- Email: [email protected]
The retzius cells in the leech: A review of their properties and synaptic connections
Camp. Eiochem. Phiol. Printed in Great B&in
Vol. 9lA,
No. 3, pp. 405413,
1988
0300-9629/88 $3.00 + 0.00 C 1988 Pergamon Press plc
MINI REVIEW THE RETZIUS CELLS IN THE LEECH: A REVIEW OF THEIR PROPERTIES AND SYNAPTIC CONNECTIONS MAURO CARRETTA Institute of Human Physiology, Via de1 Giochetto, 1, 06100 Perugia, Italy (Received 21 January 1988) Abstract-l. The Retzius cells (RCs) project an axonal branch in each anterior, posterior and dorsal segmental root. 2. RCs are the only serotonin-containing neurons projecting to the periphery. 3. RCs are activated by mechano-sensory neurons, by serotonin-containing neurons and by two pairs of subesophageal neurons, Tr 1 and Tr 2. 4. RCs also receive an excitatory and an inhibitory input from sensilla. 5. These inputs could form two systems, one converging onto RCs of each ganglion and one distributing to other ganglia after processing by RCs. 6. RCs play a role in muscle tension, in mucous release and in swimming activity.
INTRODUCI’ION In 1891, Gustav Retzius published the results of his work on the morphology and cellular structure of the nervous system of two species of leeches: Hirudo medicinalis and Aulostoma gulo. In this work he described a large number of neurons in each segmental ganglion, using a technique based on vital staining with methylene blue. Among such neurons he found a pair of bilaterally-symmetrical cells with a very large soma: Retzius called such cells “Kolossale Ganglienzellen”. Since then these cells have usually been referred to as Retzius cells (RCs). Many morphological and histological studies have been carried out by authors since 1891, but no significant new information emerged about the two RCs until the ganglia were investigated with histochemical techniques (see section on histochemistry below). Important progress has been made, in a few years, on their electrophysiological properties, their biochemistry and particularly their serotoninergic nature (see section on electrophysioiogy below). Some functional roles of these large neurons in the integrated behavior of the leech have been suggested, e.g. the control of mucus secretion from the skin (Lent, 1973), the effects on the tension of body wall muscle (Mason er al., 1979; Leake et al., 1981; Mason and Kristan, 1982) and a probable role in the regulation of swimming (Willard, 1981; Nusbaum and Kristan, 1983). The Retzius cells appear to be multi-functional and even though their specific role is not yet clear, many pieces of this functional puzzle have begun to fit into place.
ANATOMY
The Retzius cells are a pair of unipolar neurons, situated on the mid-ventral surface of each segmental ganglion of the leech. They have a soma of 50-80 pm diameter and a single major axonal process. This axon runs posteriorly and dorsally through the neuropile, then it turns laterally and bifurcates into two branches. Using methylene blue staining, Retzius (1891) described these branches and he found that they leave the ganglion via the two pairs of segmental roots and run towards the body wall. Occasionally he noted an axonal branch that also projects into the posterior ipsilateral connective. This description was confirmed by the observations of Nicholls and Baylor (1968) who utilized the same histological technique and their preparations showed neurons that were virtually identical to those originally described by Retzius. Only in 1978, however, using horseradish peroxidase as an intracellular stain, Mason and Leake demonstrated that a small axonal branch departs from each major branch and projects into each ipsilateral connective both posteriorly and anteriorly. These branches emerge from the adjacent ganglia towards the periphery via both lateral roots. This organization of axonal branches of each RC can be found in every ganglion of the leech, except in the fifth and sixth ganglia, where the connectival branches are not present (Glover and Mason, 1986). Recently Carretta and Grassi (1987) showed the presence of an axonal branch also in the dorsalposterior root (DP) using various techniques: electrophysiological, catecholamine fluorescence and intracellular injection of Lucifer yellow. These authors demonstrated that electrical stimulation of the DP
MALJRO CARRETTA
406
root elicited an antidromic spike in the ipsilateral RC soma followed by depolarizing synaptic potentials. This response was abolished when the DP root was crushed just proximally to the stimulating electrode. This showed that the response was not due to stimulus spread to the posterior root, but to the presence of an axonal branch running along the DP. The histochemical study of the RC axonal branches was made possible by catecholamine fluorescence (Falk et al., 1962; Marsden and Kerkut, 1969). In all preparations it was possible to observe one or two fluorescing axons in the anterior and posterior roots. One of the axons running in the posterior root gave rise to a collateral branch which entered in the DP. This evidence was fully confirmed by the study of the RC’s axonal branches after intracellular injection of the fluorescent dye Lucifer Yellow CH. Therefore, we can conclude that the anatomical description by Gustav Retzius, even though substantially accurate, today is augmented by new findings that give us a more detailed picture of their anatomy (Fig. 1). HISTOCHEMISTRY
After the work of Ketzius (1891), no significant new information emerged on the colossal cells until the ganglia were investigated with histochemical techniques. Poll and Sommer (1903) were the first to notice that the two RCs and four smaller neurons, located more laterally within each segmental ganglion, had the capacity to reduce chromium salts. At that time such chromium affinity was considered to be strong evidence for the presence of catecholamines, particularly epinephrine. The validity of this theory came into doubt as the techniques of fluorescence histochemistry for monoamines were developed (Falck et al., 1962). This technique is based upon the reaction of monoamines with formaldehyde vapour to produce fluorophores. Excitation of these fluorophores by ultraviolet light produces emissions with characteristic spectra: the products of catecholamines emit blue-green colors while those of the indolealkylamine serotonin (5HT) fluoresce a brilliant yellow colour. Taking advantage of these facts, the authors could demonstrate that the RCs contain serotonin, moreover they found other five serotonin containing neurons in each
Fig. 1. Schematic
representation
of the RC’s branching
ganglion of the leech: two symmetrically paired and one posterior-medial unpaired. These data were confirmed by many studies of various authors (Kerkut et al., 1967; Ehinger et al., 1968; Rude, 1969; Marsden and Kerkut, 1969); they confirmed also the presence of other serotonincontaining neurons in addition to the two RCs. The number of these neurons ranges from five to eleven in each segmental ganglion; generally the number decreases proceeding from the anterior to the posterior end of the ventral nerve cord (Rude, 1969; Lent, 1977): Within a typical middle ganglion it is possible to find the two RCs, two bilaterally symmetrical pairs of neurons, one dorsal (DL) and one ventral (VL), and a single posterior-medial neuron which is absent in the caudal ganglia. Moreover, two other serotonincontaining neurons are localized in the first three ganglia of the ventral nerve cord, immediately anterior with respect to the RCs; they are called E cells and are also present in the sub-esophageal ganglia (Sub EG 14) (Lent, 1982; Glover, 1984). Two other neurons can be found in the Sub EG 1 ganglion, they have a large soma and are called lateral cells (LL) (Lent, 1977; Lent and Dickinson, 1984). Finally a dopamine-containing neuron is located in the anterior segmental root of each ganglion (Kerkut et al., 1967; Ehinger et al., 1968; Rude, 1969; Marsden and Kerkut, 1969). It projects its axon towards the ganglion (Stuart et al., 1974), where it bifurcates in two branches and each of them runs in the ipsilateral connective both anteriorly and posteriorly. These histochemical investigations have also demonstrated that only the RCs project their axons to the periphery, whereas the other serotonincontaining neurons have their axons confined within the CNS of the leech and so they are assumed to be interneurons (Lent and Frazer, 1977; Lent, 1981). Recently, Nusbaum and Kristan (1986) have shown that the axons of two serotonin-containing neurons, cells 61 and 21 (Nusbaum and Kristan, 1982), leave the ganglion via the anterior and the posterior connectives and extend as far as the two nearest ganglia (Fig. 2). In other studies it has been shown that 5-HT reacts specifically with anti-S-HT antibody (Osborne et al., 1982; Stuart et al., 1984) and the S-HT present in the RCs in eccentric dense core vesicles label when the ganglia are incubated with radioactive 5-HT (Coggeshall, 1972). In agreement with these techniques, Leake et al. (1985) have used a specific
pattern,
drawn
from the data of histofluorescence
Retzius cells in the leech
407
Fig. 2. Schematic representation of the serotonin-containing neurons’ (cells 21 and 61) branching pattern and synaptic connections with their homologues of the adjacent ganglia. fluorescent-labelled 5-HT antibody to investigate the distribution of 5-HT-immunoreactive fibres in a variety of peripheral tissues in the medicinal leech. All the tissues examined contain immunoreactive fibres although the density of innervation in the different tissues varies (Leake, 1986); particularly, immunoreactive fibres were found in visceral tissues, such as lateral blood vessels, reproductive organs, nephridial bladder and in the body wall muscles. These findings lend further support to the hypothesis that there is a serotoninergic system in each ganglion of the leech; such a system is composed of the RCs which project their axons to the periphery and other serotonin-containing neurons which are probably devoted to interganghonic connections. ELECTROPHYSIOLOGY
Hagiwara and Morita were the first authors who recorded from the RCs with intracellular microelectrodes (1962) and their findings were followed nearly simultaneously by similar reports from Eckert (1963). These studies allowed study of the many electrophysiological properties of the RCs. These neurons have membrane resting potentials of - 30 to - 60 mV and their action potentials range between 20 and 50 mV. Thus, the action potentials, recorded within the Retzius soma, do not overshoot the electrical zero (Lebedev, 1967). This observation has led to hypothesis that impulses do not actively invade the Retzius cell soma, but are generated at a locus some distance from the soma, which is invaded only by electrotonic spread (West and Lent, 1974); this active locus has not yet been localized. Since the first studies it is known that the two RCs of each ganglion are coupled by an electrotonic non-rectifying junction (Hagiwara and Morita, 1962; Eckert, 1963). This junction maintains the two neurons at a nearly identical membrane potential and, moreover, maintains a nearly synchronous impulse activity, although the more posterior RC usually has a higher impulse frequency and tends therefore to behave as the pacemaker (Lent, 1972). Moreover, Beleslin (1968) noted that oxidative metabolism appears essential to the junctional continuity of RCs; in fact if the oxidative phosphorylation of ADP to ATP is interrupted by 2,4-dinitrophenol, they are uncoupled by an nreversible increase in junctional resistance. Many investigations on the morphology of the RCs have utilized intracellular stains and markers in an
attempt to localize the possible site of their coupling: Procion yellow (Lent, 1972), tritiated amino acids (Globus et al., 1973), cobalt chloride (Sunderland et al., 1974) and horseradish peroxidase (Mason and Leake, 1978). Only investigations made by Injections of Procion yellow and tritiated amino acids gave results: in fact a very fine process could be discerned bridging the initial segment of the axons of the two RCs. As this process is the only site of contact between the two cells, it is reasonable to assume that it is the probable site of electrical coupling, although no definite evidence is so far available. The studies concerning the action potential of the RCs have shown that various ions are involved in the rising and falling phases of the spike (Hagiwara and Morita, 1962; Eckert, 1963; Lent, 1972; West and Lent, 1974). The increases in sodium conductance (gNa+) during the rising phase of the spike and in potassium (gK+) during the falling phase, were considered responsible for the spike for many years (West and Lent, 1974). If sodium ions are completely replaced by lithium ions in the solution bathing the RCs, there are no action potentials; however, if the ion replacement is partial the RCs are still capable to generate impulses. Moreover, the sodium channels, responsible for action potential generation, are unaffected by tetrodotoxin (TTX) and saxitoxin (STX), that are known for their capacity to block selectively such channels (Lent, 1977). More recently, Kleinhaus (1976) has demonstrated that the depolarizing phase of the action potential is mediated by two kinds of voltage-dependent channels for inward current: a fast, tetrodotoxin-sensitive sodium channel and a slower sodium-calcium channel which is not blocked by TTX but by cobalt, manganese or lanthanum (Kleinhaus, 1976; Kleinhaus and Prichard, 1976). Outward ion currents involved in the repolarizing phase of the action potential use different voltage-dependent and calcium-dependent potassium channels (Kleinhaus and Prichard, 1977; Yang and Lent, 1983).
SYNAPTIC INPUTS TO THE RETZIUS CELLS
Central inputs Since the first investigations by Hagiwara and Morita (1962) it was demonstrated that electrical stimulation of the connectives elicits synaptic potentials in the RCs. Later on, it was shown that single
408
MAUROCARRETTA
shock stimulation of the ventral nerve cord activates the RCs transynaptically through ascending and descending pathways (Gerasimov, 1967; Tereshkov and Fomima, 1969; Lent, 1972; Bagnoli and Magni, 1975). It was also postulated that synaptic coupling between intersegmental pathways and RCs was chemical in nature (Lent, 1972). Such a hypothesis was based on the latency of the postsynaptic signals and the use of high Mg*+ and low Ca* + saline. Recent studies have demonstrated that electrical stimulation of the ventral cord of Hirudo medicinalis elicits in both RCs of each segmental ganglion an EPSP which is composed of an early and a late component (Carretta et al., 1981). The early EPSP is electrical in nature since it is unaffected by displacements of the membrane potential and by high Mg*+ saline, whereas the late one is chemical, being reversed in sign by membrane depolarization and suppressed by high Mg*+ saline. Latency measurements show that electrical and chemical EPSP components are mediated by two different pathways characterized by conduction velocities of 0.5 and 0.3 m/set, respectively. Both pathways run in each of the lateral connectives and propagate impulses in an anterior and posterior direction. Collision experiments show that ascending and descending impulses along the pathway mediating the electrical EPSP travel along the same fibres. Separate stimulation of the lateral connectives and selective inactivation of one RC by pronase (Parnas and Bowling, 1977) show that both pathways converge onto each RC. This convergence is a probable mechanism contributing to the synchrony of firing of the RCs. The location of the neuron somata giving origin to the two excitatory pathways and their characteristics are still unknown. Concerning the connections between various neurones and the RCs, it has been shown that the RCs receive an excitatory input, probably polysynaptic, from mechano-sensory neurons (T,P,N) (Mason, personal communication; Carretta, unpublished) and from monoamine-containing neurons. The latter neurons are electrotonically connected with each other and with the RCs (Lent and Frazer, 1977) by non-rectifying electrical junctions. In a work devoted to the initiation of swimming activity, Brodfuehrer and Friesen (1986) have found two bilaterally symmetrical pairs of interneurons located in the subesophageal ganglion and named them Tr 1 and Tr 2. These cells are considered trigger neurons because the swimming episodes elicited by their electrical stimulation outlasts the stimulus duration and the length of the swim episodes is largely independent of the intensity of Tr 1 and Tr 2 stimulation. Tr 1 and Tr 2 axons extend as far posterior as the segmental ganglion 18 of the ventral nerve cord. Tr 1 elicits excitatory input to three groups of segemntal neurons: swim-initiating neurons (cells 204 and 205, Weeks, 1982b), serotonin-containing neurons (cells 61 and 21, Nusbaum and Kristan, 1982) and the RCs. Moreover, all RCs in the subesophageal ganglion are excited directly by Tr 1. On the contrary, Tr 2 stimulation elicits transient inhibition in swim-initiating and serotonin-containing neurons and has little effect on RCs.
Peripheral inputs In 1967 Gerasimov demonstrated that light (photic) and mechanical stimulation of the anterior and posterior suckers activate the RCs producing an EPSP. Laverak (1969) showed the existence of a bidirectional fast conducting system (FCS) activated by photic and mechanical stimulation of the skin but could not find any relation between such system and the activity of the RCs. After these studies, many authors turned their attention to the question of the peripheral inputs to the RCs. Bagnoli and Magni (1975) showed that photic and mechanical stimulation of the anterior and posterior parts of the body wall activated the RCs transynaptically through ascending and descending pathways running along the lateral connectives (Bagnoli et al., 1972, 1973). Moreover, these authors could demonstrate that, although some similarities exist between the FCS and the pathways which activate the RCs, namely bidirectionality of conduction and mode of activation, the response of the RCs to natural stimuli was not mediated by the FCS. Recently, some studies demonstrated that waterwave stimulation activates the sensillar movement receptors (SMR) (Friesen, 1981; Brodfuehrer and Friesen, 1984); the sensilla are also activated by photic stimulation (Kretz et al., 1976). Water-wave stimulation elicits swimming in the intact leech and in semi-intact preparation if serotonin is added to the physiological saline (Young et al., 1981; Kristan and Nusbaum, 1983). On the other hand it is known that intracellular stimulation of the RCs can elicit swimming by releasing serotonin in the blood (Willard, 1981) and that they are activated by mechanosensory inputs (Bagnoli and Magni, 1975), by vibrational and thermal stimuli applied to the prostomium (Lent and Dickinson, 1984). The results of studies performed in our laboratory are in keeping with these findings. In a first study we have shown that electrical stimulation of the segmental roots of each ganglion of Hirudo medicinalis, elicits in both RCs inhibitory and excitatory effects (Carretta et al., 1985). The IPSP and EPSP are chemical in nature, being dependent on the membrane potential and suppressed by high Mg*+. Selective inactivation of one RC by pronase, shows that the responses of the controlateral RC are not due to electrotonic coupling between the two cells, but to synaptic actions impinging upon the membrane of both RCs. The two synaptic potentials appear to be mediated by two sets of fibres with a different threshold to electrical stimulation. Their actions on the RCs appear to be polysynaptic on the basis of central latency. Simultaneous stimulation of two roots shows evidence for occlusion for IPSP and summation for EPSP, confirming the polysynaptic nature of the effects. In a further work (Carretta and Zampolini, 1987) the source of the afferent inputs to the RCs was investigated: in particular, whether the IPSP and EPSP could be mediated by some of the mechanosensory neurons (T,P,N) or by other afferent systems, such as the sensilla. The first possibility gave negative results; in fact there is no relation between the activity
Retzius cells in the leech of T, P or N cells and IPSP in the RCs, whereas intracellular stimulation of these neurons can sometimes elicit an EPSP in the RCs (Carretta, unpublished data). On the contrary, water-wave and photic stimulation of the sensilla elicits synaptic potentials identical to those elicited by electrical stimulation of the segmental roots. Mechanical stimuli produced an IPSP in the RCs when a single-segment preparation is used, only an EPSP when the preparation consists of three or more segments. Such an IPSP is the only inhibitory input to the RCs described up to now. Photic stimulation gives always rise to an EPSP independently of the type of preparation. The impulse discharge elicited in the afferent fibres by the two kinds of stimuli is transmitted along the ventral cord both anteriorly and posteriorly to the stimulated segment. This implies that the afferent impulse excites a common pool of intersegmental neurons in each ganglion, which distribute their discharges to the adjacent ganglia. The evidence for occlusion between cordal and photically elicited volleys indicates that the power shares with the sensillar input, a common pool of interneurons. Some evidence has been produced that the serotonin-containing neurons are a component of the common pool of excitatory interneurons (Carretta and Zampolini, 1987). It has been found that waterwave and photic stimulation of the sensilla elicits excitatory post-synaptic potentials in cells 61 and 21. Such potentials are chemical in nature and precede the EPSP of the RCs. These electrophysiological data are presently being verified by an attempt to label cells 61 and 21 following cobalt injection into the sensilla. Since we have found that the intracellular stimulation of cell 61 elicits an IPSP in the T cells it can be postulated that serotonin-containing neurons are one component of the pool of excitatory interneurons to the RCs and that they are also responsible for the inhibition of other afferent inputs to the RCs, thus allowing the access of only the sensillar inputs to their target neurons.
OUTPUTS
OF THE RETZIUS
CELLS
Several anatomical investigations have demonstrated that the RCs send axon branches to the periphery via all segmental roots (see above section on Anatomy). Recently Leake et al. (1985) have found 5-HT immunoreactive fibres in many peripheral tissues in the leech, using a specific fluorescent 5-HT-antibody. Since the RCs are the only serotoninergic neurons which send their axons out of the ganglion, they are the likely source of such 5-HT immunoreactive fibres. On this anatomical basis three different actions of the RCs have been demonstrated in the leech, Hirudo medicinalis. Lent (1973) showed that mucous release from the skin of leeches is under the control of the pair of RCs of each segmental ganglion. In his work Lent could find a relation between the rate of mucous release and number of RC spikes, but he concluded that the data reported did not allow any decision as to whether such an effect is mediated by a synaptic or neurosecretory process.
409
Later Mason et al. (1979) showed that the RCs have some effect on body wall muscles. Particularly, they found that the electrical stimulation of the RC caused a decrease in basal tension of longitudinal muscles of the ipsilateral body wall and the accelerated relaxation following contractions elicited by the motorneuron of longitudinal muscle (L cell). These effects persisted when the ganglion was bathed in 20mM Mg*+ saline. All these effects outlasted the period of impulse protection by the RCs by several minutes and were mimicked by serotonin, the probable transmitter of the RC. More recently Kristan and Nusbaum (1983) have demonstrated a possible neurosecretory role for the RCs. In fact, intracellular stimulation of a RC produces a release of 5-HT in blood vessels and this promotes swimming activity. The same effect was obtained by the addition of 5-HT to the bathing saline (Willard, 1981). No connections are known to exist between the RCs and other neurons in the central nervous system of the leech in vivo. The connections between the RCs and other neurons have been the subject of several investigations carried out on isolated leech neurons and RCs cultivated in vitro (Fuchs et al., 1981, 1982; Henderson et al., 1983). Particularly, these studies have shown that the RCs make a monosynaptic inhibitory connection with the P cell, and that the transmitter liberated by isolated RCs, serotonin, evokes chloride-dependent inhibitory post-synaptic potentials on the P cell. Some discrepancies remain, however, between the results obtained in vivo and in vitro; the most important of them is the question about the differences of electrical properties of these neurons shown in the two experimental situations. NEUROTRANSMIITERS
The nature of the neurotransmitters which excite and/or inhibit the RCs and of their cellular receptors, has received much attention. Kerkut and Walker (1967) demonstrated that the spontaneous activity of the RCs was increased following the addition of acetylcholine (Ach) to the bathing saline or its iontophoretic application. The response to Ach could be reduced or completely blocked by pretreating the preparation with atropine: such a block was readily reversed. The same authors showed that the spontaneous activity of the RCs decreased following the application of 5-HT; the threshold amount of 5-HT was 100 times lower than that of Ach. Further studies have suggested that the RCs possess both nicotinic excitatory receptors and muscarinic inhibitory receptors (Woodruff et al., 1971). The ionic mechanism associated with the inhibition induced by 5-HT on the RCs has been investigated by Walker and Smith (1973) and found to be predominantly a chloride event. Chloride dependent inhibitory post-synaptic potentials (IPSPs) could be recorded from the RCs and these IPSPs could be reversibly blocked by morphine. Pretreatment of the ganglion with imipramine potentiated the duration of the 5-HT inhibition, suggesting the presence of an uptake system for 5-HT. It was also shown (Kerkut and Walker, 1967) that the RCs are hyperpolarized
MAURO CARRETTA
410
by dopamine (DA) and its mechanism of action was investigated and compared with that of 5-HT (Sunderland et al., 1979). Like 5-HT, it was found that the DA hyperpolarization is a chloride conductance increase event. It was concluded that DA and 5-HT activated the same type of chloride ionophore in the membrane of the RCs. Since it was clear that both 5-HT and DA act via chloride ionophores on RCs, an attempt was made to clarify whether each amine is acting on its own specific receptors or not (Sunderland et al.. 1980; Leake et al., 1980). The evidence so far produced suggests, rather surprisngly, that DA and 5-HT are acting via the same receptors. DlSCUSSlON
Since the first morphological description of the two Kolossale zellen by Gustav Retzius (1891), this pair of neurons has been studied at all levels, ranging from histology to histochemistry, and from electrophysiology to synaptic inputs; so today we have a large amount of data on the functioning of RCs as neuronal entities. Morphologically, we know the whole axonal organization-the projections of the axonal branches to the periphery (Retzius, 1891; Falck et al., 1962; Kerkut et al., 1967; Nicholls and Baylor, 1968; Rude, 1969; Stuart et al., 1974; Lent, 1977; Mason and Leake, 1978; Carretta and Grassi, 1987) and there are some interesting observations about the distribution of their terminal fibres (Leake, 1986). Histochemically, the serotoninergic nature of the RCs has been demonstrated (Falck et al., 1962; Kerkut et al., 1967; Ehinger et al., 1968; Rude, 1969; Marsden and Kerkut, 1969) and the role of 5-HT as their putative neurotransmitter has been firmly established (Leake et al., 1980; Leake et al., 1981; Gardner and Walker, 1982; Henderson, 1983; Leake, 1986). The electrophysiological properties of the RCs have been known since Hagiwara and Morita (1962) recorded from the neuron somata by microelectrodes. Therefore, we know their resting potentials, the characteristics of action potentials and ions responsible for the raising and falling phases, the properties of electrotonic coupling between the two cells (Eckert, 1963; Lebedev, 1967; Beleslin, 1968; Lent, 1972; West and Lent, 1974; Kleinhaus, 1976; Kleinhaus and Prichard, 1977; Yang and Lent, 1983; Leake et al., 1986). The only open question is the localization of the electrotonic junction.
Only now are we arriving at an acceptable picture of the synaptic inputs to the RCs. It is known that the RCs are activated by ascending and descending pathways running along the ventral cord (Hagiwara and Morita, 1962; Gerasimov, 1967; Lent, 1972; Bagnoh and Magni, 1975; Carretta et al., 1981); they also receive excitatory inputs from: (a) mechano-sensory neurons, cells T,P,N (Mason, personal communication; Carretta unpublished data); (b) serotonin-containing neurons (Lent and Frazer, 1977; Lent and Dickinson, 1984; Nusbaum and Kristan, 1986); (c) two pairs of neurons located in the subesophageal ganglion, Tr 1 and Tr 2 (Brodfuehrer and Friesen, 1986a); (d) sensillar receptors activated by both photic and water-wave stimulation (Kretz et al., 1976; Friesen, 1981; Brodfuehrer and Friesen, 1984; Lent and Dickinson, 1984; Carretta and Zampolini, 1987). The RCs also receive an inhibitory input from the sensilla (Carretta et al., 1985; Carretta and Zampolini, 1987); this is the only inhibitory input to these neurons described up to now. From all these data it is possible to make an attempt at a physiological interpretation of the inputs described above. One could speculate that these inputs form two different systems, one converging onto the two RCs of each ganglion and one distributing to several ganglia after processing by the RCs and/or other serotonin-containing neurons. The converging system could be the Tr 1 and Tr 2 system that runs from the sub-esophageal ganglion, throughout the whole ventral nerve cord and provides direct excitatory drive to cells 21, 61 and RCs, as well as to the swim-initiating interneurons (cells 204 and 205) (Brodfuehrer and Friesen, 1986a); such excitatory input to the RCs is chemical in nature. The distributing system could be considered the 21 and 61 cells system with their axonal branches running into both anterior and posterior connectives as far as the two nearest ganglia in each direction (Nusbaum and Kristan, 1986). These cells are electrically coupled to each other, with contralateral homologues and with RCs, even though the electrical coupling between a cell 21 or 61 and any other cell 21, 61 or RC is sufficiently weak that each cell may be considered as functionally independent (Fig. 3). These two systems, a chemical one and an electrical one, have some similarity with the two bidirectional nerve cord systems converging with electrical and
Fig. 3. Diagrammatic representation of the proposed synaptic connections of the Tr 1-Tr 2 and cells 21 and 61 systems
to RCs. Open arrows indicate chemical inhibitory synapses, closed electrical synapses and lines indicate chemical excitatory synapses.
arrows
indicate
Retzius
cells in the leech
chemical synapses onto the Retzius cells of the leech described by Carretta et al. (1981). The two systems have an important role in the swimming activity of the leech and are characterized by a common point of contact: the two Retzius neurons. Another aspect of this question is the sensory inputs to the two systems. Tr 1 and Tr 2 neurons are excited by weak mechanical stimulation of body wall such as by pinching of the skin and this activation of Tr 1 and Tr 2 may be limited to tactile mechano-sensory neurons (Brodfuehrer and Friesen, 1986b). Mechanosensory stimulation of the leech body wall in semi-intact preparations, sufficient to initiate swimming, excited cells 21, 61 and RCs (Nusbaum and Kristan, 1986). Such stimulation also directly activates identified T, P, and N neurons; intracellular stimulation of these neurons mimics the effect of body wall stimulation onto cells 21 and 61. All these effects were found to be polysynaptic. Finally, it has been shown that vibrational stimuli applied to the sensilla excite RCs (Lent and Dickinson, 1986) water-wave and photic stimulation of sensilla elicits synaptic potentials in the RCs and in cells 21 and 61 (Carretta and Zampolini, 1987a, b; Carretta and Zampolini, in press). In the light of all these observations, one is led to speculate that the RCs are part of a complex neuronal system playing a modulatory role in regulating sensory afferent inputs necessary for swimming activity of the leech. Moreover, they are subject to a general neuronal control by subesophageal ganglia and to a local segmental control by sensilla and/or other afferent sensory inputs. The functional significance of this double control in the integrated behaviour of the animal can be envisaged as follows: (a) A local response leading to the adjustment at segmental level of the tension of the longitudinal muscle, which is inhibited by the RCs (Mason et al., 1979). (b) A generalized intersegmental excitation which could cause an inhibition of the longitudinal muscle (by accelerating its relaxation rate) (Mason et al., 1979) brought about by the extensive activation of the RCs. These effects are probably the pre-requisite for the initiation of swimming. Obviously all these responses are expected to interact with each other and/or with other stimuli present in a complex situation, such as that occurring in the natural environemnt. It is interesting to observe that the RCs play their modulatory role in the swimming of the leech by two different actions: one directly on the longitudinal muscle of the body wall (Mason et al., 1979; Leake et al., 1981; Mason and Kristan, 1982) and a generalized one on the whole animal by releasing 5-HT in the blood vessel (Willard, 1981). Since a clear connection between the axon of RCs and serotoninergic receptors on the longitudinal muscle of the leech has not been demonstrated, it is conceivable that this action might also be due to a neuro-secretory effect. It should be pointed out that this model is certainly oversimplified, in so far as it neglects possible interactions of other non-identified inputs and/or outputs.
411
In spite of these short-comings, however, it may provide an unifying neurophysiological explanation of the functional significance of these two “Kolossale Gangiienzellen” discovered last century by Gustav Retzius. Acknowledgements--I
thank Prof. Franc0 Magni both for reading and making helpful suggestions about this manuscript, and Dr Silvarosa Grassi and Dr Mauro Zampolini
for collaboration
in some experiments.
REFERENCES
Bagnoli P. and Magni F. (1975) Synaptic inputs to Retzius cells in the leech. Brain Res. 96, 147-152. Bagnoli P., Brunelli M. and Magni F. (1973) Afferent connections to the fast conduction pathway in the central nervous system of the leech Hirudo medicinalis. Archs Ital. Biol. 111, 58-75. Bagnoli P., Brunelli M., Magni F. and Pellegrino M. (1975) The neurons of the fast conducting system in Hirudo medicinalis: identification and synaptic connections with primary afferent neurons. Archs. Ital. Biol. 113, 2143. Beleslin B. (1977) Lack of electrotonic transmission between Retzius nerve cells in the fifth and sixth free ganglion of horse leech Haemopis sanguisuga. Comp. Biochem. Physiol. S6A, 509-5 12. Brodfuehrer P. D. and Friesen 0. W. (1984) A sensory system initiating swimming activity in the medicinal leech. J. exp. Biol. 108, 341-355. Brodfuehrer P. D. and Friesen 0. W. (1986) Initiation of swimming activity by trigger neurons in the leech subesophageal ganglia. I. Outputs connections of Tr 1 and Tr 2. J. camp. Physiol. 159A, 489-502. Brodfuehrer P. D. and Friesen 0. W. (1986) Intitiation of swimming activity by trigger neurons in the leech subesophageal ganglia. III. Sensory inputs to Tr 1 and Tr 2. J. Comp. Physiol. 159A, 51 l-519. Carretta M., Grassi S. and Magni F. (1981) Two bidirectional nerve cord systems converging with electrical and chemical synapses on the Retzius cells of the leech Hirudo medicinalis. Archs Ital. Biol. 119, l-177. Carretta M., Grassi S. and Magni F. (1985) Synaptic effect elicited in the Retzius ceils of the leech Hirudo medicinalis by stimulation of the segmental roots. Archs Ital. Biol. 123, 227-239. Carretta M. and Grassi S. (1987) Axonal projections of the Retzius cells to the dorsal segmental root in the leech Hirudo medicinalis. Archs Ital. Biol. 125, 3845. Carretta M. and Zampolini M. (1987) Afferent sensillar inputs to the Retzius cells of the leech Hirudo medicinalis. Archs Ital. Biol. 125, 45-57. Carretta M. and Zamoolini M. (19871 Identification of some interneurons impinging upon the Retzius cells in the leech. Riunione primaverile Sot. Ital. Fisiol., Florence. Coggeshall R. E. (1972) Autoradiographic and chemical localization of 5-hydroxytryptamine in identified neurons in the leech. Anat. Rec. 172, 489498. Eckert R. (1963) Electrical interaction of paired ganglion cells in the leech. J. Gen. Physiol. 46, 573-587. Ehinger B., Falck B. and Myhrberg H. E. (1968) Biogenic monoamines in Hirudo medicinalis. Histochemie 15, 14c-149. Falck B., Hillarp N. A., Thieme G. and Torp A. (1962) Fluorescence of catecholamines and related compounds condensed with formaldehyde. Histochem. Cytochem. 10, 348-354. Friesen 0. W. (1981) Physiology of water motion detection in the medicinal leech. J. exp. Biol. 92, 255-275. Fuchs P. A., Nicholls J. G. and Ready D. F. (1981) Membrane properties and selective connexions of identified leech neurons in culture. J. Physiol. 316, 203-225.
MAUROCARREITA
412
Fuchs P. A., Henderson L. P. and Nicholls J. G. (1982) Chemical transmission between individual Retzius and sensory neurones of the leech in culture. J. Physiol. 323, 195-211. Gardner C. R. and Walker R. J. (1982) The roles of putative neurotransmitters and neuromodulators in annelids and related invertebrates. Prog. Neurobiol. 18, 81-120. Gaskell J. F. (1914) The chroafline system of annelids and the relationship of this system to the contractile vascular system in the leech Hirudo medicinalis. A contribution to the comparative physiology of the contractile vascular system and its regulators, the adrenalin secreting system and sympathetic nervous system. Phil. Trans. R. Sot. 205, 153-211.
Gerasimov V. D. (1967) Electrical properties and connections of CNS giant nerve cells of Hirudo medicinalis. In Symposium on Neurobiology of Inveriebrares (Edited bv Salanki J.), pp. 285-292. __ _ Globus H.. Lux H. D. and Schubert P. (1973) Transfer of \ amino acids between neuroglia cells and neurons in the leech ganglion. Exp. Neural 40, 104-l 13. Haaiwara S. and Morita H. (1962) Electronic transmission between two nerve cells in leech ganglion. J. Neurophysiol. I
25, 721-731.
Henderson L. P. (1983) The role of 5hydroxytryptamine as a transmitter between identified leech neurones in culture. J. Physiol. 339, 309-324.
Henderson L. P., Kuffler D. P., Nicholls J. G. and Zhang R. J. (1983) Structural and functional analysis of synaptic transmission between identified leech neurones in culture. J. Physiol. 340, 347-358.
James V. A., Walker R. J. and Wheal H. V. (1980) Structure activity studies on an excitatory receptor for glutamate on leech Retzius neurones. Brit. J. Pharmacol. 68, 711-717.
Kerkut G. A. and Walker R. J. (1967) The action of acetylcholine, dopamine and 5-hydroxytryptamine on the spontaneous activity of the cell of Retzius of the leech Hirudo medicinalis. Brit. J. Pharmacol. Chemorher. 30, 64&654.
Kerkut G. A., Sedden C. B. and Walker R. J. (1967) Cellular localization of monoamines by fluorescence microscopy in Hirudo medicinalis and Lumbricus terrestris. Comp. Biochem. Physiol. 21, 687490.
Kleinhaus A. L. (1976) Divalent cations and the action potential of leech Retzius cells. Pfliigers Arch. 363, 97-104.
Kleinhaus A. L. and Prichard J. W. (1976) Sodium dependent tetrodotoxin resistant action potentials in a leech neurone. Brain Res. 102, 368-372. Kleinhaus A. L. and Prichard J. W. (1977) Close relation between Tea response and Ca-dependent membrane phenomena of four identified leech neurones. J. Physiol. 270, 181-194.
Kretz J. R., Stent G. S. and Kristan W. B. Jr (1976) Photosensory input pathways in the medicinal leech. J. camp. Physiol. 106, 1-37. Kristan W. B. Jr and Nusbaum M. P. (1983) The dual role of serotonin in leech swimming. J. Physiol. Paris 78, 743-747.
Leake L. D. (1986) Leech Retzius cells and 5-hydroxytryptamine. Comp. Biochem. Physiol. 83C, 229-239. Leake L. D., Sunderland A. J. and Walker R. J. (1980) Is there a common receptor mediating the inhibitory actions of 5 HT and dopamine on leech Retzius cells? Brit. J. Pharmac. 70, 139-149.
Leake L. D., Mason A. J. and Sunderland A. J. (1981) Is S-hydroxytryptamine a nueromuscular transmitter in the leech? Adr. Physiol. Sci. 22, 391406. Leake L. D., Griffith S. G. and Bumstock G. (1985) 5-hydroxytryptamine-like-immunoreactivity in the peripheral and central nervous system of the leech Hirudo medicinalis. Cell. Tiss. Res. 239, 123-130.
Lebedev A. N. (1967) Spontaneous impulse potentials of single neurons in namisolated nerve ganglia of Hirudo medicinalis. Fiziol. Zh. SSSR Sechenov. 53, 915-921.
Lent C. M. (1972) Electrophysiology of Retzius cells of segmental ganglia in the horse leech Haemopis marmorata. Comp. Biochem. Physiol. 42, 857-886.
Lent C. M. (1973) Retzius cells: neuroeffectors controlling mucous release by the leech. Science 179, 693496. Lent C. M. (1977) The Retzius cells within the central nervous system of leeches. Prog. Neurobiol. 8, 81-l 17. Lent C. M. (1981) Morphology of neurons containing monoamines within leech segmental ganglia. J. exp. Zool. 216, 31 l-316.
Lent C. M. and Frazer B. M. (1977) Connectivity of the monoamine containing neurons in central nervous system of leech. Nature, Lond. 266, 844-846. Lent C. M. and Dickinson M. H. (1984) Serotonin integrates the feeding behavior of the medicinal leech. J. camp. Physiol. 154, 457471. Mason A. and Leake L. D. (1978) Morphology of leech Retzius cells demonstrated by intracellular injection of horseradish peroxidase. Comp. Biochem. Physiol. 61A, 213-216.
Mason A. and Kristan W. B. Jr (1982) Neuronal excitation, inhibition and modulation of leech longitudinal muscle. J. camp. Physiol. 146, 5277536. Mason A., Sunderland A. J. and Leake L. D. (1979) Effects of leech Retzius cells on body wall muscles. Comp. Biochem. Physiol. 63, 359-361,
Marsden C. A. and Kerkut G. A. (1969) Fluorescence microscopy of the 5-HT and catecholamine containing cells in the central nervous system of the leech Hirudo medicinalis. Comp. Biochem. Physiol. 31, 85 1-862.
Nicholls J. G. and Baylor D. A. (1968) Specific modalities and receptive fields of sensory neurons in CNS of the leech. J. Neurophysiol. 16, 740-756. Nusbaum M. B. and Kristan W. B. Jr (1982) The swiminitiating ability of intersegmental serotonin-containing leech interneurons. Neurosci. Abstr. 8, 161-168. Nusbaum M. B. and Kristan W. B. Jr (1986) Swim initiation in the leech by serotonin containing intemeurons cells 21 and cells 61. J. exp. Biol. 122, 277-302. Osborne N. N., Pate1 S. and Dockray G. (1982) Immunohistochemical demonstration of peptides, serotonin and dopamine B hydroxylase like material in the nervous system of the leech Hirudo medicinalis. Histochemistry 75, 573-583.
Parnas I. and Bowling D. (1977) Killing of single neurons by intracellular injection of proteolytic enzymes. Nature, Lond. 270, 626628.
Poll H. and Sommer A. (1903) Ueber phacchrome Zellin in centranervesystem des Blutegels. Verk. Physiol. Ges. Berl. 10, 549-550. Retzius G. (1891) Zur Kenntnis des centralen Nervensystem der Wurmer, das Nervensystem der Annulaten. Biol. Untersuch. (N:F:) 2, l-28. Rude S. (1969) Monoamine containing neurons in the central nervous system and peripheral nerves of the leech Hirudo medicinalis. J. camp. Neurol. 136. 349-372. Rude S., Coggeshall R. E.-and Van Order L. S. (1969) Chemical and ultrastructural identification of 5-hydroxytryptamine in an identified neuron. J. Cell Biol. 41, 832-854. Stuart A. E., Hudspeth A. J. and Hall Z. W. (1974) Vital staining of specific monoamine containing cells in the leech nervous
system.
Cell. Tiss. Res. 153, 55-61.
Stuart A. E., Glover J. C., Blair S. S. and Weisblat D. A. (1984) Development of leech serotonin neurons examined with serotonin antibody cell lineage tracer and cell killing. Sot. Neurosci. Abstr. i0, 605%X)5. Sunderland A. J.. Smith P. A.. Ledke L. D. and Walker R. J. (1974) Neuronal geometry of Retzius cells in Hirudo medicinalis. Experientia 30, 918-919.
Retzius
cells in the leech
Sunderland A. J., Leake L. D. and Walker R. J. (1979) The ionic mechanism of the dopamine response in Retzius cells of two leech species, Hirudo and Haemopis. Comp. Biochem. Physiol. 63C, 129-l 33. Sunderland A. J., Leake L. D. and Walker R. J. (1980) Evidence for an amine receptor on the Retzius cells of the leeches Hirudo and Haemopis. Comp. Biochem. Physiol. 63C, 129-133. Tereshkov 0. D. and Fomina M. S. (1969/70) Electrophysiological investigation of the system of paired giant cells in the ventral chain of Autostoma gulo. Neurosc. Transl. 12, l&16. Yang J. and Lent C. M. (1981) Calcium depletion produces Na+-dependent sustained depolarization of Retzius cells membranes in the leech CNS. J. camp. Physiol. 144, 111L116. Walker R. J. and Smith P. A. (1973) Ionic mechanism for S-hydroxytryptamine inhibition on Retzius cells of the leech Hirudo medicinalis. Comp. Biochem. Physiol. 45, 979-99 I. Weeks J. C. (1981) Neuronal basis of leech swimming separation of swim initiation pattern generation and
413
intersegmental coordination by selective lesions. J. Neurophysiol. 45, 698-724. Weeks J. C. (1982) Segmental specialization of a leech swim initiating interneuron cell 205. J. Neuroscience 2, 972-986. Weeks J. C. (1982) Synaptic basis of swim initiation in the leech. II: A pattern generating neuron (cell 208) which mediates motor effects of swim initiating neurons. J. camp. Physiol. 148, 265-279. Weeks J. C. and Kristan W. B. Jr (1978) Initiation, maintenance and modulation of swimming in the medicinal leech by the activity of a single neuron. J. exp. Biol. 77, 71-88. West C. H. K. and Lent C. M. (1974) Effects of temperature on the electrophysiology of Retzius cells in the leech. Comp. Biochem. Physiol. 47A, 27-38. Willard A. L. (1981) Effects of serotonin on the generation of the motor program for swimming by the medicinal leech. J. Neurosci. 1, 936-944. Woodruff G. N., Walker R. J. and Newton L. C. (1971) The action of same muscarinic and nicotinic agonists on the Retzius cells of the leech. Comp. gen. Pharmac. 2, 106117.
Vol. 9lA,
No. 3, pp. 405413,
1988
0300-9629/88 $3.00 + 0.00 C 1988 Pergamon Press plc
MINI REVIEW THE RETZIUS CELLS IN THE LEECH: A REVIEW OF THEIR PROPERTIES AND SYNAPTIC CONNECTIONS MAURO CARRETTA Institute of Human Physiology, Via de1 Giochetto, 1, 06100 Perugia, Italy (Received 21 January 1988) Abstract-l. The Retzius cells (RCs) project an axonal branch in each anterior, posterior and dorsal segmental root. 2. RCs are the only serotonin-containing neurons projecting to the periphery. 3. RCs are activated by mechano-sensory neurons, by serotonin-containing neurons and by two pairs of subesophageal neurons, Tr 1 and Tr 2. 4. RCs also receive an excitatory and an inhibitory input from sensilla. 5. These inputs could form two systems, one converging onto RCs of each ganglion and one distributing to other ganglia after processing by RCs. 6. RCs play a role in muscle tension, in mucous release and in swimming activity.
INTRODUCI’ION In 1891, Gustav Retzius published the results of his work on the morphology and cellular structure of the nervous system of two species of leeches: Hirudo medicinalis and Aulostoma gulo. In this work he described a large number of neurons in each segmental ganglion, using a technique based on vital staining with methylene blue. Among such neurons he found a pair of bilaterally-symmetrical cells with a very large soma: Retzius called such cells “Kolossale Ganglienzellen”. Since then these cells have usually been referred to as Retzius cells (RCs). Many morphological and histological studies have been carried out by authors since 1891, but no significant new information emerged about the two RCs until the ganglia were investigated with histochemical techniques (see section on histochemistry below). Important progress has been made, in a few years, on their electrophysiological properties, their biochemistry and particularly their serotoninergic nature (see section on electrophysioiogy below). Some functional roles of these large neurons in the integrated behavior of the leech have been suggested, e.g. the control of mucus secretion from the skin (Lent, 1973), the effects on the tension of body wall muscle (Mason er al., 1979; Leake et al., 1981; Mason and Kristan, 1982) and a probable role in the regulation of swimming (Willard, 1981; Nusbaum and Kristan, 1983). The Retzius cells appear to be multi-functional and even though their specific role is not yet clear, many pieces of this functional puzzle have begun to fit into place.
ANATOMY
The Retzius cells are a pair of unipolar neurons, situated on the mid-ventral surface of each segmental ganglion of the leech. They have a soma of 50-80 pm diameter and a single major axonal process. This axon runs posteriorly and dorsally through the neuropile, then it turns laterally and bifurcates into two branches. Using methylene blue staining, Retzius (1891) described these branches and he found that they leave the ganglion via the two pairs of segmental roots and run towards the body wall. Occasionally he noted an axonal branch that also projects into the posterior ipsilateral connective. This description was confirmed by the observations of Nicholls and Baylor (1968) who utilized the same histological technique and their preparations showed neurons that were virtually identical to those originally described by Retzius. Only in 1978, however, using horseradish peroxidase as an intracellular stain, Mason and Leake demonstrated that a small axonal branch departs from each major branch and projects into each ipsilateral connective both posteriorly and anteriorly. These branches emerge from the adjacent ganglia towards the periphery via both lateral roots. This organization of axonal branches of each RC can be found in every ganglion of the leech, except in the fifth and sixth ganglia, where the connectival branches are not present (Glover and Mason, 1986). Recently Carretta and Grassi (1987) showed the presence of an axonal branch also in the dorsalposterior root (DP) using various techniques: electrophysiological, catecholamine fluorescence and intracellular injection of Lucifer yellow. These authors demonstrated that electrical stimulation of the DP
MALJRO CARRETTA
406
root elicited an antidromic spike in the ipsilateral RC soma followed by depolarizing synaptic potentials. This response was abolished when the DP root was crushed just proximally to the stimulating electrode. This showed that the response was not due to stimulus spread to the posterior root, but to the presence of an axonal branch running along the DP. The histochemical study of the RC axonal branches was made possible by catecholamine fluorescence (Falk et al., 1962; Marsden and Kerkut, 1969). In all preparations it was possible to observe one or two fluorescing axons in the anterior and posterior roots. One of the axons running in the posterior root gave rise to a collateral branch which entered in the DP. This evidence was fully confirmed by the study of the RC’s axonal branches after intracellular injection of the fluorescent dye Lucifer Yellow CH. Therefore, we can conclude that the anatomical description by Gustav Retzius, even though substantially accurate, today is augmented by new findings that give us a more detailed picture of their anatomy (Fig. 1). HISTOCHEMISTRY
After the work of Ketzius (1891), no significant new information emerged on the colossal cells until the ganglia were investigated with histochemical techniques. Poll and Sommer (1903) were the first to notice that the two RCs and four smaller neurons, located more laterally within each segmental ganglion, had the capacity to reduce chromium salts. At that time such chromium affinity was considered to be strong evidence for the presence of catecholamines, particularly epinephrine. The validity of this theory came into doubt as the techniques of fluorescence histochemistry for monoamines were developed (Falck et al., 1962). This technique is based upon the reaction of monoamines with formaldehyde vapour to produce fluorophores. Excitation of these fluorophores by ultraviolet light produces emissions with characteristic spectra: the products of catecholamines emit blue-green colors while those of the indolealkylamine serotonin (5HT) fluoresce a brilliant yellow colour. Taking advantage of these facts, the authors could demonstrate that the RCs contain serotonin, moreover they found other five serotonin containing neurons in each
Fig. 1. Schematic
representation
of the RC’s branching
ganglion of the leech: two symmetrically paired and one posterior-medial unpaired. These data were confirmed by many studies of various authors (Kerkut et al., 1967; Ehinger et al., 1968; Rude, 1969; Marsden and Kerkut, 1969); they confirmed also the presence of other serotonincontaining neurons in addition to the two RCs. The number of these neurons ranges from five to eleven in each segmental ganglion; generally the number decreases proceeding from the anterior to the posterior end of the ventral nerve cord (Rude, 1969; Lent, 1977): Within a typical middle ganglion it is possible to find the two RCs, two bilaterally symmetrical pairs of neurons, one dorsal (DL) and one ventral (VL), and a single posterior-medial neuron which is absent in the caudal ganglia. Moreover, two other serotonincontaining neurons are localized in the first three ganglia of the ventral nerve cord, immediately anterior with respect to the RCs; they are called E cells and are also present in the sub-esophageal ganglia (Sub EG 14) (Lent, 1982; Glover, 1984). Two other neurons can be found in the Sub EG 1 ganglion, they have a large soma and are called lateral cells (LL) (Lent, 1977; Lent and Dickinson, 1984). Finally a dopamine-containing neuron is located in the anterior segmental root of each ganglion (Kerkut et al., 1967; Ehinger et al., 1968; Rude, 1969; Marsden and Kerkut, 1969). It projects its axon towards the ganglion (Stuart et al., 1974), where it bifurcates in two branches and each of them runs in the ipsilateral connective both anteriorly and posteriorly. These histochemical investigations have also demonstrated that only the RCs project their axons to the periphery, whereas the other serotonincontaining neurons have their axons confined within the CNS of the leech and so they are assumed to be interneurons (Lent and Frazer, 1977; Lent, 1981). Recently, Nusbaum and Kristan (1986) have shown that the axons of two serotonin-containing neurons, cells 61 and 21 (Nusbaum and Kristan, 1982), leave the ganglion via the anterior and the posterior connectives and extend as far as the two nearest ganglia (Fig. 2). In other studies it has been shown that 5-HT reacts specifically with anti-S-HT antibody (Osborne et al., 1982; Stuart et al., 1984) and the S-HT present in the RCs in eccentric dense core vesicles label when the ganglia are incubated with radioactive 5-HT (Coggeshall, 1972). In agreement with these techniques, Leake et al. (1985) have used a specific
pattern,
drawn
from the data of histofluorescence
Retzius cells in the leech
407
Fig. 2. Schematic representation of the serotonin-containing neurons’ (cells 21 and 61) branching pattern and synaptic connections with their homologues of the adjacent ganglia. fluorescent-labelled 5-HT antibody to investigate the distribution of 5-HT-immunoreactive fibres in a variety of peripheral tissues in the medicinal leech. All the tissues examined contain immunoreactive fibres although the density of innervation in the different tissues varies (Leake, 1986); particularly, immunoreactive fibres were found in visceral tissues, such as lateral blood vessels, reproductive organs, nephridial bladder and in the body wall muscles. These findings lend further support to the hypothesis that there is a serotoninergic system in each ganglion of the leech; such a system is composed of the RCs which project their axons to the periphery and other serotonin-containing neurons which are probably devoted to interganghonic connections. ELECTROPHYSIOLOGY
Hagiwara and Morita were the first authors who recorded from the RCs with intracellular microelectrodes (1962) and their findings were followed nearly simultaneously by similar reports from Eckert (1963). These studies allowed study of the many electrophysiological properties of the RCs. These neurons have membrane resting potentials of - 30 to - 60 mV and their action potentials range between 20 and 50 mV. Thus, the action potentials, recorded within the Retzius soma, do not overshoot the electrical zero (Lebedev, 1967). This observation has led to hypothesis that impulses do not actively invade the Retzius cell soma, but are generated at a locus some distance from the soma, which is invaded only by electrotonic spread (West and Lent, 1974); this active locus has not yet been localized. Since the first studies it is known that the two RCs of each ganglion are coupled by an electrotonic non-rectifying junction (Hagiwara and Morita, 1962; Eckert, 1963). This junction maintains the two neurons at a nearly identical membrane potential and, moreover, maintains a nearly synchronous impulse activity, although the more posterior RC usually has a higher impulse frequency and tends therefore to behave as the pacemaker (Lent, 1972). Moreover, Beleslin (1968) noted that oxidative metabolism appears essential to the junctional continuity of RCs; in fact if the oxidative phosphorylation of ADP to ATP is interrupted by 2,4-dinitrophenol, they are uncoupled by an nreversible increase in junctional resistance. Many investigations on the morphology of the RCs have utilized intracellular stains and markers in an
attempt to localize the possible site of their coupling: Procion yellow (Lent, 1972), tritiated amino acids (Globus et al., 1973), cobalt chloride (Sunderland et al., 1974) and horseradish peroxidase (Mason and Leake, 1978). Only investigations made by Injections of Procion yellow and tritiated amino acids gave results: in fact a very fine process could be discerned bridging the initial segment of the axons of the two RCs. As this process is the only site of contact between the two cells, it is reasonable to assume that it is the probable site of electrical coupling, although no definite evidence is so far available. The studies concerning the action potential of the RCs have shown that various ions are involved in the rising and falling phases of the spike (Hagiwara and Morita, 1962; Eckert, 1963; Lent, 1972; West and Lent, 1974). The increases in sodium conductance (gNa+) during the rising phase of the spike and in potassium (gK+) during the falling phase, were considered responsible for the spike for many years (West and Lent, 1974). If sodium ions are completely replaced by lithium ions in the solution bathing the RCs, there are no action potentials; however, if the ion replacement is partial the RCs are still capable to generate impulses. Moreover, the sodium channels, responsible for action potential generation, are unaffected by tetrodotoxin (TTX) and saxitoxin (STX), that are known for their capacity to block selectively such channels (Lent, 1977). More recently, Kleinhaus (1976) has demonstrated that the depolarizing phase of the action potential is mediated by two kinds of voltage-dependent channels for inward current: a fast, tetrodotoxin-sensitive sodium channel and a slower sodium-calcium channel which is not blocked by TTX but by cobalt, manganese or lanthanum (Kleinhaus, 1976; Kleinhaus and Prichard, 1976). Outward ion currents involved in the repolarizing phase of the action potential use different voltage-dependent and calcium-dependent potassium channels (Kleinhaus and Prichard, 1977; Yang and Lent, 1983).
SYNAPTIC INPUTS TO THE RETZIUS CELLS
Central inputs Since the first investigations by Hagiwara and Morita (1962) it was demonstrated that electrical stimulation of the connectives elicits synaptic potentials in the RCs. Later on, it was shown that single
408
MAUROCARRETTA
shock stimulation of the ventral nerve cord activates the RCs transynaptically through ascending and descending pathways (Gerasimov, 1967; Tereshkov and Fomima, 1969; Lent, 1972; Bagnoli and Magni, 1975). It was also postulated that synaptic coupling between intersegmental pathways and RCs was chemical in nature (Lent, 1972). Such a hypothesis was based on the latency of the postsynaptic signals and the use of high Mg*+ and low Ca* + saline. Recent studies have demonstrated that electrical stimulation of the ventral cord of Hirudo medicinalis elicits in both RCs of each segmental ganglion an EPSP which is composed of an early and a late component (Carretta et al., 1981). The early EPSP is electrical in nature since it is unaffected by displacements of the membrane potential and by high Mg*+ saline, whereas the late one is chemical, being reversed in sign by membrane depolarization and suppressed by high Mg*+ saline. Latency measurements show that electrical and chemical EPSP components are mediated by two different pathways characterized by conduction velocities of 0.5 and 0.3 m/set, respectively. Both pathways run in each of the lateral connectives and propagate impulses in an anterior and posterior direction. Collision experiments show that ascending and descending impulses along the pathway mediating the electrical EPSP travel along the same fibres. Separate stimulation of the lateral connectives and selective inactivation of one RC by pronase (Parnas and Bowling, 1977) show that both pathways converge onto each RC. This convergence is a probable mechanism contributing to the synchrony of firing of the RCs. The location of the neuron somata giving origin to the two excitatory pathways and their characteristics are still unknown. Concerning the connections between various neurones and the RCs, it has been shown that the RCs receive an excitatory input, probably polysynaptic, from mechano-sensory neurons (T,P,N) (Mason, personal communication; Carretta, unpublished) and from monoamine-containing neurons. The latter neurons are electrotonically connected with each other and with the RCs (Lent and Frazer, 1977) by non-rectifying electrical junctions. In a work devoted to the initiation of swimming activity, Brodfuehrer and Friesen (1986) have found two bilaterally symmetrical pairs of interneurons located in the subesophageal ganglion and named them Tr 1 and Tr 2. These cells are considered trigger neurons because the swimming episodes elicited by their electrical stimulation outlasts the stimulus duration and the length of the swim episodes is largely independent of the intensity of Tr 1 and Tr 2 stimulation. Tr 1 and Tr 2 axons extend as far posterior as the segmental ganglion 18 of the ventral nerve cord. Tr 1 elicits excitatory input to three groups of segemntal neurons: swim-initiating neurons (cells 204 and 205, Weeks, 1982b), serotonin-containing neurons (cells 61 and 21, Nusbaum and Kristan, 1982) and the RCs. Moreover, all RCs in the subesophageal ganglion are excited directly by Tr 1. On the contrary, Tr 2 stimulation elicits transient inhibition in swim-initiating and serotonin-containing neurons and has little effect on RCs.
Peripheral inputs In 1967 Gerasimov demonstrated that light (photic) and mechanical stimulation of the anterior and posterior suckers activate the RCs producing an EPSP. Laverak (1969) showed the existence of a bidirectional fast conducting system (FCS) activated by photic and mechanical stimulation of the skin but could not find any relation between such system and the activity of the RCs. After these studies, many authors turned their attention to the question of the peripheral inputs to the RCs. Bagnoli and Magni (1975) showed that photic and mechanical stimulation of the anterior and posterior parts of the body wall activated the RCs transynaptically through ascending and descending pathways running along the lateral connectives (Bagnoli et al., 1972, 1973). Moreover, these authors could demonstrate that, although some similarities exist between the FCS and the pathways which activate the RCs, namely bidirectionality of conduction and mode of activation, the response of the RCs to natural stimuli was not mediated by the FCS. Recently, some studies demonstrated that waterwave stimulation activates the sensillar movement receptors (SMR) (Friesen, 1981; Brodfuehrer and Friesen, 1984); the sensilla are also activated by photic stimulation (Kretz et al., 1976). Water-wave stimulation elicits swimming in the intact leech and in semi-intact preparation if serotonin is added to the physiological saline (Young et al., 1981; Kristan and Nusbaum, 1983). On the other hand it is known that intracellular stimulation of the RCs can elicit swimming by releasing serotonin in the blood (Willard, 1981) and that they are activated by mechanosensory inputs (Bagnoli and Magni, 1975), by vibrational and thermal stimuli applied to the prostomium (Lent and Dickinson, 1984). The results of studies performed in our laboratory are in keeping with these findings. In a first study we have shown that electrical stimulation of the segmental roots of each ganglion of Hirudo medicinalis, elicits in both RCs inhibitory and excitatory effects (Carretta et al., 1985). The IPSP and EPSP are chemical in nature, being dependent on the membrane potential and suppressed by high Mg*+. Selective inactivation of one RC by pronase, shows that the responses of the controlateral RC are not due to electrotonic coupling between the two cells, but to synaptic actions impinging upon the membrane of both RCs. The two synaptic potentials appear to be mediated by two sets of fibres with a different threshold to electrical stimulation. Their actions on the RCs appear to be polysynaptic on the basis of central latency. Simultaneous stimulation of two roots shows evidence for occlusion for IPSP and summation for EPSP, confirming the polysynaptic nature of the effects. In a further work (Carretta and Zampolini, 1987) the source of the afferent inputs to the RCs was investigated: in particular, whether the IPSP and EPSP could be mediated by some of the mechanosensory neurons (T,P,N) or by other afferent systems, such as the sensilla. The first possibility gave negative results; in fact there is no relation between the activity
Retzius cells in the leech of T, P or N cells and IPSP in the RCs, whereas intracellular stimulation of these neurons can sometimes elicit an EPSP in the RCs (Carretta, unpublished data). On the contrary, water-wave and photic stimulation of the sensilla elicits synaptic potentials identical to those elicited by electrical stimulation of the segmental roots. Mechanical stimuli produced an IPSP in the RCs when a single-segment preparation is used, only an EPSP when the preparation consists of three or more segments. Such an IPSP is the only inhibitory input to the RCs described up to now. Photic stimulation gives always rise to an EPSP independently of the type of preparation. The impulse discharge elicited in the afferent fibres by the two kinds of stimuli is transmitted along the ventral cord both anteriorly and posteriorly to the stimulated segment. This implies that the afferent impulse excites a common pool of intersegmental neurons in each ganglion, which distribute their discharges to the adjacent ganglia. The evidence for occlusion between cordal and photically elicited volleys indicates that the power shares with the sensillar input, a common pool of interneurons. Some evidence has been produced that the serotonin-containing neurons are a component of the common pool of excitatory interneurons (Carretta and Zampolini, 1987). It has been found that waterwave and photic stimulation of the sensilla elicits excitatory post-synaptic potentials in cells 61 and 21. Such potentials are chemical in nature and precede the EPSP of the RCs. These electrophysiological data are presently being verified by an attempt to label cells 61 and 21 following cobalt injection into the sensilla. Since we have found that the intracellular stimulation of cell 61 elicits an IPSP in the T cells it can be postulated that serotonin-containing neurons are one component of the pool of excitatory interneurons to the RCs and that they are also responsible for the inhibition of other afferent inputs to the RCs, thus allowing the access of only the sensillar inputs to their target neurons.
OUTPUTS
OF THE RETZIUS
CELLS
Several anatomical investigations have demonstrated that the RCs send axon branches to the periphery via all segmental roots (see above section on Anatomy). Recently Leake et al. (1985) have found 5-HT immunoreactive fibres in many peripheral tissues in the leech, using a specific fluorescent 5-HT-antibody. Since the RCs are the only serotoninergic neurons which send their axons out of the ganglion, they are the likely source of such 5-HT immunoreactive fibres. On this anatomical basis three different actions of the RCs have been demonstrated in the leech, Hirudo medicinalis. Lent (1973) showed that mucous release from the skin of leeches is under the control of the pair of RCs of each segmental ganglion. In his work Lent could find a relation between the rate of mucous release and number of RC spikes, but he concluded that the data reported did not allow any decision as to whether such an effect is mediated by a synaptic or neurosecretory process.
409
Later Mason et al. (1979) showed that the RCs have some effect on body wall muscles. Particularly, they found that the electrical stimulation of the RC caused a decrease in basal tension of longitudinal muscles of the ipsilateral body wall and the accelerated relaxation following contractions elicited by the motorneuron of longitudinal muscle (L cell). These effects persisted when the ganglion was bathed in 20mM Mg*+ saline. All these effects outlasted the period of impulse protection by the RCs by several minutes and were mimicked by serotonin, the probable transmitter of the RC. More recently Kristan and Nusbaum (1983) have demonstrated a possible neurosecretory role for the RCs. In fact, intracellular stimulation of a RC produces a release of 5-HT in blood vessels and this promotes swimming activity. The same effect was obtained by the addition of 5-HT to the bathing saline (Willard, 1981). No connections are known to exist between the RCs and other neurons in the central nervous system of the leech in vivo. The connections between the RCs and other neurons have been the subject of several investigations carried out on isolated leech neurons and RCs cultivated in vitro (Fuchs et al., 1981, 1982; Henderson et al., 1983). Particularly, these studies have shown that the RCs make a monosynaptic inhibitory connection with the P cell, and that the transmitter liberated by isolated RCs, serotonin, evokes chloride-dependent inhibitory post-synaptic potentials on the P cell. Some discrepancies remain, however, between the results obtained in vivo and in vitro; the most important of them is the question about the differences of electrical properties of these neurons shown in the two experimental situations. NEUROTRANSMIITERS
The nature of the neurotransmitters which excite and/or inhibit the RCs and of their cellular receptors, has received much attention. Kerkut and Walker (1967) demonstrated that the spontaneous activity of the RCs was increased following the addition of acetylcholine (Ach) to the bathing saline or its iontophoretic application. The response to Ach could be reduced or completely blocked by pretreating the preparation with atropine: such a block was readily reversed. The same authors showed that the spontaneous activity of the RCs decreased following the application of 5-HT; the threshold amount of 5-HT was 100 times lower than that of Ach. Further studies have suggested that the RCs possess both nicotinic excitatory receptors and muscarinic inhibitory receptors (Woodruff et al., 1971). The ionic mechanism associated with the inhibition induced by 5-HT on the RCs has been investigated by Walker and Smith (1973) and found to be predominantly a chloride event. Chloride dependent inhibitory post-synaptic potentials (IPSPs) could be recorded from the RCs and these IPSPs could be reversibly blocked by morphine. Pretreatment of the ganglion with imipramine potentiated the duration of the 5-HT inhibition, suggesting the presence of an uptake system for 5-HT. It was also shown (Kerkut and Walker, 1967) that the RCs are hyperpolarized
MAURO CARRETTA
410
by dopamine (DA) and its mechanism of action was investigated and compared with that of 5-HT (Sunderland et al., 1979). Like 5-HT, it was found that the DA hyperpolarization is a chloride conductance increase event. It was concluded that DA and 5-HT activated the same type of chloride ionophore in the membrane of the RCs. Since it was clear that both 5-HT and DA act via chloride ionophores on RCs, an attempt was made to clarify whether each amine is acting on its own specific receptors or not (Sunderland et al.. 1980; Leake et al., 1980). The evidence so far produced suggests, rather surprisngly, that DA and 5-HT are acting via the same receptors. DlSCUSSlON
Since the first morphological description of the two Kolossale zellen by Gustav Retzius (1891), this pair of neurons has been studied at all levels, ranging from histology to histochemistry, and from electrophysiology to synaptic inputs; so today we have a large amount of data on the functioning of RCs as neuronal entities. Morphologically, we know the whole axonal organization-the projections of the axonal branches to the periphery (Retzius, 1891; Falck et al., 1962; Kerkut et al., 1967; Nicholls and Baylor, 1968; Rude, 1969; Stuart et al., 1974; Lent, 1977; Mason and Leake, 1978; Carretta and Grassi, 1987) and there are some interesting observations about the distribution of their terminal fibres (Leake, 1986). Histochemically, the serotoninergic nature of the RCs has been demonstrated (Falck et al., 1962; Kerkut et al., 1967; Ehinger et al., 1968; Rude, 1969; Marsden and Kerkut, 1969) and the role of 5-HT as their putative neurotransmitter has been firmly established (Leake et al., 1980; Leake et al., 1981; Gardner and Walker, 1982; Henderson, 1983; Leake, 1986). The electrophysiological properties of the RCs have been known since Hagiwara and Morita (1962) recorded from the neuron somata by microelectrodes. Therefore, we know their resting potentials, the characteristics of action potentials and ions responsible for the raising and falling phases, the properties of electrotonic coupling between the two cells (Eckert, 1963; Lebedev, 1967; Beleslin, 1968; Lent, 1972; West and Lent, 1974; Kleinhaus, 1976; Kleinhaus and Prichard, 1977; Yang and Lent, 1983; Leake et al., 1986). The only open question is the localization of the electrotonic junction.
Only now are we arriving at an acceptable picture of the synaptic inputs to the RCs. It is known that the RCs are activated by ascending and descending pathways running along the ventral cord (Hagiwara and Morita, 1962; Gerasimov, 1967; Lent, 1972; Bagnoh and Magni, 1975; Carretta et al., 1981); they also receive excitatory inputs from: (a) mechano-sensory neurons, cells T,P,N (Mason, personal communication; Carretta unpublished data); (b) serotonin-containing neurons (Lent and Frazer, 1977; Lent and Dickinson, 1984; Nusbaum and Kristan, 1986); (c) two pairs of neurons located in the subesophageal ganglion, Tr 1 and Tr 2 (Brodfuehrer and Friesen, 1986a); (d) sensillar receptors activated by both photic and water-wave stimulation (Kretz et al., 1976; Friesen, 1981; Brodfuehrer and Friesen, 1984; Lent and Dickinson, 1984; Carretta and Zampolini, 1987). The RCs also receive an inhibitory input from the sensilla (Carretta et al., 1985; Carretta and Zampolini, 1987); this is the only inhibitory input to these neurons described up to now. From all these data it is possible to make an attempt at a physiological interpretation of the inputs described above. One could speculate that these inputs form two different systems, one converging onto the two RCs of each ganglion and one distributing to several ganglia after processing by the RCs and/or other serotonin-containing neurons. The converging system could be the Tr 1 and Tr 2 system that runs from the sub-esophageal ganglion, throughout the whole ventral nerve cord and provides direct excitatory drive to cells 21, 61 and RCs, as well as to the swim-initiating interneurons (cells 204 and 205) (Brodfuehrer and Friesen, 1986a); such excitatory input to the RCs is chemical in nature. The distributing system could be considered the 21 and 61 cells system with their axonal branches running into both anterior and posterior connectives as far as the two nearest ganglia in each direction (Nusbaum and Kristan, 1986). These cells are electrically coupled to each other, with contralateral homologues and with RCs, even though the electrical coupling between a cell 21 or 61 and any other cell 21, 61 or RC is sufficiently weak that each cell may be considered as functionally independent (Fig. 3). These two systems, a chemical one and an electrical one, have some similarity with the two bidirectional nerve cord systems converging with electrical and
Fig. 3. Diagrammatic representation of the proposed synaptic connections of the Tr 1-Tr 2 and cells 21 and 61 systems
to RCs. Open arrows indicate chemical inhibitory synapses, closed electrical synapses and lines indicate chemical excitatory synapses.
arrows
indicate
Retzius
cells in the leech
chemical synapses onto the Retzius cells of the leech described by Carretta et al. (1981). The two systems have an important role in the swimming activity of the leech and are characterized by a common point of contact: the two Retzius neurons. Another aspect of this question is the sensory inputs to the two systems. Tr 1 and Tr 2 neurons are excited by weak mechanical stimulation of body wall such as by pinching of the skin and this activation of Tr 1 and Tr 2 may be limited to tactile mechano-sensory neurons (Brodfuehrer and Friesen, 1986b). Mechanosensory stimulation of the leech body wall in semi-intact preparations, sufficient to initiate swimming, excited cells 21, 61 and RCs (Nusbaum and Kristan, 1986). Such stimulation also directly activates identified T, P, and N neurons; intracellular stimulation of these neurons mimics the effect of body wall stimulation onto cells 21 and 61. All these effects were found to be polysynaptic. Finally, it has been shown that vibrational stimuli applied to the sensilla excite RCs (Lent and Dickinson, 1986) water-wave and photic stimulation of sensilla elicits synaptic potentials in the RCs and in cells 21 and 61 (Carretta and Zampolini, 1987a, b; Carretta and Zampolini, in press). In the light of all these observations, one is led to speculate that the RCs are part of a complex neuronal system playing a modulatory role in regulating sensory afferent inputs necessary for swimming activity of the leech. Moreover, they are subject to a general neuronal control by subesophageal ganglia and to a local segmental control by sensilla and/or other afferent sensory inputs. The functional significance of this double control in the integrated behaviour of the animal can be envisaged as follows: (a) A local response leading to the adjustment at segmental level of the tension of the longitudinal muscle, which is inhibited by the RCs (Mason et al., 1979). (b) A generalized intersegmental excitation which could cause an inhibition of the longitudinal muscle (by accelerating its relaxation rate) (Mason et al., 1979) brought about by the extensive activation of the RCs. These effects are probably the pre-requisite for the initiation of swimming. Obviously all these responses are expected to interact with each other and/or with other stimuli present in a complex situation, such as that occurring in the natural environemnt. It is interesting to observe that the RCs play their modulatory role in the swimming of the leech by two different actions: one directly on the longitudinal muscle of the body wall (Mason et al., 1979; Leake et al., 1981; Mason and Kristan, 1982) and a generalized one on the whole animal by releasing 5-HT in the blood vessel (Willard, 1981). Since a clear connection between the axon of RCs and serotoninergic receptors on the longitudinal muscle of the leech has not been demonstrated, it is conceivable that this action might also be due to a neuro-secretory effect. It should be pointed out that this model is certainly oversimplified, in so far as it neglects possible interactions of other non-identified inputs and/or outputs.
411
In spite of these short-comings, however, it may provide an unifying neurophysiological explanation of the functional significance of these two “Kolossale Gangiienzellen” discovered last century by Gustav Retzius. Acknowledgements--I
thank Prof. Franc0 Magni both for reading and making helpful suggestions about this manuscript, and Dr Silvarosa Grassi and Dr Mauro Zampolini
for collaboration
in some experiments.
REFERENCES
Bagnoli P. and Magni F. (1975) Synaptic inputs to Retzius cells in the leech. Brain Res. 96, 147-152. Bagnoli P., Brunelli M. and Magni F. (1973) Afferent connections to the fast conduction pathway in the central nervous system of the leech Hirudo medicinalis. Archs Ital. Biol. 111, 58-75. Bagnoli P., Brunelli M., Magni F. and Pellegrino M. (1975) The neurons of the fast conducting system in Hirudo medicinalis: identification and synaptic connections with primary afferent neurons. Archs. Ital. Biol. 113, 2143. Beleslin B. (1977) Lack of electrotonic transmission between Retzius nerve cells in the fifth and sixth free ganglion of horse leech Haemopis sanguisuga. Comp. Biochem. Physiol. S6A, 509-5 12. Brodfuehrer P. D. and Friesen 0. W. (1984) A sensory system initiating swimming activity in the medicinal leech. J. exp. Biol. 108, 341-355. Brodfuehrer P. D. and Friesen 0. W. (1986) Initiation of swimming activity by trigger neurons in the leech subesophageal ganglia. I. Outputs connections of Tr 1 and Tr 2. J. camp. Physiol. 159A, 489-502. Brodfuehrer P. D. and Friesen 0. W. (1986) Intitiation of swimming activity by trigger neurons in the leech subesophageal ganglia. III. Sensory inputs to Tr 1 and Tr 2. J. Comp. Physiol. 159A, 51 l-519. Carretta M., Grassi S. and Magni F. (1981) Two bidirectional nerve cord systems converging with electrical and chemical synapses on the Retzius cells of the leech Hirudo medicinalis. Archs Ital. Biol. 119, l-177. Carretta M., Grassi S. and Magni F. (1985) Synaptic effect elicited in the Retzius ceils of the leech Hirudo medicinalis by stimulation of the segmental roots. Archs Ital. Biol. 123, 227-239. Carretta M. and Grassi S. (1987) Axonal projections of the Retzius cells to the dorsal segmental root in the leech Hirudo medicinalis. Archs Ital. Biol. 125, 3845. Carretta M. and Zampolini M. (1987) Afferent sensillar inputs to the Retzius cells of the leech Hirudo medicinalis. Archs Ital. Biol. 125, 45-57. Carretta M. and Zamoolini M. (19871 Identification of some interneurons impinging upon the Retzius cells in the leech. Riunione primaverile Sot. Ital. Fisiol., Florence. Coggeshall R. E. (1972) Autoradiographic and chemical localization of 5-hydroxytryptamine in identified neurons in the leech. Anat. Rec. 172, 489498. Eckert R. (1963) Electrical interaction of paired ganglion cells in the leech. J. Gen. Physiol. 46, 573-587. Ehinger B., Falck B. and Myhrberg H. E. (1968) Biogenic monoamines in Hirudo medicinalis. Histochemie 15, 14c-149. Falck B., Hillarp N. A., Thieme G. and Torp A. (1962) Fluorescence of catecholamines and related compounds condensed with formaldehyde. Histochem. Cytochem. 10, 348-354. Friesen 0. W. (1981) Physiology of water motion detection in the medicinal leech. J. exp. Biol. 92, 255-275. Fuchs P. A., Nicholls J. G. and Ready D. F. (1981) Membrane properties and selective connexions of identified leech neurons in culture. J. Physiol. 316, 203-225.
MAUROCARREITA
412
Fuchs P. A., Henderson L. P. and Nicholls J. G. (1982) Chemical transmission between individual Retzius and sensory neurones of the leech in culture. J. Physiol. 323, 195-211. Gardner C. R. and Walker R. J. (1982) The roles of putative neurotransmitters and neuromodulators in annelids and related invertebrates. Prog. Neurobiol. 18, 81-120. Gaskell J. F. (1914) The chroafline system of annelids and the relationship of this system to the contractile vascular system in the leech Hirudo medicinalis. A contribution to the comparative physiology of the contractile vascular system and its regulators, the adrenalin secreting system and sympathetic nervous system. Phil. Trans. R. Sot. 205, 153-211.
Gerasimov V. D. (1967) Electrical properties and connections of CNS giant nerve cells of Hirudo medicinalis. In Symposium on Neurobiology of Inveriebrares (Edited bv Salanki J.), pp. 285-292. __ _ Globus H.. Lux H. D. and Schubert P. (1973) Transfer of \ amino acids between neuroglia cells and neurons in the leech ganglion. Exp. Neural 40, 104-l 13. Haaiwara S. and Morita H. (1962) Electronic transmission between two nerve cells in leech ganglion. J. Neurophysiol. I
25, 721-731.
Henderson L. P. (1983) The role of 5hydroxytryptamine as a transmitter between identified leech neurones in culture. J. Physiol. 339, 309-324.
Henderson L. P., Kuffler D. P., Nicholls J. G. and Zhang R. J. (1983) Structural and functional analysis of synaptic transmission between identified leech neurones in culture. J. Physiol. 340, 347-358.
James V. A., Walker R. J. and Wheal H. V. (1980) Structure activity studies on an excitatory receptor for glutamate on leech Retzius neurones. Brit. J. Pharmacol. 68, 711-717.
Kerkut G. A. and Walker R. J. (1967) The action of acetylcholine, dopamine and 5-hydroxytryptamine on the spontaneous activity of the cell of Retzius of the leech Hirudo medicinalis. Brit. J. Pharmacol. Chemorher. 30, 64&654.
Kerkut G. A., Sedden C. B. and Walker R. J. (1967) Cellular localization of monoamines by fluorescence microscopy in Hirudo medicinalis and Lumbricus terrestris. Comp. Biochem. Physiol. 21, 687490.
Kleinhaus A. L. (1976) Divalent cations and the action potential of leech Retzius cells. Pfliigers Arch. 363, 97-104.
Kleinhaus A. L. and Prichard J. W. (1976) Sodium dependent tetrodotoxin resistant action potentials in a leech neurone. Brain Res. 102, 368-372. Kleinhaus A. L. and Prichard J. W. (1977) Close relation between Tea response and Ca-dependent membrane phenomena of four identified leech neurones. J. Physiol. 270, 181-194.
Kretz J. R., Stent G. S. and Kristan W. B. Jr (1976) Photosensory input pathways in the medicinal leech. J. camp. Physiol. 106, 1-37. Kristan W. B. Jr and Nusbaum M. P. (1983) The dual role of serotonin in leech swimming. J. Physiol. Paris 78, 743-747.
Leake L. D. (1986) Leech Retzius cells and 5-hydroxytryptamine. Comp. Biochem. Physiol. 83C, 229-239. Leake L. D., Sunderland A. J. and Walker R. J. (1980) Is there a common receptor mediating the inhibitory actions of 5 HT and dopamine on leech Retzius cells? Brit. J. Pharmac. 70, 139-149.
Leake L. D., Mason A. J. and Sunderland A. J. (1981) Is S-hydroxytryptamine a nueromuscular transmitter in the leech? Adr. Physiol. Sci. 22, 391406. Leake L. D., Griffith S. G. and Bumstock G. (1985) 5-hydroxytryptamine-like-immunoreactivity in the peripheral and central nervous system of the leech Hirudo medicinalis. Cell. Tiss. Res. 239, 123-130.
Lebedev A. N. (1967) Spontaneous impulse potentials of single neurons in namisolated nerve ganglia of Hirudo medicinalis. Fiziol. Zh. SSSR Sechenov. 53, 915-921.
Lent C. M. (1972) Electrophysiology of Retzius cells of segmental ganglia in the horse leech Haemopis marmorata. Comp. Biochem. Physiol. 42, 857-886.
Lent C. M. (1973) Retzius cells: neuroeffectors controlling mucous release by the leech. Science 179, 693496. Lent C. M. (1977) The Retzius cells within the central nervous system of leeches. Prog. Neurobiol. 8, 81-l 17. Lent C. M. (1981) Morphology of neurons containing monoamines within leech segmental ganglia. J. exp. Zool. 216, 31 l-316.
Lent C. M. and Frazer B. M. (1977) Connectivity of the monoamine containing neurons in central nervous system of leech. Nature, Lond. 266, 844-846. Lent C. M. and Dickinson M. H. (1984) Serotonin integrates the feeding behavior of the medicinal leech. J. camp. Physiol. 154, 457471. Mason A. and Leake L. D. (1978) Morphology of leech Retzius cells demonstrated by intracellular injection of horseradish peroxidase. Comp. Biochem. Physiol. 61A, 213-216.
Mason A. and Kristan W. B. Jr (1982) Neuronal excitation, inhibition and modulation of leech longitudinal muscle. J. camp. Physiol. 146, 5277536. Mason A., Sunderland A. J. and Leake L. D. (1979) Effects of leech Retzius cells on body wall muscles. Comp. Biochem. Physiol. 63, 359-361,
Marsden C. A. and Kerkut G. A. (1969) Fluorescence microscopy of the 5-HT and catecholamine containing cells in the central nervous system of the leech Hirudo medicinalis. Comp. Biochem. Physiol. 31, 85 1-862.
Nicholls J. G. and Baylor D. A. (1968) Specific modalities and receptive fields of sensory neurons in CNS of the leech. J. Neurophysiol. 16, 740-756. Nusbaum M. B. and Kristan W. B. Jr (1982) The swiminitiating ability of intersegmental serotonin-containing leech interneurons. Neurosci. Abstr. 8, 161-168. Nusbaum M. B. and Kristan W. B. Jr (1986) Swim initiation in the leech by serotonin containing intemeurons cells 21 and cells 61. J. exp. Biol. 122, 277-302. Osborne N. N., Pate1 S. and Dockray G. (1982) Immunohistochemical demonstration of peptides, serotonin and dopamine B hydroxylase like material in the nervous system of the leech Hirudo medicinalis. Histochemistry 75, 573-583.
Parnas I. and Bowling D. (1977) Killing of single neurons by intracellular injection of proteolytic enzymes. Nature, Lond. 270, 626628.
Poll H. and Sommer A. (1903) Ueber phacchrome Zellin in centranervesystem des Blutegels. Verk. Physiol. Ges. Berl. 10, 549-550. Retzius G. (1891) Zur Kenntnis des centralen Nervensystem der Wurmer, das Nervensystem der Annulaten. Biol. Untersuch. (N:F:) 2, l-28. Rude S. (1969) Monoamine containing neurons in the central nervous system and peripheral nerves of the leech Hirudo medicinalis. J. camp. Neurol. 136. 349-372. Rude S., Coggeshall R. E.-and Van Order L. S. (1969) Chemical and ultrastructural identification of 5-hydroxytryptamine in an identified neuron. J. Cell Biol. 41, 832-854. Stuart A. E., Hudspeth A. J. and Hall Z. W. (1974) Vital staining of specific monoamine containing cells in the leech nervous
system.
Cell. Tiss. Res. 153, 55-61.
Stuart A. E., Glover J. C., Blair S. S. and Weisblat D. A. (1984) Development of leech serotonin neurons examined with serotonin antibody cell lineage tracer and cell killing. Sot. Neurosci. Abstr. i0, 605%X)5. Sunderland A. J.. Smith P. A.. Ledke L. D. and Walker R. J. (1974) Neuronal geometry of Retzius cells in Hirudo medicinalis. Experientia 30, 918-919.
Retzius
cells in the leech
Sunderland A. J., Leake L. D. and Walker R. J. (1979) The ionic mechanism of the dopamine response in Retzius cells of two leech species, Hirudo and Haemopis. Comp. Biochem. Physiol. 63C, 129-l 33. Sunderland A. J., Leake L. D. and Walker R. J. (1980) Evidence for an amine receptor on the Retzius cells of the leeches Hirudo and Haemopis. Comp. Biochem. Physiol. 63C, 129-133. Tereshkov 0. D. and Fomina M. S. (1969/70) Electrophysiological investigation of the system of paired giant cells in the ventral chain of Autostoma gulo. Neurosc. Transl. 12, l&16. Yang J. and Lent C. M. (1981) Calcium depletion produces Na+-dependent sustained depolarization of Retzius cells membranes in the leech CNS. J. camp. Physiol. 144, 111L116. Walker R. J. and Smith P. A. (1973) Ionic mechanism for S-hydroxytryptamine inhibition on Retzius cells of the leech Hirudo medicinalis. Comp. Biochem. Physiol. 45, 979-99 I. Weeks J. C. (1981) Neuronal basis of leech swimming separation of swim initiation pattern generation and
413
intersegmental coordination by selective lesions. J. Neurophysiol. 45, 698-724. Weeks J. C. (1982) Segmental specialization of a leech swim initiating interneuron cell 205. J. Neuroscience 2, 972-986. Weeks J. C. (1982) Synaptic basis of swim initiation in the leech. II: A pattern generating neuron (cell 208) which mediates motor effects of swim initiating neurons. J. camp. Physiol. 148, 265-279. Weeks J. C. and Kristan W. B. Jr (1978) Initiation, maintenance and modulation of swimming in the medicinal leech by the activity of a single neuron. J. exp. Biol. 77, 71-88. West C. H. K. and Lent C. M. (1974) Effects of temperature on the electrophysiology of Retzius cells in the leech. Comp. Biochem. Physiol. 47A, 27-38. Willard A. L. (1981) Effects of serotonin on the generation of the motor program for swimming by the medicinal leech. J. Neurosci. 1, 936-944. Woodruff G. N., Walker R. J. and Newton L. C. (1971) The action of same muscarinic and nicotinic agonists on the Retzius cells of the leech. Comp. gen. Pharmac. 2, 106117.