Enteric and extrinsic sensory neurons respond to similar stimuli. Thus they may be activated in series or in parallel. Because signal transmission via synapses or mediator release would depend on calcium, we investigated its role for extrinsic afferent sensitivity to chemical and mechanical stimulation. Extracellular multiunit afferent recordings were made in vitro from paravascular nerve bundles supplying the mouse jejunum. Intraluminal pressure and afferent nerve responses were recorded under control conditions and under four conditions designed to interfere with enteric neurotransmission. We found that phasic intestinal contractions ceased after switching perfusion to Ca2+-free buffer with or without a purinergic P2 receptor antagonist, pyridoxal phosphate-6-azo(benzene-2,4-disulfonic acid) (PPADS) or cadmium (blocking all Ca2+-channels) but not following ω-conotoxin GVIA (N-type Ca2+-channel blocker). Luminal HCl (pH 2) and 5-HT (500 μM) evoked peak firing of 17 ± 4 impulses per second (imp/s) (n = 10) and 21 ± 4 imp/s (n = 13) under control conditions. These responses were reduced to 4 ± 2 imp/s and 5 ± 2 imp/s by cadmium (n = 7, P < 0.05), to 7 ± 2 imp/s and 6 ± 1 imp/s by Ca2+-free perfusion (n = 6, P < 0.05), and to 3 ± 1 imp/s and 4 ± 1 imp/s by Ca2+-free perfusion with PPADS (n = 6, P < 0.05). Responses were unchanged by ω-conotoxin GVIA. Mechanical ramp distension of the intestinal segment to 60 cmH2O was not altered by any of the experimental conditions. We concluded that HCl and 5-HT activate extrinsic afferents via a calcium-dependent mechanism, which is unlikely to involve enteric neurons carrying N-type calcium channels. Extrinsic mechanosensitivity is independent of enteric neurotransmission. It appears that cross talk from the enteric to the extrinsic nervous system does not mediate extrinsic afferent sensitivity.
the gastrointestinal tract is endowed with an elaborate network of surveillance systems that comprise both intrinsic and extrinsic sensory neurons as well as an array of immune cells and specialized epithelial cells (9). Sensory neurons can be classified according to the location of their cell bodies and the pattern or projection of their axons. Intrinsic primary afferent neurons (IPANs) are estimated to make up about 30% of the enteric nervous system (11). These have cell bodies either in the submucosal or the myenteric plexus and afferent axons that innervate both the mucosa and the muscle layers of the gut wall. IPANs synapse with each other and in this way form a self-reinforcing network, providing sensory input from the mucosa and muscle that in turn modulates the enteric reflex circuitry that controls gastrointestinal function. The sensory nerve endings have a number of transduction channels that transform stimulus energy into a coded sequence of action potentials that are then transmitted synaptically through the enteric network (3, 5, 37). However, in addition, these sensory endings have ligand-gated receptors and channels that transmit sensory information from other structures within the gut wall. One important source of such input arises from the proximity of sensory terminals in the lamina propria to enterochromaffin cells (EC cells) in the mucosal epithelium. 5-HT released from these cells has been shown to contribute to the mechanical and chemical sensitivity of IPANs supplying the intestinal mucosal (2). In addition to these intrinsic sensory neurons, there are two groups of extrinsic sensory neurons that project information out of the gut wall to the central nervous system. Vagal afferents have their cell bodies in the jugular and nodose ganglia, whereas spinal afferents have their cell bodies in the dorsal root ganglia and project into the spinal cord. The visceral endings of the vagal and spinal afferents are located in different sites within the gut wall, namely mucosa, submucosa, muscle, myenteric plexus, serosal, and mesenteric connections (1, 21). These terminals are positioned to respond to changes in the mechanical and chemical environment both within the lumen and within the gut wall and its mesenteric connections (49). Intrinsic and extrinsic afferents share a number of characteristics, i.e., both groups have similar innervation territories and are responsive to similar modalities of mechanical and chemical stimulation (11, 14, 21).
Both spinal and vagal afferents can be seen coursing through the enteric nervous system, where they are potentially exposed to chemical mediators released during enteric reflex activation (5, 14, 15). Vagal afferents in particular form elaborate structures around enteric ganglia. These so-called intraganglionic lamina endings form basket-like structures around ganglia and have been shown by Brookes and colleagues to be the site of mechanosensitive hotspot (49). This proximity between enteric and extrinsic sensory neurons raises the intriguing possibility that they may share a common mechanism of activation with enteric neural activity being transmitted to the terminals of extrinsic afferents following release of chemical mediators into the enteric ganglionic neuropil. This would give rise to “cross talk” between the enteric and extrinsic nervous system, and as such sensory input to enteric reflex circuits would also be available to the central nervous system. Such cross talk would likely depend on calcium-dependent mechanisms responsible for vesicular release of mediators into the synaptic neuropil. Calcium entry through mainly N-type calcium channels is required for transmitter release (29), and, therefore, removal of extracellular calcium or calcium channel blockers should attenuate this process. However, extracellular calcium can interact with some ion channels to increase membrane conductances (20). In contrast, increases in intracellular calcium may lead to the modulation of various ion channels including calcium-gated potassium channels that decrease the excitability of neurons (8). Therefore, calcium can have both positive and negative effects on sensory neuron excitability, and the relative contribution of these counteractive influences can only be determined empirically.
In the present study, we aimed to manipulate calcium channels by different techniques to examine the effect on gastrointestinal afferent sensitivity. By recording from mesenteric paravascular bundles in vitro, we simultaneously sampled the activity of both vagal and spinal afferents supplying the jejunum to determine the sensitivity in both mechanical and chemical stimulation. We hypothesized that mesenteric afferent sensitivity depends on functional calcium channels in the intestinal wall, i.e., on intact synaptic transmission in the enteric nervous system.
MATERIALS AND METHODS
All experiments were performed with female mice (C57BL6, 20–30 g), which were fed a standard laboratory diet. Animals were anesthetized with pentobarbitone (60 mg/kg ip). Following laparotomy, a 2-cm segment of jejunum with mesenteric attachment was excised and placed into a custom-made organ bath. In one chamber of the organ bath (perfusion chamber), the jejunal segment was superfused with Krebs buffer equilibrated with 95% O2 and 5% CO2 [composition (in mM): 143.5 Na+, 5.9 K+, 126 Cl−, 2.5 Ca2+, 1.2 Mg2+, 1.2 H2PO4, 1.2 SO4, 25 HCO3−, 10 glucose, 1 and sodium butyrate, pH 7.4; rate 10 ml/min, temperature 32°C; a maximum of 2 segments was used per animal]. The mesenteric arcade was pulled through an aperture into a separate chamber (recording chamber). The aperture was sealed with Vaseline and the recording chamber filled with colorless heavy liquid paraffin, prewarmed to 32°C to insulate the recording electrodes. The gut lumen was cannulated at both ends and perfused with isotonic saline from the proximal end (perfusion rate 10 ml/h), while the distal cannula remained open to the atmosphere for drainage except during mechanical ramp distension (see below). The intraluminal pressure was recorded continuously at the proximal end of the intestinal segment via a Y-piece, which was connected to a pressure transducer (Ohmeda DTX Plus Transducer; Ohmeda, Liberty Corner, NJ). The pressure was typically 1–3 cmH2O at baseline on top of which phasic contractile events could be observed. A separate segment of jejunum was used for each experiment. The committee for animal experiments at the University of Tuebingen approved of the protocol before these procedures. N numbers vary for technical reasons because it was not always possible to run the whole protocol in a single preparation.
A single paravascular nerve bundle was dissected out from the mesenteric arcade and attached to one of a pair of platinum recording electrodes. A strip of connective tissue was wrapped around the other electrode to act as a reference. The electrodes were connected to a CED single channel 1902 preamplifier/filter (Cambridge Electronic Design, CED, Cambridge, UK), and the signal was differentially amplified 10,000 times and filtered with a bandwidth of 100 Hz to 1 kHz. The output from the 1902 amplifier was passed into a power Micro 1401 interface system (CED), captured, and viewed online by a PC running Spike 2 software (version 4.01; CED), together with the output from the pressure transducer showing intraluminal pressure.
HCl (pH 2 for 4 min) and 5-HT (500 μM for 2 min) were perfused through the gut lumen at a rate of 10 ml/h separated by an interval of 10 min, which served for washout with normal saline at the same perfusion rate. Subsequently, the outlet cannula was closed, and the jejunal segment was distended with normal saline, which was infused at a rate of 1.5 ml/min by a perfusion pump (IVAC 711; IVAC, San Diego, CA) until the maximum intraluminal pressure of 60 cmH2O was reached.
In control experiments, this protocol with HCl, 5-HT, and mechanical ramp distension was run 20 min after the preparation was allowed to stabilize during Krebs buffer perfusion. In separate experimental subgroups, the protocol was run 10–20 min after perfusion was switched to one of the following solutions: 1) Ca2+-free Krebs (containing 2.5 mM Mg2+ + 0.1 mM EGTA, a Ca2+ chelator) to inhibit synaptic transmission by removal of extracellular Ca2+; 2) Ca2+-free Krebs (containing 2.5 mM Mg2+ + 0.1 mM EGTA) + pyridoxal phosphate-6-azo(benzene-2,4-disulfonic acid) (PPADS) (50 μM); because ATP is released by removal of extracellular Ca2+ and this may influence afferent firing indirectly, PPADS was added as a purinergic P2 receptor antagonist to block any activation of afferents by ATP; 3) Krebs plus ω-conotoxin GVIA (500 nM) to block N-type Ca2+ channels; and 4) cadmium (250 μM) in HEPES buffer (50 mM) to block all Ca2+ channels unselectively.
Use of HEPES buffer in the fourth experimental group was necessary to maintain cadmium in solution. In pilot experiments, afferent nerve responses to HCl, 5-HT, and distension were unchanged in HEPES compared with Krebs buffer. This allowed comparison of the four experimental subgroups with each other. Ca2+-free solutions (Solutions 1 and 2) were perfused for at least 20 min before subsequent stimuli (HCl, 5-HT, ramp distension) to ensure complete Ca2+ washout in the preparation.
Pentobarbitone was obtained from Rhône Mérieux, Lyon, France. 5-HT, cadmium, heavy liquid paraffin, PPADS, EGTA, and all buffer salts were from Sigma Chemicals (St. Louis, MO). ω-Conotoxin GVIA was obtained from Alomone Laboratories (Jerusalem, Israel).
Baseline discharge frequency (impulse per second, imp/s) during perfusion with Krebs solution was calculated by averaging afferent nerve firing for a 60-s period following a 20-min period for signal stabilization. After perfusion was changed to a modified solution according to different experimental subgroups (see above, Solutions 1–4), baseline discharge was again quantified in this manner.
Following HCl, 5-HT, and mechanical ramp distension, peak increase in afferent nerve discharge frequency above baseline was determined. These peak responses were calculated as mean discharge per second from a 3-s bin analyzed during peak response. Intestinal motility was quantified as mean of peak amplitudes from phasic contractile events observed during 200 s following stabilization of afferent firing in each of the solutions in the perfusion chamber. Data are presented as means ± SE and compared by one-way ANOVA followed by Dunnett's corrections for multiple comparisons. P < 0.05 was considered statistically significant.
Under control conditions with Krebs buffer perfusion, phasic increases in intraluminal pressure were recorded with a peak pressure of 11.2 ± 0.9 cmH2O (n = 23). When Ca2+-free Krebs buffer was perfused, peak pressures were reduced to 1.2 ± 0.5 cmH2O and to 1.9 ± 0.3 cmH2O during perfusion with Ca2+-free Krebs and PPADS (50 μM, n = 6, P < 0.05). Contrary to Ca2+-free conditions, contractile activity continued with peak pressures of 9.1 ± 3.4 cmH2O when the perfusion was switched to a solution containing the N-type calcium channel antagonist ω-conotoxin GVIA (500 nM, n = 6). A reduction of phasic pressure events to 2.4 ± 0.2 cmH2O was also observed during perfusion with cadmium containing HEPES buffer (n = 6, P < 0.05), which was similar to Ca2+-free conditions (Fig. 1, A–C).
Baseline afferent discharge.
Under control conditions with Krebs buffer perfusion, spontaneous multiunit afferent nerve discharge was observed at baseline. This afferent discharge was maintained over the time course of the experiments and consisted of spikes with different amplitudes and waveforms indicative of multiunit recordings. Baseline afferent nerve discharge was 14 ± 4 imp/s, when afferent discharge was evaluated during the initial control conditions with Krebs buffer perfusion from all experimental subgroups (n = 23).
Afferent firing was increased from 14 ± 3 imp/s during the initial recording period under control conditions to 24 ± 3 imp/s when perfusion was switched to Ca2+-free Krebs solution (P < 0.05, n = 6). This increase in afferent firing was from 12 ± 2 imp/s to 26 ± 4 imp/s when perfusion was switched to Ca2+-free Krebs buffer containing PPADS (50 μM, P < 0.05, n = 6). Baseline afferent firing was unchanged compared with control conditions when perfusion was continued with a solution containing ω-conotoxin GVIA (14 ± 2 imp/s, n = 6) or cadmium (12 ± 3 imp/s, n = 6; Fig. 1, A–C).
Afferent sensitivity to HCl, 5-HT, and mechanical ramp distension.
Intraluminal acid at pH 2 evoked a robust increase in afferent nerve discharge which peaked at 17 ± 4 imp/s above baseline (n = 10). This response to HCl was attenuated following perfusion with Ca2+-free buffer to 7 ± 2 imp/s (n = 6, P < 0.05) and to 3 ± 1 imp/s (n = 6, P < 0.05) following Ca2+-free buffer containing PPADS. It was unchanged following ω-conotoxin GVIA with a peak firing of 20 ± 4 imp/s (n = 6) but also attenuated following cadmium (4 ± 2 imp/s: n = 7, P < 0.05; Fig. 2A). Intestinal motility was unchanged during intraluminal perfusion with HCl (Fig. 3).
Similar to HCl, intraluminal 5-HT (500 μM) was followed by a robust increase in afferent discharge to 21 ± 4 imp/s above baseline (n = 13). The response to 5-HT was reduced to 6 ± 1 imp/s in Ca2+-free conditions (n = 6, P < 0.05) and to 4 ± 1 imp/s in Ca2+-free buffer with PPADS (n = 6, P < 0.01), whereas ω-conotoxin GVIA had no effect (27 ± 6 imp/s, n = 6). Afferent discharge to 5-HT was reduced to 5 ± 2 imp/s during perfusion with a cadmium containing solution (n = 7, P < 0.05; Fig. 2B). Intraluminal perfusion with 5-HT did not alter intestinal motility (Fig. 3).
Mechanical ramp distension of the intestinal segment to 60 cmH2O stimulated a pressure-dependent increase in afferent discharge (n = 12). Afferent responses to distension were similar in all experimental subgroups regarding response profile and peak firing, independent of the solution perfused in the organ bath (Fig. 4).
The present study investigated the role of Ca2+-dependent mechanisms for intestinal motility and mesenteric afferent nerve discharge at baseline and following mechanical and chemical stimuli.
Phasic intestinal contractions were abolished under all Ca2+-free conditions. Calcium is needed to maintain the calcium balance of the smooth muscles and for contractile processes (36). It is, therefore, not surprising that intestinal motor events were attenuated in the absence of extracellular calcium because L-type calcium channels play an important role in excitation/contraction coupling in intestinal smooth muscle (22). Similarly, cadmium, which will block all voltage-gated calcium channels, including L-type channels, causes a pronounced inhibition of intestinal motility. In addition, Ca2+-dependent mechanisms are necessary for synaptic transmission, which will be attenuated in the absence of extracellular calcium or in the presence of cadmium (27). Synaptic transmission relies more on N-type calcium channels, and these can be selectively blocked by ω-conotoxin GVIA. However, in contrast to Ca2+-free and cadmium solutions, ω-conotoxin had no effect on intestinal motility, indicating that under the present experimental conditions the motor activity was driven largely by myogenic mechanisms.
Spontaneous mesenteric afferent nerve discharge at baseline was unchanged when cadmium was added to the buffer solution despite the marked reduction in phasic intestinal motor activity. Under baseline conditions, therefore, it appears that contractile activity does not drive afferent discharge. Similarly, blocking N-type Ca2+ channels with ω-conotoxin GVIA failed to alter spontaneous afferent firing. Because both cadmium and ω-conotoxin will attenuate synaptic transmission within the enteric nervous system (41, 44, 45), their lack of effect on baseline afferent discharge would indicate that spontaneous afferent firing does not require mediator release or synaptic transmission in the enteric nervous system and rather suggests that it is a property of the afferent ending itself. Strikingly, baseline afferent firing was increased under Ca2+-free conditions at a time when intestinal motility was attenuated. The most likely explanation is that excitability of the afferent endings is increased in the absence of external Ca2+ potentially subsequent to disinhibition of certain ion channels in the afferent membrane. Indeed, there is considerable evidence from sensory neurons maintained in cell culture that extracellular Ca2+ modulates the gating properties of a number of ion channels that regulate neuronal excitability (40). Moreover, Ca2+ is required by ectonucleotidases for the breakdown of ATP to adenosine (33, 34, 50, 51). In the absence of extracellular calcium, there is an increase in ATP content (33), and, because intestinal afferents express P2X receptors (26), there may be a secondary increase in afferent excitability via ligand-gated activation. This possibility, however, was ruled out because adding the purinergic P2 receptor antagonist PPADS to the Ca2+-free perfusate did not alter increased afferent discharge at baseline under Ca2+-free conditions.
Mechanical ramp distension of the jejunal segment used in the present study was designed to assess the contribution of low-threshold and high-threshold afferents traveling via vagal and spinal afferent pathways (4, 12, 31). The afferent response during ramp distension was unchanged following the various conditions designed to manipulate calcium metabolism. This may not be surprising regarding spinal afferents because many have mechanosensitive endings and mesenteric connections at some distance from enteric ganglia (1). However, the response to distension was also unchanged at lower distending pressure that might be considered more physiological. Vagal afferents that terminate as intraganglionic laminar endings (49) are considered the site of low-threshold mechanosensitivity. In this location, these endings may be exposed to the chemical environment in the synaptic neuropil and therefore are a likely candidate for cross talk between extrinsic and enteric sensory endings. The absence of any attenuation of mechanosensitivity following calcium manipulation suggests that this, too, is independent of synaptic activity, which is consistent with previous observations (48). Nevertheless, the multiunit signal does not completely rule out the possibility that a minor subpopulation of extrinsic afferents depends on synaptic activity. A third population of afferents in the mesenteric nerve bundles arise from intestinofugal fibers that project to the prevertebral ganglia. These receive cholinergic synaptic input following mechanical stimulation (39), which would be eliminated under Ca2+-free conditions and during Cd superfusion. That the response to distension is independent of calcium implies that these intestinofugal fibers do not contribute greatly to the sensory signals conveyed in the mesenteric nerve bundles at the level of the mouse jejunum, which is in keeping with their anatomical distribution, which is predominantly in the ileum and colon (10). In summary, mechanosensitivity in the mouse jejunum seems to be independent of calcium regardless of whether it is stimulated by a physiological or noxious trigger.
Although mechanosensitivity appears to be clearly mediated by direct activation of afferent endings, there may be a good rationale to suggest that chemosensitivity may be indirect, in particular, the evidence that EC cells are responsible for luminal “tasting” and influence both enteric and extrinsic afferent following the release of 5-HT (2, 13). We therefore examined luminal sensitivity to acid and 5-HT to assess potential involvement at the level of the EC cell-sensory terminal interface or following activation of enteric sensory neurons and transmission within enteric ganglia.
We found that the afferent nerve response to acid and 5-HT was attenuated by removal of calcium from the perfusate and by adding the nonselective calcium channel blocker cadmium, whereas the N-type calcium channel blocker ω-conotoxin GVIA had no effect. Thus, irrespective of changes in excitability at the level of the sensory terminal, the absence of any influence of ω-conotoxin (17) suggests that cross talk from intrinsic primary afferent neurons to extrinsic afferents does not contribute to the serotonin sensitivity of extrinsic afferents.
However, how is the attenuated response to 5-HT explained during Ca2+-free conditions? 5-HT may activate enteric neurons in the myenteric plexus via 5-HT3, 5-HT4, and 5-HT1P receptor (2, 36, 43). Because cross talk from the enteric nervous to the extrinsic nervous system does not seem to occur, extrinsic afferent nerve fibers were probably directly stimulated by luminal 5-HT. Indeed, it was shown previously that extrinsic afferents respond to 5-HT directly via the 5-HT3 receptor subtype (18), which is located on vagal afferents (17). Considering that the 5-HT3 receptor is a ligand-gated ion channel (6) that, when activated by 5-HT, permits Ca2+ influx (16), it is likely that this activation is attenuated in the absence of extracellular calcium. A potential secondary effect by accumulated ATP during Ca2+-free conditions (33) was again ruled out by adding the purinergic P2 receptor antagonist PPADS in a different series of experiments.
This mechanism of activation of the 5-HT3 receptor does not explain inhibition of the 5-HT response in the presence of cadmium, which blocks calcium channels unselectively. It was shown, however, that desensitization of the 5-HT3 receptor is regulated by other voltage-gated Ca2+ channels (24). In other words, Ca2+ channels in the cell wall of extrinsic afferents may have been blocked by cadmium with a subsequently enhanced desensitization of the 5-HT3 receptor. Alternatively, cadmium may have bound directly to the 5-HT3 receptor (23) with subsequent potential inhibition, or it may change overall excitability of intestinal afferents, which depends on external calcium (40). This mechanism may involve TTX-resistant Na+ channels, which have been described in sensory neurons (Nav1.8 and 1.9) and potentially contribute to the mechanisms that regulate sensory neuronal function, especially in the context of hypersensitivity (19). That Nav1.8 is sensitive to cadmium has also been demonstrated (28), and indeed the current that these channels mediate in isolated dorsal root ganglions is more sensitive to cadmium than that mediated by Nav1.9. Therefore, cadmium at concentrations below those used in the present study markedly attenuate both Nav1.8 and Nav1.9 currents and potentially result in attenuated excitability. What is more difficult to explain with this concept is why the effects of cadmium are seen only with responses to luminal chemical stimulation, while baseline firing and mechanosensitivity were unaffected. One possibility is that the latter drives the afferents beyond the level at which subtle changes in excitability might become manifest. However, this remains speculative pending further investigation.
Several possible explanations exist that may account for the observation that the afferent nerve response to acid was reduced during Ca2+-free perfusion and in the presence of cadmium. Although minor subpopulations of afferents may behave differently, cross talk from the enteric to the extrinsic nervous system seems again unlikely because blocking neuronal N-type calcium channels in the enteric nervous system was without effect. One possible mechanism to explain the effect observed during acid exposure is that acid has the potential to activate transient receptor potential vanilloid 1 receptors or acid-sensitive ion channels on afferent nerve terminals directly (35, 38, 42). Both mechanisms would be attenuated under Ca2+-free conditions because calcium as a substrate for these ion channels would be missing (46). However, this does not explain why the response was attenuated during Cd exposure, rendering this suggested explanation unlikely. A second possible explanation is that acid triggers the release of 5-HT from EC-cells, which was shown in previous investigations (25, 47). The mechanism involves the transient receptor potential ankyrin 1 receptor (32), which seems to be sensitive to alterations in pH (7). This mechanism involves calcium influx in the EC-cell (32), which explains the attenuation of the acid response during Ca-free conditions and Cd by reduced 5-HT release. 5-HT would then be subject to the mechanisms illustrated above for 5-HT sensitivity, which would elegantly explain why a similar modulation of the afferent response to acid and 5-HT occurred during different types of calcium manipulation.
In conclusion, extrinsic afferent nerve discharge to acid and 5-HT is stimulated by a calcium-dependent mechanism, suggesting mediator release with subsequent activation of afferents. This mediator release does not involve N-type calcium channels and is, therefore, unlikely to stem from neurons in the enteric nervous system, rendering EC cells as likely candidates. Extrinsic afferent nerve discharge to mechanical stimulation was independent of calcium, indicating that mechanosensitive afferents were directly activated. These observations suggest that cross talk from the enteric to the extrinsic nervous system does not occur. It is of note, however, that mechanisms of afferent activation may be different for other stimuli and that, in contrast to naive animals, cross talk may occur under pathological conditions such as intestinal inflammation.
- Copyright © 2009 the American Physiological Society