Cholinergic Action on the Heart of the Leech, Hirudo Medicinalis

The longitudinal muscle of the leech body wall is very sensitive to superfused acetylcholine (ACh) in the presence of an anticholinesterase such as eserine (Bacq & Coppee, 1937; Flacke & Yeoh, 1968a). Superfused (Walker, Woodruff & Kerkut, 1970) and iontophoretically applied (Kuffler, 1978) ACh depolarizes the muscle, and these responses are reversibly blocked by curare (Walker et al. 1970; Kuffler, 1978). More recently, significant amounts of choline acetyltransferase (CAT) and acetyl cholinesterase (AChE) have been detected in the cell bodies of several motor neurones, including those that innervate the longitudinal muscles (Wallace, 1981a,b; Wallace & Gillon, 1982). These observations, taken together, strongly implicate ACh as an excitatory neuromuscular transmitter in the leech. Nevertheless, the pharmacological properties of the muscular acetylcholine receptors (Flacke & Yeoh, 1968a,b) and the nature of their ionic channels remain relatively unexplored.

The leech’s paired hearts are tube-like vessels that extend the length of this segmented worm (Mann, 1962). The walls of these vessels are made up of spiral muscle cells that receive direct innervation from two identified classes of neurones, the HE motor neurones and the HA modulatory neurones (Maranto & Calabrese, 1984a) (Fig. 1). Pairs of HE motor neurones occur in nearly every segmental ganglion of the nerve cord (Thompson & Stent, 1976) ; each HE cell innervates the ipsilateral heart in its own segment. Pairs of HA modulatory neurones occur only in the fifth and sixth segmental ganglia (Calabrese & Maranto, 1984); each HA cell innervates the contralateral heart in several segments. The neuromuscular contacts of both types of neurones have been described at the ultrastructural level (Hammersen & Staudte, 1969; Maranto & Calabrese, 1984a) and are similar to those described for the longitudinal muscles (Rosenbluth, 1973; Yaksta-Sauerland & Coggeshall, 1973) and the muscles in the central nervous system (Tulsi & Coggeshall, 1971).

The cell bodies of the HE motor neurones contain both CAT and AChE (Wallace, 1981a,b; Wallace & Gillon, 1982) and action potentials in these motor neurones elicit excitatory junctional potentials in heart muscle cells (Thompson & Stent, 1976; Maranto & Calabrese, 1984a). These observations suggest that there are cholinergic receptors on leech heart muscle cells associated with the HE cell’s neuromuscular contacts. HE motor neurones also contain an FMRFamide-like peptide (Kuhlman, Li & Calabrese, 1985a). Thus, the HE motor neurones may use both ACh and an FMRFamide-like peptide as transmitters.

The HA modulatory neurones also contain an FMRFamide-like peptide (Kuhlman et al. 1985a), and their activity modulates the myogenic properties and beat tension of the heart without apparent junctional potentials (Maranto & Calabrese, 1984a; Calabrese & Maranto, 1984). Superfused FMRFamide mimics these modulatory actions on the heart (Kuhlman et al. 1985b). Thus, like the HE motor neurones, the HA modulatory neurones may use an FMRFamide-like peptide as a transmitter. Since HA cells have not been assayed for CAT or AChE, it is not clear whether they use ACh as a transmitter.

In this study we localize acetylcholinesterase in the heart, characterize cholinergic receptors on the heart muscle cells and test neuromuscular transmission by the HE motor neurones and HA modulatory neurones in the presence of curare, a potent antagonist of those receptors.

Leeches, Hirudo medicinalis, were purchased from commercial suppliers and maintained in artificial pond water at 15 °C for up to 4 months before use.

Enzymatically isolated heart muscle cells, isolated hearts and innervated hearts were prepared and maintained as described previously (Maranto & Calabrese, 1984a,b; Calabrese & Maranto, 1984). Intracellular recordings from these preparations were also made as previously described (Maranto & Calabrese, 1984b; Calabrese & Maranto, 1984).

The original method of Karnovsky & Roots (1964) or a modification of this method for the leech (Wallace & Gillon, 1982) was used to stain for acetyl-cholinesterase activity. Only tissues stained by the first method were suitable for ultrastructural examination, because the other method uses Triton X-100 in the reaction medium. Pseudocholinesterases were blocked with 0·03 % ethopropazine in the first method and with 0·1 mmol 1⁻¹ phospholine iodide in the second. Controls for either staining method consisted of tissue incubated in a reaction mixture supplemented with 1 mmol 1⁻¹ eserine sulphate to block acetylcholinesterase, or tissue incubated in a reaction mixture lacking the substrate (acetylthiocholine iodide). Fixation and staining procedures were those described by the authors for each method. After treatment the tissue was prepared for light or electron microscopy as previously described (Maranto & Calabrese, 1984a). Criteria for the identification of neuromuscular junctions were those of Hammersen & Staudte (1969) and Maranto & Calabrese (1984a). At neuromuscular junctions axon terminals contained clear synaptic vesicles approximately 50 nm in diameter and large dense-core vesicles 80–120 nm in diameter. The junctional cleft was occupied by the basement membrane and was 50–60 nm thick. The post junctional membrane often had filamentous material associated with it.

Pharmacological agents to be applied to heart muscle preparations were dissolved in cold leech saline prior to each experiment and allowed to equilibrate to room temperature. The pH of each drug solution was then readjusted to the original pH of the saline. Drug solutions and normal saline were then alternately superfused over the preparation at flow rates of 2–6 ml min⁻¹.

Acetylcholine was microiontophoretically applied according to the method of Curtis & Eccles (1958). Short-shanked (10 mm) micropipettes were back-filled with a 3 mol 1⁻¹ solution of acetylcholine chloride in distilled water. The micropipette was positioned as close as possible to the cell without penetrating it, and acetylcholine was ejected with positive current supplied by a WPI microiontophoresis unit. The dose (measured in nanoCoulombs, nC) was controlled by altering the duration or amplitude of the current pulse. A negative backing current of up to 40 nA was applied to prevent leakage of acetylcholine from the tip of the pipette; the backing current was adjusted to the minimum current that permitted the electrode to approach the recorded muscle cell without depolarizing it.

In a few experiments drugs were applied from the tip of a micropipette with pressure. Drugs dissolved in saline were ejected with 1-s (timed manually) pulses of pressure applied to the back of the pipette.

Alpha-bungarotoxin was purchased from Biotoxins, Inc., St Cloud, Florida, USA.

The histochemical methods of either Karnovsky & Roots (1964) or Wallace & Gillon (1982) indicated the presence of AChE in nerves and endplates of the heart (Fig. 2A). An interconnecting network of stained nerves was present over the entire length of the stained tissue. The stained endplates were small (approximately 5 μm in diameter), circular and more densely stained than the nerves. All staining was blocked by adding eserine to the reaction medium, or removing acetylthiocholine.

In preparations stained for AChE, axon terminals contained clusters of clear synaptic vesicles approximately 50 μm in diameter and large, dense-core vesicles 80–120 nm in diameter (Fig. 2B). Glycogen particles and mitochondria were also present. The axon terminal was surrounded by a glial covering except at the region of the synaptic cleft. The junctional cleft was occupied by a basement membrane and was approximately 60 nm thick. In all respects such terminals were indistinguishable from the terminals of the HE motor neurones identified by intracellular labelling with horseradish peroxidase (HRP) (Maranto & Calabrese, 1984a). The granular quality of the AChE reaction product observed here was similar in appearance, though less densely packed, than that observed in vertebrates (Karnovsky & Roots, 1964; McMahan, Sanes & Marshall, 1978). Particles of AChE stain were often found in the cleft of the neuromuscular junction in these preparations, just as they are at vertebrate neuromuscular junctions (Karnovsky & Roots, 1964; McMahan et al. 1978). Control preparations contained no such particles and were indistinguishable from those previously reported by Hammersen & Staudte (1969) and ourselves (Maranto & Calabrese, 1984a). Intracellular AChE was not observed in nerve terminals, although a considerable amount of stain was observed in larger axons running along the heart (Fig. 2C).

Although many isolated heart muscle cells could express a polarization rhythm (Maranto & Calabrese, 1984b), many cells were initially quiescent or became quiescent during intracellular recording. Quiescent cells had resting potentials of approximately −40 mV and input resistances of 50–150 MΩ. These cells responded to iontophoretic application of acetylcholine (ACh) with a dose-dependent depolarization (Fig. 3). Maximum responses were 20–30 mV in amplitude and these large depolarizations often triggered action potentials (Fig. 3A). The rise time to the peak response was 300–350 ms for iontophoretic pulses varying in duration from 10 to 30 ms. To eliminate conductance changes associated with action potentials, iontophoretic pulses of ACh were adjusted to produce subthreshold depolarizations in succeeding experiments. Most of the thick central portion of the cells seemed sensitive to ACh without well-defined hot spots.

To investigate whether ACh opened ion channels, positive and negative current pulses (5 s in duration) of equally spaced amplitudes were injected into a cell to alter membrane potential. One second after the start of each current pulse, a pulse of ACh was applied from an iontophoretic pipette. The arrangement of voltage, current and ACh pipettes is illustrated in Fig. 4A. Fig. 4B shows superimposed records from a typical experiment in which ACh was applied during seven hyperpolarizing and seven depolarizing current pulses. At more negative membrane potentials the iontophoretic responses became larger, while at more positive membrane potentials they became smaller.

We tested the effect of membrane potential on the amplitudes of the iontophoretic responses to ACh in eight cells. In each case, although the amplitude of the iontophoretic response varied with membrane potential as shown (Fig. 4B), we were unable to reverse the iontophoretic responses with the currents injected. We estimated the reversal potential by extrapolating the data to the abscissa. The mean reversal potential thus determined was −9 mV (range −12mV to −5 mV), approximately 30 mV more positive than resting potential. Graphic analysis of the data from one cell is shown in Fig. 4D. Although the cell’s input resistance changed over this range of membrane potential (Fig. 4C), the resistance change can only account for a fraction of the amplitude change in the iontophoretic response. At the largest hyperpolarization the cell’s input resistance was approximately three times the input resistance at the largest depolarization, yet the amplitude of the iontophoretic response was more than ten times the amplitude of the response at the largest depolarization. Changes in the cell’s input resistance would, of course, reduce the accuracy of the extrapolation, but the data clearly indicate that ACh acts by causing a conductance increase with an equilibrium potential near 0 mV. Assuming that the ionic equilibrium potentials for these cells resemble those of other muscle cells, a reversal potential near 0 mV suggests that ACh increases conductance to both sodium and potassium ions.

We tested the effect of the cholinergic antagonist curare (d-tubocurarine) on the iontophoretic responses of isolated heart muscle cells to ACh (Fig. 5). Superfusion of the cell culture with 10⁻⁴moll⁻¹ curare reversibly reduced the amplitude of the iontophoretic response at all doses tested. Higher concentrations of curare completely blocked the iontophoretic response.

To examine the responses of heart muscle cells to a variety of cholinergic agents including agonists, antagonists and cholinesterase inhibitors, intracellular recordings were made from muscle cells in an isolated heart while these agents were superfused over the preparation. Polarization rhythms observed in these preparations usually had interburst intervals in excess of 30 s, and thus did not interfere with our measurements of membrane potential. Membrane potential was constant during the interburst intervals and was approximately −40 mV. Input resistance of the cells was approximately 4MΩ. This low input resistance probably reflects the extensive electrical coupling between cells in the intact heart (Maranto & Calabrese, 1984a).

Muscle cells were rapidly depolarized when ACh was applied at concentrations as low as 1 μmoll⁻¹ in normal saline. Maximum responses at high ACh concentrations were between 20 and 30 mV in amplitude. Such large depolarizations normally did not elicit action potentials in the intact heart, as they did in isolated muscle cells. When the superfused saline contained 50 μmol 1⁻¹ eserine sulphate, muscle cells depolarized in response to ACh concentrations as low as 10 nmol 1⁻¹. This enhanced sensitivity was probably caused by inhibition of the above described acetylcholinesterase by eserine. Dose-response data for four separate experiments are plotted in Fig. 6. The sensitivity of the heart muscle to ACh is similar to the sensitivity of leech longitudinal muscle reported by others (Bacq & Coppee, 1937; Walker et al. 1968; Flacke & Yeoh, 1968a; Kuffler, 1978). Although the heart became more sensitive to ACh in the presence of eserine, the slope of the dose-response curve declined; receptor desensitization at high ACh concentrations in the absence of active AChE could account for this decrease in slope. (See the discussion of the action of other anticholinesterases below.)

Iontophoretic application of ACh to muscle cells in isolated hearts gave dose -dependent responses (Fig. 7A) similar to those observed in isolated muscle cells, except that no action potentials were elicited even at large doses. The data of Fig. 7A yield a dose-response curve that is linear over a broad dosage range (Fig. 7B), but begins to saturate with large iontophoretic pulses. The rise time to the peak response was l·2–1·3s for iontophoretic pulses varying in duration from 40 to 300 ms.

The ability to record in heart muscle cells stable, dose-dependent responses to ACh applied either by superfusion (Fig. 6) or by iontophoresis (Fig. 7) permitted us to use the isolated heart preparation to study in more detail the pharmacology of the ACh receptor. Cholinergic agents were superfused over the heart and tentatively classified as: (1) agonists, if they depolarized muscle cells; (2) antagonists, if they reduced the size of iontophoretic responses to ACh without affecting resting potential or resting input resistance; (3) AChE inhibitors, if they increased the size and duration of iontophoretic responses to ACh (Table 1). In each case the reported responses were reversed by superfusion with normal saline.

The agonistic effects of carbachol (carbamylcholine), DMPP (1,1-dimethyl-4-phenylpiperazinium), nicotine and TMA (tetramethylammonium) suggest the presence of a nicotinic receptor similar to the type found in vertebrate autonomic ganglia. The ability of curare (d-tubocurarine) (Fig. 8), decamethonium and hexa-methonium to block iontophoretic responses supports this notion (Chen, Portman & Wickel, 1951; Ginsborg & Guerrero, 1964), although the inhibitory effect of nicotine at lower concentrations (Fig. 8) was inconsistent. Muscarinic agonists, methacholine and pilocarpine, and the muscarinic antagonist, scopolamine, had no effects even at millimolar concentrations. Atropine (another muscarinic antagonist) at reasonably low concentrations (50–100 μmol 1⁻¹) did block iontophoretic responses to ACh (Fig. 8), while 1 mmol l⁻¹ atropine depolarized the muscle.

As expected, neostigmine blocked AChE at the heart but also appeared to antagonize the iontophoretic response to ACh. Decamethonium and hexamethonium produced similar effects. The dual effects were observed as an initial augmentation, followed by a decrement, of the iontophoretic response to ACh. The apparent antagonistic action of these compounds on the heart ACh receptor might result from desensitization of the receptor in the absence of active AChE.

Three antagonists, atropine, nicotine and curare, were singled out for further study because they were free from other effects at low concentrations. We measured the relative potency of these antagonists by determining the concentration of the antagonist which inhibited 50 % of iontophoretic responses from the middle of the linear portion of the ACh dose-response curve ascertained for each muscle cell recorded (Fig. 8). Curare was found to be the most potent antagonist, inhibiting 50% of the iontophoretic responses at approximately 0·5μmoll⁻¹. Nicotine and atropine blocked 50% of the response at 50 and 100μmol l⁻¹, respectively.

Alpha-bungarotoxin, even at a concentration of lμmoll⁻¹, did not antagonize the iontophoretic response of heart muscle cells to ACh. At higher concentrations the toxin reversibly depolarized heart muscle cells. To determine if this effect was mediated through the ACh receptors, the ability of curare to block the response to toxin was tested. Alpha-bungarotoxin was applied to a muscle cell in the heart by ejecting it from a micropipette with pressure. The concentration of toxin in the pipette was 1 mmol 1⁻¹. Iontophoretic responses to ACh were elicited as usual. Pulses of ACh and alpha-bungarotoxin depolarized the heart rapidly and reversibly (Fig. 9). When the heart was superfused with 5 μmol l⁻¹ curare, the responses to both compounds were attenuated. The antagonistic effects of curare on both the ACh- and toxin-induced depolarizations were reversed by superfusion with normal saline.

Since curare was a potent antagonist of the acetylcholine receptor described above, we used curare to determine if either the HE motor neurones or the HA modulatory neurones engaged such receptors in neuromuscular transmission. We used a preparation in which a single segment of a heart was left innervated by its segmental ganglion. Previous studies have shown that in such preparations the heart often expresses a myogenic rhythm, and that impulses in HE motor neurones produce excitatory junctional potentials in heart muscle cells (Maranto & Calabrese, 1984a). The individual junctional potentials are small, however, and are not usually resolvable, but bursts of impulses in HE motor neurones do cause rapid summated junctional potentials that clearly reset the heart’s myogenic rhythm (Maranto & Calabrese, 1984b) (Fig. 10A). Impulses in HA modulatory neurones do not cause measurable junctional potentials in heart muscle cells (Maranto & Calabrese, 1984a), but their activity can accelerate the myogenic rhythm of the heart or induce it in quiescent preparations (Calabrese & Maranto, 1984). In the present study, curare (10 ^{− 4}mol l ^{− 1}) reversibly blocked the ability of the HE motor neurone to depolarize the heart muscle rapidly and reset its myogenic rhythm (Fig. 10B,C), but did not substantially interfere with the ability of the HA modulatory neurone to accelerate the heart’s myogenic rhythm (Fig. 11).

To examine the sensitivity of the heart muscle cells to other putative neurotransmitters, intracellular recordings were made from cells in the isolated heart while the compounds were superfused over the preparation. At concentrations up to 1 mmol 1 ^{− 1}, the following compounds had no effect on membrane potential: adenosine triphosphate, aspartate, carnosine, dopamine, epinephrine, gamma-amino butyric acid, glutamate, glycine, histamine, norepinephrine and octopamine.

Serotonin, applied at 1 mmol 1 ^{− 1}, depolarized muscle cells by about 5 mV. No membrane potential changes were observed when 0 · 1 mmol 1 ^{− 1} serotonin was applied. Walker et al. (1968) have reported that 50mol l ^{− 1} serotonin reduces the size of excitatory junctional potentials and iontophoretic responses to ACh in leech longitudinal muscle. When we tested the effect of 0 · 1 mmol 1 ⁻¹ serotonin on the iontophoretic response to ACh, we observed no changes in amplitude or duration.

Superfusion and iontophoretic application of ACh to isolated heart muscle cells or muscle cells in isolated hearts demonstrated the presence of acetylcholine receptors. These receptors are linked to channels which open in the presence of ACh; the opening of these channels increases the membrane’s conductance apparently to both sodium and potassium ions, since the reversal potential of the depolarizing response is near 0 mV. Because these receptors were not localizable to distinct hot spots in iontophoretic experiments, it is likely that they represent both junctional and extra junctional receptors.

Pharmacologically, these receptors are similar to those described in the molluscan central nervous system (Kehoe, 1972), the crustacean central nervous system and muscle (Marder & Paupardin-Tritsch, 1978, 1980) and in vertebrate autonomic ganglia (Chen et al. 1951; Ginsborg & Guerrero, 1964). The features common to these acetylcholine receptors are: (1) they mediate depolarizing responses to ACh; (2) carbamylcholine-, nicotine, TMA and DMPP mimic the effects of ACh; (3) curare, decamethonium and hexamethonium are antagonists of ACh; (4) muscarinic agonists and antagonists such as methacholine and pilocarpine have either weak or no effects on these receptors.

The failure of alpha-bungarotoxin to block ACh-mediated depolarizations in leech heart muscle is also characteristic of this class of nicotinic receptors in sympathetic ganglia (Brown & Fumagalli, 1977), crustacean muscle (Marder & Paupardin-Tritsch, 1980) and the molluscan central nervous system (Kehoe, 1972).

Hassoni, Kerkut & Walker (1985) have characterized an ACh receptor on the body wall muscle of the earthworm, Lumbricus, that is sensitive to alpha bungarotoxin. This receptor is similar pharmacologically both to the receptor on leech body wall muscle (Flacke & Yeoh, 1968a,b) and to the receptor described here. The sensitivity of the leech body wall receptor to alpha-bungarotoxin has not been determined: either leech and earthworm muscular ACh receptors differ in their sensitivity to alpha-bungarotoxin, or leech cardiac muscle has an ACh receptor distinct from that on leech body wall muscle. Neural nicotinic acetylcholine receptors in insects both bind alpha-bungarotoxin and are irreversibly blocked by it (Harrow & Sattelle, 1983; Sattelle et al. 1983).

A depolarizing response to alpha-bungarotoxin, such as the one observed here with high toxin concentrations (1 mmol 1 ^{− 1}), has not been previously reported. This response appears to be mediated through acetylcholine receptors, because it is sensitive to curare, but the possibility that it is caused by a contaminant in the toxin preparation cannot be eliminated.

We have demonstrated the presence of acetylcholinesterase associated with nerves and nerve terminals on leech heart muscle using histochemical stains (Karnovsky & Roots, 1964; Wallace & Gillon, 1982). Ultrastructurally, this staining was localized inside axon branches coursing near the heart and to the vicinity of neuromuscular junctions. The sensitivity of the heart muscle to superfused ACh is greatly enhanced by anticholinesterases, such as eserine sulphate. This observation indicates that the cholinesterase may rid heart neuromuscular junctions of released ACh.

Previous studies have demonstrated that the leech’s hearts are innervated by two classes of efferent neurones, heart motor neurones (HE cells) and heart modulatory neurones (HA cells) (Maranto & Calabrese, 1984a). Since the HE motor neurones contain both CAT and AChE activity, they are probably cholinergic (Wallace, 1981a,b; Wallace & Gillon, 1982).

The results presented here demonstrate that there are cholinergic receptors on heart muscles cells which could subserve neuromuscular transmission from HE motor neurones. These receptors were sensitive to curare and the HE cell’s transmission to heart muscle was blocked by curare. These results provide further evidence that HE motor neurones, like other leech motor neurones (Walker et al. 1970; Kuffler, 1978), use ACh as a neuromuscular transmitter. HE motor neurones also contain an FMRFamide-like peptide (Kuhlman et al. 1985a); therefore, they possibly use more than one transmitter. Since FMRFamide is active on the heart (Kuhlman et al. 1985b), a slower peptidergic component of the HE cell’s transmission might remain when cholinergic transmission’s blocked by curare. We intend to use curare to explore this possibility.

The HA neurones have not been assayed for cholinergic enzymes, but they too contain an FMRFamide-like peptide (Kuhlman et al. 1985a). We have demonstrated here that the HA cell’s action on the heart is insensitive to curare, while previous work has shown that the HA cell’s effects on the heart muscle are mimicked by superfusion of micromolar amounts of FMRFamide (Kuhlman et al. 1985b). It is, therefore, likely that HA cells are peptidergic but not cholinergic.

We thank Dr Eve Marder for much helpful advice and Christine S. Cozzens for valuable editorial assistance. This research was supported by NIH grant 1 RO1 NS21232 to RLC.