal synapses rapidly lead to a long-lasting increase of their efficacy (4), a change commonly held as a cellular correlate of learning. Hebbian mechanisms relating to visual experience also strongly affect the early developmental refinement of connections in the visual pathways, although their permissive rather than instructive role is under close scrutiny (17). Another striking example of synaptic plasticity, of a short-term nature but closer to the focus of this presentation, are the findings of Lo and Poo (21) in frog neuromuscular junctions in vitro: brief tetanic stimulation of one of two spinal neurons coinnervating a single embryonic muscle cell leads to rapid suppression of synaptic transmission from the other neuron, unless it is also synchronously activated. Hebbian potentiation of active synapses and depression of nonactive ones are thus both expressed in this in vitro system. In this review, we will present evidence that similar mechanisms are operative during a well-known developmental change occurring at the neuromuscular junction, namely synapse elimination. A description of the general features of this process will be introduced first.
|
Synapse elimination in skeletal muscle
|
|---|
In neonatal animals each muscle fiber is innervated by several axons that are collaterals of different motoneurons (Fig. 1A
), whereas in the adult only one axonal branch makes contact with them. In the 1970s, this observation by Redfern (see Ref. 16 for review) introduced a new developmental phenomenon, synapse elimination, later confirmed in many other locations of the peripheral and central nervous system, such as autonomic ganglia, cerebellum, thalamus, and cerebral cortex. The occurrence of synapse elimination is comprised in rodents between the time of birth and postnatal day 15 (P15), with a few days’ variation in specific muscles. During this developmental period, the number of motoneurons in the spinal cord stays constant, programmed cell death being an embryonic process that is already over by the time of birth. Synapse elimination is therefore a peripheral process that only involves motor nerve endings and their axonal branch: each motor unit (the muscle fibers and their innervating motoneuron) undergoes a change from a large size, overlapping with other units, to a much smaller, nonoverlapping one, which is retained as a permanent feature throughout adult life (Fig. 1B). The neonatal/adult motor unit size ratio varies in different muscles and is estimated to reach a value of ~5 in soleus muscle; accordingly, in this muscle, the maximum estimated number of neonatal motor nerve endings on each fiber is close to 5.
View larger version (54K):
[in this window]
[in a new window]
|
FIGURE 1. Neonatal synapse elimination represented as the loss of polyneuronal innervation (A) and the reduction in motor unit size (B).
|
|
That neonatal muscle fibers are polyneuronally innervated can be unequivocally assessed by recording intracellularly from fibers of muscles perfused in vitro and by showing that graded single-shock electrical stimulation of the nerve evokes end-plate potentials of amplitudes increasing in discrete steps. This shows that motor axons with different thresholds exist in the nerve that innervate the same muscle fiber (Fig. 2A). In adult muscles, instead, only one step of the end-plate potential can be evoked (mononeuronal innervation). Polyneuronal innervation cannot be an artifact due to electrical coupling between muscle fibers through gap junctions, because they are no longer detectable after birth (5, 16). In addition, morphological studies confirm these conclusions by showing perinatally multiple axons reaching each muscle fiber (Fig. 2B). Morphology also uncovers a peculiar feature of polyinnervation: multiple axons do not reach separate ACh receptor (AChR) aggregates but converge instead on the same aggregate, the areas occupied by the different terminals being initially intermingled with one another in a complex fashion. During the following days the terminals converging on each AChR aggregate (albeit remaining in close contact at their boundaries), become segregated, and finally gradually withdraw, with the exception of one that takes over the relinquished AChR territory. Only occasionally does the takeover not occur, and in this case the vacated AChR region disappears (30). Elimination differs from wallerian degeneration, being a process of resorption of the nerve ending and atrophy of the axonal branch. A retraction bulb (schema in Fig. 1A
and image in Fig. 2B) is the frequent morphological indicator of a terminal undergoing elimination.
View larger version (52K):
[in this window]
[in a new window]
|
FIGURE 2. Polyneuronal innervation in soleus muscles of newborn rats (postnatal day 8) shown by electrophysiology (A) and confocal fluorescence microscopy (B). Intracellular recordings of end-plate potentials evoked by graded nerve stimulation under partial curarization are shown, as is fluorescence microscopy of neuromuscular junctions visualized with rhodamine- bungarotoxin [ACh receptors (AChRs), red], anti-neurofilament (axons, green), and anti-synaptophysin (terminals, green) fluorescent antibodies. Asterisks indicate polyneuronally innervated AChR aggregates. Arrow indicates retraction bulb with atrophic axon.
|
|
Synapse elimination is not a random process by which each motoneuron loses a percentage of its terminals, because this would lead to some muscle fibers being devoid of all motor nerve endings, at least before collateral reinnervation can take place. Never, however, during the time course of synapse elimination can one observe denervated muscle fibers, not even transiently (16). Only a competitive process, through which each terminal tries to become the only input of the muscle fiber, can lead from the initial polyneuronal pattern of innervation to the mononeuronal adult one. The competitive nature is confirmed by partial denervation performed at birth. The idea is to see the effect of early removal of part of the competitors on the evolution of motor unit size. The result is that at approximately P15, when the size has normally already shrunk to the small adult value, after partial denervation it still retains the large neonatal dimension (5). This is not accompanied by morphological signs of terminal collateral sprouting (2) and is evidently due to reduced competition.
Clearly synapse elimination, which is a widespread developmental step in the nervous system, serves a general purpose of refinement, through epigenetic mechanisms, of a highly ordered and precise pattern of connections already established under genetic instruction. Synapse elimination may then be flexibly adapted to the specific developmental needs of a particular system in the peripheral and central nervous system. Basic questions concerning the activation of this process include what kinds of signals are involved, in particular in skeletal muscle, why multiple endings are initially formed, and why they remain stable for a substantial amount of time on each target myofiber, when after birth a competition process is switched on that leads to mononeuronal innervation.
In general, signals involved in neuromuscular junction development range from action potentials to chemical factors (26), and in the case of synapse elimination previous evidence has shown a strong influence of the first factor. For example, synapse elimination is inhibited by block of neuromuscular activity (see Ref. 16 for review). Confirmation comes from the demonstration that the opposite effect, that is an accelerated elimination, is induced by increase of activity obtained through electrical stimulation of the nerve (28).
One essential point regarding this issue, however, is that a clear-cut distinction should be made between 1) indirect effects of activity on synapse elimination through the well-known changes that activity induces postsynaptically, that is in muscle fiber properties (to which in all likelihood the results just described are related), and 2) a presynaptic, much more specific role of activity related to amount and especially timing of action potential firing in different inputs competing for innervation of the same target (16). An extreme form of this differential activity is that of normally active vs. completely inactive inputs; its test on synapse competition and elimination has given conflicting results (that is advantage for the former or for the latter; see Ref. 16 for review; see, however, Conclusions and Ref. 6). In any case, this paradigm, despite its considerable interest and relevance, is certainly not one that is found in normal neuromuscular physiology either in development or in the adult condition. We therefore started our experiments on the effects of differential activity on synapse elimination by testing more physiological paradigms, that is asynchronous vs. synchronous activity. Two groups of investigations will be described: 1) a study of the effects of replacing the physiological asynchronous activity of motor axons with a synchronous one, through electrical stimulation, during the entire course of synapse elimination and 2) a study of the way motoneurons fire their action potentials early in development, to see if a synchronous firing exists before the physiological asynchronous pattern sets in.
|
Effects of evoked synchronous activity on synapse elimination
|
|---|
In adult life, spike activity of motoneurons of the same pool, which innervate the same muscle, is normally asynchronous in nature (25). This means that the time when each spike is fired by one motoneuron is not correlated with the time when the spikes of the other motoneurons are discharged. This ensures that when different motor units, each firing repetitively and inducing unfused, fluctuating tetanic contractions, are coactivated, they give rise to a smooth, nonfluctuating overall contraction. It is reasonable to assume that it is this way of firing of the competing inputs that brings about synapse elimination. Several things make this hypothesis attractive. First, the Stent’s extension of Hebb’s postulate applied to this situation implies that each time a motor terminal fires a spike and transmits it to the muscle fiber, it strengthens itself while simultaneously weakening the other terminals not firing at the same time. It is certainly true that a moment later this situation is reversed, and so on and so forth, but the essential point is that asynchronous firing creates a situation of instability in which each ending competes with the others to become the only input of the muscle fiber. Further support for this idea is that similar asynchronous activity of converging inputs during the critical period of development of ocular inputs to visual cortical neurons leads to impaired development of binocular, that is polyneuronal, inputs and thus to visual defects. This can be induced in newborn mammals by creating an artificial squint (15). Another type of support comes from experiments performed not during the process of synapse elimination but in adult monoinnervated neuromuscular junctions; here blocking only part of the receptor aggregate with -bungarotoxin brings about elimination of the overlying portion of the motor terminal, as if activation of one part of the neuromuscular junction induces the elimination of the adjacent inactive one (1). This partial motor terminal elimination occurs without takeover (see above) of the blocked receptor aggregate that actually disappears and could thus have some peculiarities of its own with respect to the mechanisms at work during the physiological elimination process. This paradigm is still, however, of great interest for the issue of differential activity. In fact, blocking the entire receptor aggregate does not destabilize the junction (1). Finally, the results of Lo and Poo (21) on neuromuscular junctions cited earlier, although obtained in a cellular in vitro model of synapse competition rather than in the in vivo process, are clear-cut proof of the operativity of these Hebbian mechanisms in real neuromuscular synapses.
We now come to our experiment of synchronous activation of competing inputs. Because mammalian muscles only have one motor nerve, it is not possible to electrically stimulate independently, that is asynchronously, two groups of axons innervating the same muscle fibers during the period of synapse elimination. However, we reasoned that if we electrically stimulated the unique motor nerve to a muscle, we would in fact replace the spontaneous action potential firing of its axons, normally asynchronous, with the synchronous firing evoked by the electrodes. This is true, however, only if the spontaneous activity is concomitantly blocked, central to the stimulation site. We could then study what influence, if any, this type of firing has on synapse elimination. Should the effect be a substantial delay or suppression altogether of the elimination process, we would have a very strong argument to conclude that, physiologically, elimination is brought about by asynchronous activity of the axons that compete to remain the only input of each muscle fiber.
Transient polyneuronal innervation not only characterizes embryonic development, but it also occurs in adult muscle during reinnervation by its own nerve after crush or during cross-innervation by a foreign nerve in an ectopic region after section of the original nerve (5). These adult models of synapse competition and elimination are very useful, because they allow manipulations, like in this case chronic electrical stimulation and central conduction block of nerves, that would be impossible in newborn rats due to their small size. The model we used for our experiment involved the following steps, all depicted in Fig. 3, A and B: 1) the central stump of a foreign nerve, superficial fibular, is transplanted on the proximal, synapse-free surface of the rat soleus muscle; 2) 2–3 wk are allowed to pass, during which foreign axons grow without making synapses; 3) the original soleus nerve is now cut, and simultaneously chronic electrical stimulation of the nerve through implanted electrodes and a centrally located block of conduction with tetrodotoxin are made operative; 4) after a further 3–4 days, many new neuromuscular junctions begin to form in the fibular nerve region, while their axons are synchronously firing action potentials; and 5) the contralateral side serves as a control, in which foreign synapses also form but their motor axons discharge asynchronously because no electrical stimulation or conduction block are applied. At varying times after cutting the soleus nerve, we isolated in vitro experimental and control contralateral muscles to assess the percentage of polyneuronally innervated fibers with respect to the total number of cross-innervated fibers in each muscle. We could thus compare synapse elimination while converging inputs on polyinnervated muscle fibers fire synchronously or asynchronously for several weeks. The results were striking and are summarized in Fig. 3C. Compared with control synapses (asynchronous activity), the experimental ones (synchronous activity) had a much higher percentage of polyneuronal innervation at all time points, on average about three times as large. Thus the initial working hypothesis was completely confirmed, and the role of asynchronous activity of competing inputs in causing synapse elimination was strongly supported (9).
View larger version (28K):
[in this window]
[in a new window]
|
FIGURE 3. Inhibition by synchronous activation of synapse competition and elimination in cross-innervated adult soleus muscles by a foreign nerve. See text for details.
|
|
|
Synchronous vs. asynchronous firing of spinal motoneurons during development
|
|---|
In the experiment described in the preceding section, we used an artificially imposed synchronous activity to assess the role of asynchronous firing of motoneurons of the same muscle pool in synapse elimination. However, could synchronous firing actually be a physiological feature of motoneuronal activity during early developmental stages? If this theory was correct, a switch to the asynchronous mode of firing characteristic of the adult motoneurons should take place somewhere during postnatal life. Would this hypothesized switch take place early enough to account for the activation of the elimination process? We tried to answer these questions by recording the physiological action potential discharge of motoneurons in awake newborn rats accomplishing simple movements, by means of electromyographic recordings of single motor unit activity. In records from a given muscle in which at least two such units are clearly identifiable by their different waveform and amplitude, we could make a cross-correlation study of the relative time of occurrence of their signals. Records were taken from tibialis anterior (TA) and soleus muscles during the period between the last day of embryonic life (E21) and postnatal day 30 (P30). Our findings, recently published (7), can be summarized as follows: 1) at E21 and during the first few days of postnatal life (up to P2 and P5 in TA and soleus, respectively) the time of occurrence of the signals of any two motor units within the same muscle was highly correlated (example in Fig. 4, A and C); 2) the window for this correlated firing was narrow (~25 ms) as judged from the averaged cross-correlogram of the entire population (30 couples), the most frequent interval being 5 ms (Fig. 4E, gray bars); 3) after this time, the correlation quickly became much looser and completely disappeared after the first decade of postnatal life (Fig. 4E, open bars, P13–P30, 47 couples); 4) the switch from correlated to uncorrelated activity occurred a few days earlier in the TA muscle than in soleus; also, the loss of polyneuronal innervation occurred earlier in the former muscle.
View larger version (15K):
[in this window]
[in a new window]
|
FIGURE 4. Electromyographic recordings of single motor unit firing (2 units in each record) in a postnatal day 3 (A) and a postnatal day 29 (B) soleus muscle. C and D: respective cross-correlograms of the time of occurrence of spikes. E: 2 averaged cross-correlograms of all unit couples of the populations at early (perinatal) and late (adult) developmental times. Each unit is marked with * or x. Calibrations (A and B): voltage, 100 µV; time, 25 ms.
|
|
|
Conclusions
|
|---|
On the basis of the above experimental evidence, we can draw the following conclusions:
- The striking effects of imposed synchronous activity, consisting of profound inhibition of synapse elimination, indicate that physiologically, asynchronous firing is important for activation of the elimination process. Detailed evidence has been presented (9) that our evoked firing is highly effective in activating the muscle and that its amount (number of spikes evoked/day of chronic stimulation) is entirely comparable with normal activity of the motoneuronal pool utilized for cross-innervation of the soleus muscle. What differs from the contralateral control muscle, undergoing cross-innervation with normal spontaneous activity, is the way that the impulse traffic is organized in time in the innervating axons, synchronous instead of asynchronous. The end result is that the neuromuscular preparation on the stimulated side is as "adult" as the control side, from all standpoints (growth of muscle fiber size, adult profile of membrane receptors and ion channels, shape of the synaptic potential) except for the presence of a prominent polyneuronal innervation on this side only. Support for the importance of asynchronous activity in elimination comes from very recent experiments in which the comparison between normally active and completely inactive terminals (see above) has been revisited by using a reliable model in mice. This is based on transgenes in which some of the motoneurons are unable to activate the muscle fibers because they do not release ACh due to depletion of choline acetyltransferase. Under confocal fluorescence microscopy, normal terminals are easily recognized by appropriate staining. In all myofibers where these terminals compete with ACh-deficient ones, they are the winners (6). This finding finally settles this question and establishes a Hebbian rule for this paradigm.
- Synchronous activity is present physiologically very early in development, to be replaced soon after birth by the well-known asynchronous firing characteristic of adult motoneurons. The switch occurs early enough to account for the role of asynchronous activity in synapse elimination. Furthermore, the fact that 1) the switch occurs earlier in TA than in soleus muscle and that 2) synapse elimination also starts sooner in TA strengthens the conclusion that asynchrony promotes elimination. That activity of neonatal motoneurons is well correlated has also been found in another study (23); the window for correlation is, however, much longer than our ~25 ms by a factor of 10–100, being ~2 s. This loose correlation has therefore a different meaning from that of our tight correlation, simply that trains of spikes of different motoneurons occur in phase, as is true also in adults. On the other hand, a very narrow window similar to ours has been found in another recent study at P1 in the in vitro spinal cord of mice (29). In the same study, the point has been established that gap junctions only partially contribute to this synchrony. In summary, the following overall picture emerges. In the embryo, multiple, synchronously active synaptic contacts of different motoneurons continue to form on each myofiber, due to low level of competition. Soon after birth, with a slight variation in time of onset for different muscles, asynchronous firing in each motor pool sets in, activating competition and causing elimination of all but one terminal.
There is evidence that, besides activity and its pattern, other factors probably participate in elimination, although their identity is unknown. For example, it has been shown that the large size induced by partial denervation in the remaining motor units is not permanent but is followed by a delayed shrinkage, despite the fact that many muscle fibers become permanently denervated (5, 16). Similarly, our data shown in Fig. 3C indicate that the high level of polyneuronal innervation induced by synchronous activity undergoes in the long run a slow decline, despite continuing stimulation. Evidently, motoneurons cannot indefinitely maintain the maximum size of their peripheral territory and finally shrink to a smaller size. Also worth mentioning are the experiments of Costanzo et al. (12) indicating that even in the complete absence of activity, competition can occur in some preparations. It should not be surprising that a complex phenomenon such as synapse elimination depends on multiple factors.
Aside from the role of the pattern of asynchronous activity, the intimate mechanism that underlies elimination is unknown. It is very likely that it is mediated through the postsynaptic side of the neuromuscular junction. Several facts more or less directly speak in favor of this contention. The strongest evidence would that obtained by Lo and Poo in the above-cited Hebbian model of a single frog myocyte innervated by two neurons were it not for the fact that it is a model of short-term functional synaptic suppression that is in part different from synapse elimination; in any case, the evidence here is that preventing the rise of ionized calcium in the myocyte blocks the suppression elicited by activation of one neuron only (see above). Another line of evidence is that competing terminals not able to release ACh are eliminated by normal ones (see above). In this model, activation of the postsynaptic AChRs appears necessary. Further evidence comes from experiments in which polyneuronal innervation is obtained in mammalian muscles by using cross-innervation paradigms in newborn and adult muscles. In these models, multiple inputs are obtained not only on a single site but also at some distance from each other. Elimination also occurs in this case, at distances that can reach ~1 mm (5, 16). This finding is also interesting for another reason: because beyond ~1 mm elimination does not occur, the postsynaptic signal mediating elimination cannot be the action potential but rather must be a diffusible signal inside the myofiber.
Further details on hypothetical scenarios about the intimate mechanism of synapse elimination, its possible molecular mediators, and the way activity might set them in motion are beyond the scope of the present review (for more details see Refs. 8, 18, 22, and the discussion in Ref. 9). We mention here that a role has been proposed for proteases (abundantly present at the neuromuscular junction together with protease inhibitors; Ref. 13) as effectors of the elimination process. These include a calcium-activated protease (11) and the serine protease thrombin (19, 20), both activity dependent. Quite recently, an extracellular matrix protein called "reelin," also possessing a serine protease activity, has been claimed to be important for the physiology of synapse elimination. In fact in the mutant reeler mouse, in which reelin is lacking, these authors have reported that synapse elimination does not take place because innervation of muscle fibers remains polyneuronal into adulthood (24). A theoretical interpretation of how this suggested physiological function of reelin could be linked to the well-established activity dependence of neuromuscular synapse elimination was offered by others (10). However, in a subsequent thorough reexamination of the issue conducted in parallel with two other laboratories (3), we have been unable to confirm the findings of Quattrocchi et al. (24), because we found that synapse elimination proceeds normally and occurs at the right time of development in reeler mice. A comprehensive detailed hypothesis on how synapse competition occurs and finally leads to the elimination of all but one motor terminal is still lacking.
Our last comments are devoted to the physiological significance of synapse elimination. Polyneuronal innervation in the adult condition is a disadvantage on many counts: 1) for force gradation by temporal summation, a regular train of impulses is best for optimal subtetanic summation of twitches; 2) more importantly, for force gradation by recruitment of motor units, polyneuronal innervation would lead to progressively smaller increments of force, precisely the opposite of what is needed physiologically; and 3) given the fact that each motor unit is homogeneous in fiber type, monoinnervation of each myofiber is necessary. A question then to ask is why polyneuronal innervation develops in the first place. It could simply be a mechanism to allow all muscle fibers to receive a minimum of innervation. The discovery of the embryonic synchronous firing of motoneurons of the same pool, however, suggests a further possible reason: initially inhibiting synapse competition allows all motoneurons to obtain a sufficient share in terms of number of target myofibers, irrespective of possible variations in the time of arrival of their axon in the muscle. A better balance of motor unit sizes, certainly also dependent on other factors, would thus be achieved.
|
Acknowledgments
|
|---|
The financial support of MIUR of Italy and Telethon-Italy (grant no. 1002) is gratefully acknowledged.
|
References
|
|---|
- Balice-Gordon RJ and Lichtman JW. Long-term synapse loss induced by focal blockade of postsynaptic receptors. Nature 372: 519–524, 1994.[CrossRef][Medline]
- Betz WJ, Caldwell JH, and Ribchester RR. The effect of partial denervation at birth on development of muscle fibres and motor units in rat lumbrical muscle. J Physiol 303: 265–279, 1980.[Abstract/Free Full Text]
- Bidoia C, Misgeld T, Weinzierl E, Finn S, Buffelli M, Feng G, Cangiano A, Lichtman JW, and Sanes JR. Comment on: "Reelin promotes peripheral synapse elimination and maturation." Science 303: 1977, 2004.[Free Full Text]
- Bliss T and Collingridge G. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361: 31–39, 1993.[CrossRef][Medline]
- Brown MC, Jansen JKS, and Van Essen DC. Polyneuronal innervation of skeletal muscle in new-born rats and its elimination during maturation. J Physiol 261: 387–424, 1976.[Abstract/Free Full Text]
- Buffelli M, Burgess RW, Feng G, Lobe C, Lichtman JW, and Sanes JR. Genetic evidence that relative synaptic efficacy biases the outcome of synaptic competition. Nature 424: 430–434, 2003.[CrossRef][Medline]
- Buffelli M, Busetto G, Cangiano L, and Cangiano A. Perinatal switch from synchronous to asynchronous activity of motoneurons: link with synapse elimination. Proc Natl Acad Sci USA 99: 13200–13205, 2002.[Abstract/Free Full Text]
- Busetto G, Buffelli M, Cangiano L, and Cangiano A. Effects of evoked and spontaneous motoneuronal firing on synapse competition and elimination in skeletal muscle. J Neurocytol 32: 795–802, 2003.[Medline]
- Busetto G, Buffelli M, Tognana E, Bellico F, and Cangiano A. Hebbian mechanisms revealed by electrical stimulation at developing rat neuromuscular junctions. J Neurosci 20: 685–695, 2000.[Abstract/Free Full Text]
- Chih B and Scheiffele P. Is reelin the answer to synapse elimination at the neuromuscular junction? Sci STKE 2003: pe45, 2003.
- Connold AL, Evers JV, and Vrbova G. Effect of low calcium and protease inhibitors on synapse elimination during postnatal development in the rat soleus muscle. Brain Res 393: 99–107, 1986.[Medline]
- Costanzo EM, Barry JA, and Ribchester RR. Competition at silent synapses in reinnervated skeletal muscle. Nat Neurosci 3: 694–700, 2000.[CrossRef][Medline]
- Hantai D, Rao JS, and Festoff BW. Serine proteases and serpins: their possible roles in the motor system. Rev Neurol (Paris) 144: 680–687, 1988.[Medline]
- Hebb D. The Organization of Behavior. New York: Wiley, 1949.
- Hubel DH and Wiesel TN. Binocular interactions in striate cortex of kittens reared with artificial squint. J Neurophysiol 28: 1041–1059, 1965.[Free Full Text]
- Jansen JK and Fladby T. The perinatal reorganization of the innervation of skeletal muscle in mammals. Prog Neurobiol 34: 39–90, 1990.[CrossRef][ISI][Medline]
- Katz LC and Crowley JC. Development of cortical circuits: lessons from ocular dominance columns. Nat Rev Neurosci 3: 34–42, 2002.[CrossRef][ISI][Medline]
- Lichtman JW and Colman H. Synapse elimination and indelible memory. Neuron 25: 269–278, 2000.[CrossRef][ISI][Medline]
- Liu Y, Fields RD, Festoff BW, and Nelson PG. Proteolytic action of thrombin is required for electrical activity-dependent synapse reduction. Proc Natl Acad Sci USA 91: 10300–10304, 1994.[Abstract/Free Full Text]
- Liu Y, Fields RD, Fitzgerald S, Festoff BW, and Nelson PG. Proteolytic activity, synapse elimination, and the Hebb synapse. J Neurobiol 25: 325–335, 1994.[CrossRef][ISI][Medline]
- Lo YJ and Poo MM. Activity-dependent synaptic competition in vitro: heterosynaptic suppression of developing synapses. Science 254: 1019–1022, 1991.[Abstract/Free Full Text]
- Nguyen QT and Lichtman JW. Mechanism of synapse disassembly at the developing neuromuscular junction. Curr Opin Neurobiol 6: 104–112, 1996.[CrossRef][ISI][Medline]
- Personius KE and Balice-Gordon RJ. Loss of correlated motor neuron activity during synaptic competition at developing neuromuscular synapses. Neuron 31: 395–408, 2001.[CrossRef][ISI][Medline]
- Quattrocchi CC, Huang C, Niu S, Sheldon M, Benhayon D, Cartwright J Jr, Mosier DR, Keller F, and D’Arcangelo G. Reelin promotes peripheral synapse elimination and maturation. Science 301: 649–653, 2003.[Abstract/Free Full Text]
- Rothwell J. Control of Human Voluntary Movement. London: Chapman & Hall, 1994.
- Sanes JR and Lichtman JW. Development of the vertebrate neuromuscular junction. Annu Rev Neurosci 22: 389–442, 1999.[CrossRef][ISI][Medline]
- Stent GS. A physiological mechanism for Hebb’s postulate of learning. Proc Natl Acad Sci USA 70: 997–1001, 1973.[Abstract/Free Full Text]
- Thompson WJ. Synapse elimination in neonatal rat muscle is sensitive to the pattern of muscle use. Nature 302: 614–616, 1983.[CrossRef][Medline]
- Tresch MC and Kiehn O. Synchronization of motor neurons during locomotion in the neonatal rat: predictors and mechanisms. J Neurosci 22: 9997–10008, 2002.[Abstract/Free Full Text]
- Walsh MK and Lichtman JW. In vivo time-lapse imaging of synaptic takeover associated with naturally occurring synapse elimination. Neuron 37: 67–73, 2003.[CrossRef][ISI][Medline]
Copyright © 2004 by the Int. Union Physiol. Sci./Am. Physiol. Soc.
Top
Introduction
Synapse elimination in skeletal...
