Topographical and electrophysiological characteristics of highly excitable S neurones in the myenteric plexus of the guinea-pig ileum
Terence K. Smith,
Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, NV 89557-0046, USA
Corresponding author T. K. Smith: Department of Physiology and Cell Biology, Anderson Medical Sciences Building/352, University of Nevada School of Medicine, Reno, NV 89557-0046, USA. Email: firstname.lastname@example.org
1Most intracellular electrical recordings from myenteric neurones have been made from the centre of large ganglia. In this study, we examined the electrophysiological properties of neurones at the corners of large ganglia close to internodal strands and in microganglia.
2Of 150 neurones in these locations: 111 were tonic S neurones; 9 were phasic S neurones and 30 were AH neurones.
3Tonic S neurones were characterized by: (i) low resting membrane potentials (−50 ± 1 mV, mean ± s.e.m.); (ii) high input impedance (522 ± 23 MΩ); (iii) low threshold for action potential (AP) generation (0.012 ± 0.004 nA); (iv) firing of APs throughout a depolarizing pulse (duration ≤ 1 s) and one to four APs following a hyperpolarizing pulse and (v) spontaneous fast excitatory postsynaptic potentials (FEPSPs). A substantial proportion of tonic S neurones (43 %) also fired APs spontaneously (7.6 ± 0.6 Hz; range, 0.3–19 Hz). All APs were blocked by tetrodotoxin (1 μm).
4Tonic S neurones were subclassified, according to their post-stimulus responses, as SAH or SAD neurones. Following a burst of APs, SAH neurones exhibited a prominent after-hyperpolarization (duration, 711 ± 10 ms) and SAD neurones an after-depolarization (duration, 170 ± 10 ms). The after-hyperpolarization was reduced in four of ten neurones by apamin (0.3 μm).
5FEPSPs were evoked in 20 of 38 S neurones by electrical stimulation applied both oral and anal to the recording site. Repetitive stimuli evoked slow excitatory postsynaptic potentials (SEPSPs) in some tonic S neurones.
6Three functional classes of S neurones were identified after injection of neurobiotin through the recording microelectrode: (i) longitudinal muscle motor neurones, (ii) short circular muscle motor neurones, and (iii) ascending interneurones.
7In conclusion, there appears to be topographical organization of highly excitable, tonic S neurones within the myenteric plexus, since, in contrast to other S neurones, they can be readily impaled in myenteric ganglia close to internodal strands and in microganglia.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Myenteric neurones are generally classified electrophysiologically into AH/type II and S/type I neurones (Nishi & North, 1973; Hirst et al. 1974). Action potentials in AH neurones are calcium dependent. It is the influx of calcium with the action potential which leads to the opening of calcium-dependent potassium channels that are responsible for the long after-hyperpolarization (AH) which characterizes these neurones (Nishi & North, 1973; Hirst et al. 1974). The calcium-activated potassium current renders the soma refractory to further stimuli (Nishi & North, 1973; Hirst et al. 1974). In contrast, action potentials in S neurones are dependent on Na+ influx since they are completely blocked by tetrodotoxin (Nishi & North, 1973; Hirst et al. 1974; Shuttleworth & Smith, 1999). Fast excitatory postsynaptic potentials (FEPSPs) are readily evoked in S neurones (S for synaptic input), but rarely in AH neurones (Nishi & North, 1973; Hirst et al. 1974, 1975). Prominent slow excitatory postsynaptic potentials (SEPSPs), however, can be readily generated in AH neurones by repetitive nerve stimulation (see Wood, 1994b). AH neurones are usually large multipolar neurones (Dogiel type II morphology) which send processes down to the mucosa and appear to be primary afferent neurones (Furness et al. 1990). They respond to a variety of mucosal stimuli such as serotonin and HCl (Smith, 1994; Kunze et al. 1995). AH neurones communicate with S neurones and other AH neurones mainly via slow excitatory postsynaptic potentials (SEPSPs) (Kunze et al. 1993), whereas, S neurones communicate with other S neurones mainly via FEPSPs (Hirst et al. 1975; Bornstein et al. 1991; Smith et al. 1992). As a group, S neurones are functionally more diverse than AH neurones. They are uniaxonal neurones, which exhibit a variety of dendritic morphologies and soma sizes, and comprise the motor neurones and interneurones within the reflex pathways that supply both the longitudinal and circular muscle layers (Bornstein et al. 1991; Smith et al. 1992; Costa et al. 1996). S neurones, which respond to mucosal stimulation or distension with bursts of FEPSPs, usually lie mainly within either ascending or descending nervous pathways, although some S neurones (e.g. longitudinal muscle motor neurones) are common to both pathways (Smith et al. 1992). In recognition of this wide functional diversity amongst myenteric neurones, some investigators have tried to subdivide them further into several electrophysiological classes (see Wood, 1994a). Under typical recording conditions, most (85–92 %) myenteric S neurones exhibit a phasic or rapidly adapting firing pattern, whereas some others have a slowly adapting firing pattern (reviewed in Bornstein et al. 1994; Kunze et al. 1994; Kunze et al. 1997). Although, Brookes et al. (1997) recently reported that ascending interneurones, identified by 1,1′-didodecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) retrograde labelling techniques, were tonic S neurones.
Early studies using extracellular electrodes have reported that many myenteric neurones fired action potentials spontaneously and that this activity increased with stretch (reviewed in Wood, 1994b). Since spontaneous action potentials were rarely, if ever, observed in myenteric neurones with intracellular microelectrodes it has been suggested that this activity resulted from the mechanical disturbance applied by pressure or suction electrodes to mechanically sensitive somas (North & Williams, 1977). Later Wood & Mayer (1978), using intracellular recording techniques, found a number of myenteric S type neurones that fired action potentials spontaneously. They attributed these differences to the fact that they did not use drugs to paralyse the muscle or because they sampled neurones in most regions of the ganglia. More recently it has been shown that some phasic neurones increase their firing rates in response to stretch applied to the tissue in the absence of drugs to paralyse the muscle (Kunze et al. 1998).
The idea of a topographical organization of myenteric neurones that is related to function, as in the spinal cord or spinal ganglia of invertebrates (see Levitan & Kaczmarek, 1991), is very attractive but in the past has not gained much support. Immunohistochemical and DiI retrograde labelling studies in recent years, however, suggest the possibility that some degree of neuronal organization occurs within myenteric ganglia. Motor neurones that innervate the longitudinal muscle are generally found at the edges of myenteric ganglia near the junction with internodal strands or in microganglia (Bornstein et al. 1991; Brookes et al. 1992). Also, calretinin (a calcium-binding protein) immunoreactive neurones, which include both cholinergic longitudinal muscle motor neurones and ascending interneurones, are often located near the edges of ganglia and are often seen in clusters of two to four cells near the internodal strands (Brookes et al. 1991a).
In this study, we examined the electrophysiological and morphological characteristics of S neurones located in the corners of large ganglia close to where internodal strands emerge from the ganglion, and in microganglia formed at the intersection of several internodal strands. We report that highly excitable tonic S neurones can be readily accessed in these locations.
Sixty male guinea-pigs (Duncan Hartley) in the weight range 200–300 g were killed by exposure to a rising concentration of CO2 gas, as approved by the Animal Ethics Committee of the University of Nevada, and a 3 cm segment of ileum was removed from a region situated 15 cm above the caecum. The lumen of the segment was then flushed with saline with the aid of a syringe inserted into the oral end. The segment of ileum was then cut open along its mesenteric border, and pinned with the mucosa uppermost to the bottom of a Petri dish lined with Sylgard (Dow Corning Corp.). The mucosa, submucosa and the circular muscle were removed by sharp dissection to create a standard longitudinal muscle myenteric plexus (LMMP) preparation, which was then stretched and pinned to the floor of an electrophysiological chamber with the myenteric surface uppermost. The electrophysiological chamber was mounted on the stage of an inverted condensing microscope (Diaphot, Nikon) and continuously perfused with oxygenated Krebs solution (see below for composition).
Myenteric neurones were impaled with standard microelectrodes fabricated from borosilicate glass capillary tubing (1.0 mm o.d., 0.7 mm i.d., 1B100F-4: WPI Corp.) on a Brown-Flaming Puller (Sutter Instruments). The electrodes, which were filled with a solution of 2 % neurobiotin in 1 M KCl, had a resistance of between 200 and 300 MΩ. Electrical activity recorded with the microelectrode was sampled using an Axoprobe-2A dual channel amplifier (Axon Instruments). Depolarizing or hyperpolarizing current was injected using one of the amplifier's bridge circuits, and the pulse duration and frequency regulated using an output from a programmable eight-channel pulse generator (Master-8, Jerusalem, Israel). The intracellular recordings were displayed on a four-channel storage oscilloscope with a built-in chart recorder (1600 series, Gould Electronics) and also recorded on video cassette for off-line analysis (PCM 3000, A. R. Vetter Corp., Rebersburg, PA, USA). Microelectrodes were positioned using a micromanipulator (Narashige M15, Pacer Scientific, Los Angeles, CA, USA) and cells impaled using the variable capacitance compensation of the Axoprobe-2A.
The synaptic inputs onto neurones were characterized in two ways: (i) by stimulating an internodal strand emerging from the ganglia from which recordings were made, with a focal stimulating electrode made from tungsten wire (diameter, 100 μm) and (ii) by stimulating transmurally, using a pair of platinum stimulating electrodes (0.3 mm o.d.) placed 15 ± 2 mm both oral and anal to the recording site (Fig. 1C).
Most of the impaled neurones were located in myenteric ganglia close to where internodal strands entered ganglia (internodal hillock region; Fig. 1B) and in microganglia consisting of 5–20 neurones located at the junctions of several (2–4) internodal strands.
The electrophysiological chamber was continuously perfused with modified Krebs solution at 36 ± 0.5°C at a rate of 10 ml min−1 for at least 2 h before recordings were made. The Krebs solution contained nicardipine (2 μm), an L-type calcium channel blocker, and atropine (1 μm), a muscarinic antagonist, to block spontaneous and evoked contractions of the longitudinal muscle.
The morphology of neurones was determined by coupling the injected tracer neurobiotin with streptavidin (ABC Kit PK 6100; Vector Laboratories Corp., Burlingame, CA, USA) and then with horseradish peroxidase (HRP) in a standard HRP-diaminobenzadine (DAB) reaction (DAB Kit SK-4100; Vector Laboratories Corp.) (see Lunam & Smith, 1996; Smith & Lunam, 1998). The area of the soma of neurones (with and without the area of dendrites) was traced and calculated using a Metamorph imaging system (Universal Imaging Corp., West Chester, PA, USA). In the measurement of neuronal soma + dendrite area we excluded the long processes of multipolar AH neurones (see Furness et al. 1990).
Drugs and solutions
Atropine hydrochloride, hexamethonium, nicardipine and tetrodotoxin (TTX) were purchased from Sigma. Neurobiotin was purchased from Vector Laboratories Corp. All drugs were diluted from stock solutions made up at 10−2 M in either distilled H2O or physiological saline except nicardipine which was diluted in ethanol.
The composition of the modified Krebs solution was (mM): NaCl, 120.35; KCl, 5.9; NaHCO3, 15.5; NaH2PO4, 1.2; MgSO4, 1.2; CaCl2, 2.5; glucose, 11.5. The solution was gassed continuously with a mixture containing 3 % CO2-97 % O2 (v/v), pH 7.3–7.4.
Where appropriate Student's paired and unpaired t tests were used to determine significant differences (P < 0.05 level for significance). The data are presented as means ±s.e.m., where n is the number of cells.
In this study we primarily focused on determining the electrophysiological characteristics of myenteric neurones located in the corners of large ganglia close to the junction where internodal strands emerge from ganglia, and in microganglia formed at the intersection of several internodal strands (Fig. 1B). In addition, we also compared the electrophysiological properties of S neurones in these locations with those of 30 S neurones in the body of the ganglion (Fig. 1A; Table 1) in order to ensure that the recording conditions were not biasing our results.
Table 1. Comparison of electrophysiological properties of tonic S neurones and AH neurons impaled in large ganglia near junctions with internodal strands (INS) and in microganglia(MG) with those of S neurones impaled in the centre of large ganglia
Type and location of impaled neurons
Tonic S neurones at junctions with INS and in MG
−50 ± 1
522 ± 23
0.012 ± 0.004
AH neurones at junctions with INS and in MG
−69 ± 6
120 ± 10
0.28 ± 0.01
S neurones in the centre of large ganglia
−57 ± 5
295 ± 89
0.14 ± 0.01
Neurones were classified as S or AH using well-established criteria (Hirst et al. 1974; Bornstein et al. 1994). S neurones exhibited prominent FEPSPs in response to focal stimulation applied to an internodal strand emerging from the ganglia from which recordings were made. AH neurones exhibited a prolonged after-hyperpolarization following a single action potential and a characteristic hump during the repolarization phase of the action potential.
S neurones in peripheral locations
Of the 150 neurones sampled in peripheral locations, 120 were characterized as S neurones, of which 111 were tonic (slowly adapting) and 9 were phasic (rapidly adapting), and 30 as AH neurones.
During impalements hyperpolarizing current pulses (100 ms, 0.05 nA, 1 Hz) were injected continuously through the microelectrode. Tonic S neurones at the corners of ganglia could often be identified immediately prior to actual impalement as once the electrode touched the cell trains of extracellular action potentials followed each hyperpolarizing pulse. Electrophysiological characterization of these neurones was performed after the membrane potential had stabilized following removal of any applied holding current. This was important, otherwise tonic neurones could appear phasic when hyperpolarized by continuous current injection (see Smith & Lunam, 1998). Tonic S neurones were characterized by low resting membrane potential (RMP) and high input impedance (Rin) (Table 1). They were easily excited since the threshold for generating action potentials with depolarizing current pulses was low (< 0.01–0.02 nA) when compared with S neurones impaled in the centre of ganglia (see Table 1 and below). Also, tonic S neurones fired action potentials continuously (up to 42 Hz) in response to depolarizing current pulses of ⩽ 1 s duration (Fig. 2E). They responded to hyperpolarizing current pulses with one to eight ‘off’ spikes (Fig. 2B and D) or with a transient increase in spontaneous action potentials following termination of the pulse. The hyperpolarizing voltage response in these neurones had a time constant of 15.6 ± 1.4 ms (range, 8.8–29.4 ms; median, 13.3 ms). Many of these particular S neurones appeared to have a prominent A-type potassium current, which manifested itself as a lag in the action potential response to depolarizing current after the membrane was hyperpolarized to more negative values (Fig. 2F; see Levitan & Kaczmarek, 1991). Forty-eight tonic neurones (43 %) also fired action potentials spontaneously (7.6 ± 0.6 Hz; range, 0.3–19 Hz; n= 20) when any holding current was removed, an example of which is shown in Fig. 2H. Spontaneous action potentials, which were abolished by membrane hyperpolarization, were recorded for up to 2 h in some neurones before the impalement was lost.
Extreme care had to be taken when impaling tonic neurones because they were easily damaged owing to their relatively small size (Table 3). It was essential, therefore, to use high resistance microelectrodes (≥ 200 MΩ) otherwise neurones visibly swelled soon after impalement, had lower RMPs, lower input resistance, lost their spontaneous activity and became rapidly adapting in response to depolarizing current pulses. Two such neurones that were initially tonic and spontaneously active and that later became phasic were identified as longitudinal muscle motor neurones (see below). Some of the nine phasic S neurones observed in peripheral locations were probably tonic neurones which were damaged during penetration, since they required holding current to maintain impalement and their cell bodies appeared swollen and readily visible under the microscope.
Table 3. Parameters associated with S neurones
Functional class of S neurone
Measurement excluding dendrites
Measurement including dendrites
Mean ± S.E.M.
193 ± 48
15 ± 1
18 ± 4
304 ± 79
24 ± 3
27 ± 7
Mean ± S.E.M.
177 ± 18
15 ± 1
18 ± 1
356 ± 35
26 ± 2
31 ± 2
Mean ± S.E.M.
246 ± 97
16 ± 4
20 ± 3
651 ± 220
35 ± 8
40 ± 3
AH/type II(n= 10)
Mean ± S.E.M.
381 ± 51
20 ± 2
29 ± 3
679 ± 62
32 ± 2
57 ± 5
Subtypes of tonic S neurones
These tonic, highly excitable S neurones were further subclassified into two electrophysiological classes, SAH and SAD neurones, according to whether their post-stimulus responses to an evoked burst of action potentials was an after-hyperpolarization (AH) or an after-depolarization (AD).
SAH neurones, which account for 90 % of these S neurones (n= 100 of 111), exhibited a prominent after-hyperpolarization which depended on the number of preceding action potentials (Fig. 2A and C). At least two action potentials were required to observe any after-hyperpolarization in these neurones (Fig. 2A). The after-hyperpolarization had a mean duration of 711 ± 10 ms (n= 20) following an evoked burst of 10–12 action potentials at the RMP. The bee venom apamin (0.3 μm), which blocks one type of small conductance calcium-sensitive potassium (SK) channel in autonomic and CNS neurones (see Sah, 1996), significantly reduced the area under the after-hyperpolarization from 4.2 ± 0.2 to 1.0 ± 0.1 mV s (P > 0.01) in four of ten tonic S neurones, an example of which is shown in Fig. 3. In the other six tonic S neurones apamin had no detectable effect on the after-hyperpolarization.
SAD neurones, on the other hand, which account for the remaining 10 % of tonic S neurones (n= 11 of 111), were characterized by a small after-depolarization of 170 ± 10 ms duration (n= 11) following a burst of action potentials (Fig. 4A and B). The after-depolarization could even be observed following the ‘off’ spikes associated with a hyperpolarizing pulse (Fig. 4E). The after-depolarization was membrane potential dependent, increasing in size and duration with hyperpolarization and decreasing with depolarization (Figs 4A and C and 5); it had a reversal potential of between −35 and −45 mV (Fig. 5). In four neurones the post-stimulus response varied, initially being an after-depolarization and then giving way to an after-hyperpolarization. In these four neurones an after-depolarization was revealed after hyperpolarization was applied by constant current injection.
AH neurones in peripheral locations
AH neurones in peripheral locations, on the other hand, had properties similar to those previously reported: low RMP, low imput resistance and a fairly high threshold for action potential generation (Table 1; see Bornstein et al. 1994; Wood, 1994a). Seven of the AH neurones were initially classified as non-spiking neurones; however, once a slow excitatory post synaptic potential (SEPSP) was evoked in these neurones by focal stimulation applied close to the ganglia they changed their state of excitability and exhibited characteristics that were similar to other AH neurones.
S neurones in the centre of large of ganglia
The electrophysiological characteristics of 30 neurones within the middle of large myenteric ganglia, i.e. away from the corners, were determined. Twenty-six of these neurones were characterized as phasic S neurones and four as tonic neurones. These S neurones tended to have higher RMP, lower Rin and a higher threshold for action potential generation than the tonic, highly excitable S neurones described above (Table 1).
Synaptic potentials in tonic S neurones
Both fast and slow excitatory postsynaptic potentials (FEPSPs and SEPSPs, respectively) were recorded in tonic, highly excitable S neurones.
One to six action potentials superimposed on several suprathreshold FEPSPs were evoked in all S neurones by single pulse focal stimuli applied to internodal strands around the ganglia in which impalements were made (Fig. 6A and B). These neurones also responded with one or four action potentials to single pulse transmural stimulation applied either oral or anal to the recording site. Some neurones responded to both oral (Fig. 4F) and anal (Fig. 4G) stimuli and some just to oral (in a descending pathway) or anal (in an ascending pathway) stimuli (Table 2). The FEPSPs underlying the evoked action potentials were more readily observed after hyperpolarizing with continuous intracellular current injection prior to nerve stimulation (see Fig. 4H, oral, and Fig. 4I, anal). All SAD neurones (n= 4) tested received inputs from both oral and anal stimuli (Fig. 4F-I), suggesting they might receive convergent inputs from both descending and ascending nervous pathways (see Smith et al. 1992). FEPSPs evoked by either focal or transmural stimulation could be graded by varying the stimulus strength.
Table 2. Characteristics of highly excitable S neurones receiving FEPSPs
Functional class of neurone
Type of stimulation
Oral + anal
AH + AD
LMMN, longitudinal muscle motor neurone; CMMN, circular muscle motor neurone; AIN, ascending interneurone; UTSN, unidentified tonic S neurone; Circum, circumferentially applied stimulation; Oral/Anal, transmural stimulation applied oral and/or anal to the recording electrode; AH, after-hyperpolarization; AD, after-depolarization; AP, spontaneous action potentials.
Seventy-two (65 %) tonic S neurones exhibited a low frequency (0.06 ± 0.04 Hz; n= 10) discharge of spontaneous FEPSPs, which were more readily observed when the membrane was hyperpolarized (−80 to −90 mV) by continuous injection of current (Fig. 6C).
Trains of focal stimuli (0.05 ms, 20 Hz for 0.5 s, 10–15 V) applied to circumferentially arranged internodal strands emerging from the ganglia from which recordings were made evoked slow excitatory postsynaptic potentials (SEPSPs) in 19 of the 22 neurones tested. The SEPSPs gave rise to the continuous generation of action potentials. The SEPSPs had a duration of 18.3 ± 0.7 s (range, 13–30 s) (Fig. 6B). In contrast, repetitive transmural stimulation (0.05 ms, 20 Hz for 0.5 s, 10–15 V) applied either oral or anal to the recording site evoked only FEPSPs in all 10 neurones tested. These results suggest that much of the slow synaptic input originated from circumferentially directed processes, presumably from AH sensory neurones (see Kunze et al. 1993). Three of the neurones that responded with SEPSPs were positively identified as longitudinal muscle motor neurones and two as circular muscle motor neurones (see below).
Morphology of neurones
Thirty-seven neurones in peripheral locations that had been injected with neurobiotin were judged to be adequately filled after processing with DAB (see Methods and Lunam & Smith, 1996; Smith & Lunam, 1998) to allow thorough examination of their morphology and projections. Twenty-seven of these neurones were classified as highly excitable tonic S neurones and ten as AH neurones. The tonic S neurones comprised three functional classes (Fig. 7): (i) short circular muscle motor neurones (CMMNs; n= 10; Fig. 7A and B), which did not extend further than one or two rows of ganglia, and ended in two to four expansion bulbs (Fig. 7A and C; see Brookes et al. 1991b; Smith et al. 1992); (ii) longitudinal muscle motor neurones (LMMNs; n= 9), which had extensive arborization (⩽ 0.25 mm2) over the surface of the longitudinal muscle (Fig. 7D and E; see Bornstein et al. 1991; Brookes et al. 1992); and (iii) ascending interneurones (AINs; Fig. 7F and G; n= 8; Brookes et al. 1991a, 1997; Costa et al. 1996). The axons of some ascending interneurones could be traced up to 4500 μm within the myenteric plexus as they passed through several rows of ganglia. Each axon gradually narrowed before terminating in fine varicose fibres within a ganglion (Fig. 7F). The AH neurones had Dogiel type II morphology since they were oval with two or more branching processes (see Furness et al. 1990; Bornstein et al. 1994). Table 3 compares the relative sizes of these three classes of S neurone and the AH neurones, which had been found in these peripheral locations.
Five uniaxonal, phasic S neurones impaled near the centre of large ganglia were also identified. All five projected anally for more than four rows of ganglia. Three ended in expansion bulbs suggesting they were circular muscle motor neurones (probably inhibitory motor neurones) and two gave off small varicose side branches in ganglia suggesting they were descending interneurones (see Bornstein et al. 1991; Smith et al. 1992).
In this study, we concentrated on impaling myenteric neurones in the corners of large ganglia near the junction with an internodal strand and in microganglia. Both AH and S neurones were found in these locations. The majority of these S neurones were, however, atypical; they were classified as highly excitable tonic neurones because of the following characteristics: (i) firing of action potentials continuously throughout a depolarizing pulse; (ii) low RMP; (iii) high input resistance; (iv) low threshold for the generation of action potentials; (v) ‘off’ spikes following the repolarization phase of a hyperpolarizing current pulse; (vi) some fired action potentials spontaneously; and (vii) spontaneous FEPSPs. Focal stimulation of fibre tracts close to the ganglia in which impalements were made evoked suprathreshold FEPSPs and SEPSPs in these highly excitable tonic S neurones. The high excitability of these S neurones is probably related to their high Rin which is a consequence of their small size (Table 3).
In contrast, most of the S neurones that we impaled in the middle of ganglia had characteristics that were similar to those reported by others in preparations in which spontaneous muscle contractions had been paralysed with nicardipine (Bornstein et al. 1994; Kunze et al. 1994, 1997) or in unparalysed tissues without stretch (Kunze et al. 1998), namely more negative RMP, lower input resistance, mostly phasic, and no spontaneous action potentials.
The highly excitable tonic S neurones found in our study had electrophysiological characteristics that were similar to some myenteric neurones reported by Wood & Mayer (1978) to be spontaneously active. Wood & Mayer attributed the enhanced excitability of their atypical neurones to the fact that either (i) drugs were not used to prevent muscle contraction, or (ii) neurones were sampled throughout the ganglia as opposed to previous investigations in which impalements were usually confined to the central regions of large ganglia. In support of their first hypothesis, it has recently been shown that Dogiel type II neurones and some uniaxonal phasic S neurones increase their action potential responses to intracellular current and fire action potentials spontaneously when the tissue is stretched and that this activity was suppressed by muscle relaxants such as nicardipine and isoprenaline (Kunze et al. 1998).
In our study, however, we employed fairly standard techniques to facilitate intracellular impalement into neurones since an L-type channel blocker (nicardipine) and a muscarinic antagonist (atropine) were added to the bathing medium to inhibit spontaneous and evoked contractions of the muscle (nicardipine and a muscarinic antagonist: Bornstein et al. 1991; Smith et al. 1992; Brookes et al. 1997; Shuttleworth & Smith, 1999; nicardipine only: Kunz et al. 1993, 1994, 1995, 1997). Even under these conditions, our highly excitable, tonic S neurones exhibited many of the features of the excitable neurones described by Wood & Mayer (1978). Our results suggest, therefore, that location is also important since the atypical tonic S neurones were readily impaled at the corners of ganglia and in microganglia.
Subclasses of S neurones
Surprisingly, the excitable tonic S neurones could be further subdivided into two electrophysiological classes (SAH and SAD neurones) according to their post-stimulus responses following a burst of action potentials evoked by a depolarizing current pulse.
The after-hyperpolarization, although not generally recognized as a property of myenteric S neurones until recently (Shuttleworth & Smith, 1999), is a common property of autonomic and CNS tonic neurones (see review by Sah, 1996; and also Davies et al. 1966; Smith & Lunam, 1998). The post-stimulus after-hyperpolarization was observed in all three functional classes of these excitable S neurones. The after-hyperpolarization, in some of these tonic S neurones, was clearly due to the opening of SK potassium channels since it was blocked by apamin (see Sah, 1996). This was surprising since apamin is generally considered not to affect the excitability of myenteric S or AH neurones (Kunze et al. 1994). We have recently found that repetitive bursts of action potentials, which are blocked by TTX, in tonic S neurones produced a robust calcium influx through high threshold N-type calcium channels (Shuttleworth & Smith, 1999). The duration of the resulting increase in intracellular calcium was proportional to the duration of the after-hyperpolarization, suggesting that the after-hyperpolarization is due to the opening of calcium-dependent potassium channels.
The after-depolarization, which was found in both longitudinal muscle motor neurones and circular muscle motor neurones, has not been previously reported for myenteric neurones. The after-depolarizing response has, however, been observed in other neurones such as mammalian spinal motor neurones (Granit et al. 1963), hippocampal neurones (Wong & Prince, 1981) and autonomic neurones (Akasu et al. 1990; Smith & Lunam, 1998). The after-depolarization in these tonic S neurones, had a reversal potential of −35 to −40 mV. It could, therefore, be due to activation of a calcium-activated Cl− conductance, as it appears to be in rabbit parasympathetic neurones (Akasu et al. 1990), although activation of a non-specific cation conductance cannot be ruled out. The increase in the amplitude and duration of the after-depolarization with hyperpolarization may result from the increased driving force for Cl−, Ca2+ or Na+. It is of interest, that some SEPSPs in myenteric neurones (mainly AH neurones) are mediated by an increase in Cl− conductance as well as a decrease in potassium conductance (Bertrand & Galligan, 1994). Therefore, it is possible that the post-stimulus after-depolarization may be more strongly activated during slow synaptic input in these S neurones.
The roles of the after-hyperpolarization and after-depolarization in myenteric S neurones are unknown. However, in CNS motor neurones, hippocampal neurones and peripheral neurones both these post-stimulus events appear to regulate action potential frequency and firing patterns (spike frequency adaptation) (Granit et al. 1963; Wong & Prince, 1981; Sah, 1996). Another role of the after-depolarization may also be to reduce the duration of the after-hyperpolarization (Granit et al. 1963; Wong & Prince, 1981).
Identification of S neurones
There are many different functional classes of interneurones and motor neurones in the myenteric plexus of the guinea-pig ileum (see Costa et al. 1996). Injection of neurobiotin revealed that the highly excitable S neurones were longitudinal muscle motor neurones, short circular muscle motor neurones and ascending interneurones. All of these three groups of neurones are likely to be cholinergic. Retrograde tracing studies combined with immunohistochemistry have shown that all ascending interneurones and 97 % of longitudinal muscle motor neurones in the ileum are choline acetyltransferase immunoreactive, as are the short circular muscle motor neurones with filamentous dendrites (see Brookes et al. 1991a, b, 1992, 1997). Ascending interneurones and longitudinal muscle motor neurones are also calretinin immunoreactive (Brookes et al. 1991a). Brookes et al. (1997) recently used DiI retrograde labelling to identify and determine the electrophysiological characteristics of ascending interneurones. They also reported that ascending neurones were tonic S neurones. Kunze et al. (1997), on the other hand, reported that ascending interneurones were rapidly adapting. The reasons for such contrasting results are unclear. We found, along with Brookes et al. (1997), that the resistance of microelectrodes needed to be high in order to successfully electrophysiologically characterize small neurones; otherwise, they were easily damaged and exhibited phasic electrical properties (see Results). The resistance of our microelectrodes and those of Brookes et al. (1997) were ≥ 200 MΩ and > 140 MΩ, respectively, whereas those of Kunze et al. (1997) ranged from 80–200 MΩ.
Function of highly excitable S neurones
It seems likely that the three classes of highly excitable S neurones identified in the present study, which are likely to be found clustered together (see Brookes et al. 1991a), are functionally related. Presumably, all three classes of neurones are coupled and involved in initiating peristalsis, as postulated previously for ascending interneurones by Brookes et al. (1997). Their small size and consequent high excitability of these neurones suggests that they are responsible for the synchronous contraction of both the longitudinal and circular muscle layers which initiates peristalsis (see Smith & Robertson, 1998).
Tonic, highly excitable S neurones exhibit some degree of topographical organization since they can be readily impaled in large ganglia near the junctions of internodal strands or in microganglia formed at the junction of several (2–4) internodal strands. These neurones can be subdivided according to their post-stimulus responses into SAH and SAD neurones. It is possible that these S neurones, which comprise three distinct classes of neurones, may be in specific circuits that initiate peristalsis.
We are grateful to Matthew J. Free who helped with data analysis. This work was supported by a grant from the National Institutes of Health, USA (PO1 DK 41315).