Interactions of Serotoninergic, Cholinergic, and Tachykinin-Containing Nerve Elements in the Rabbit Small Intestine



This report presents novel results on the effects of serotonin (5-HT) on longitudinal muscle contractions in the rabbit ileum and the interactions of serotonin with some neuronal elements of the myenteric plexus. We showed previously that serotonin-triggered contractions involved two mechanisms in the rabbit ileum: neuronal excitation (via 5-HT2 receptors in the neurons) and direct muscular stimulation (via 5-HT4 receptors in the muscle). Here, we focus on the neuronal 5-HT2 receptor pathway and report further pharmacological and immunocytochemical data clarifying the details of the mechanisms. We observed that antagonists for neurokinin (NK1 and NK2) receptors partially blocked the serotonin response, but NK3 receptor antagonists had no effect. Pretreatment by atropine (ATR) eliminated the NK1 receptor antagonist resistant contractions. In contrast, the NK1 antagonist did not depress the ATR-resistant contraction when ATR was added first. 5-HT2 receptor agonist-induced contractions were partially suppressed by ATR, hexamethonium, and NK1 or NK2 receptor antagonists. In conclusion, serotonin acting through 5-HT2 receptors could stimulate interneurons and excitatory motor neurons. Immunocytochemical staining revealed an extensive tachykinin-immunoreactive (IR) network in the myenteric plexus. Approximately 52% of all myenteric neurons were labeled. 5-HT-IR fibers could be detected around both choline acetyltransferase- and tachykinin-IR cells, suggesting functional relationships between them. Consistent with our pharmacological observations, we found that immunopositive nerve elements for 5-HT2A receptor and double-labeled immunostaining revealed a remarkable overlap between tachykinin-IR neurons and 5-HT2A-IR elements. Anat Rec, 2009. © 2009 Wiley-Liss, Inc.

One of the principal regulators of digestion in mammals is the enteric nervous system. Enteric neurons coordinate a variety of functions, such as the secretion of digestive enzymes, control of absorption, and the regulation of motility. Intestinal smooth muscle contractility is organized by polarized reflex pathways through both excitatory and inhibitory neurons (Goyal and Hirano,1996; Kunze and Furness,1999; Olsson and Holmgren,2001). Excitatory effects are mediated by serotonin (5-HT), tachykinins, acetylcholine (ACh), gastrin-releasing peptide/bombesin, and so forth, whereas the transmitters that inhibit gut motility are the members of the vasoactive intestinal peptide-family, nitric oxide, γ-aminobutyric acid, and ATP (Shuttleworth and Keef,1995; Kunze and Furness,1999).

5-HT has been identified as both a local hormone and a neurotransmitter that is synthesized and stored in robust amounts in the gastrointestinal tract (Erspamer and Asero,1952; Gershon et al.,1965). Numerous pharmacological studies have provided evidence that 5-HT can act on the secretion of intestinal fluid and electrolytes (Bjorck et al.,1988) and gastrointestinal motility (Grider et al.,1996; Wade et al.,1996; Bush et al.,2001). In guinea pig, morphological studies described a rich, varicose 5-HT-immunoreactive (IR) fiber network in the myenteric and submucous plexuses and reactive nerve cells with short, broad dendrites and a single long process in the myenteric plexus (Costa et al.,1982). Furthermore, 5-HT-IR nerve elements were also found in the large and small intestines of many other mammals (Timmermans et al.,1990; Sang and Young,1996; Fujimiya et al.,1997; Toole et al.,1998). In rabbit, however, 5-HT could be detected only after preloading in the myenteric plexus. Nevertheless, nerve fibers that could accumulate 5-HT were IR for tryptophan hydroxylase, the enzyme that synthesizes 5-HT (Dénes et al.,2003). Application of exogenous 5-HT caused a concentration-dependent increase of contractions in rabbit ileum. The results indicated that 5-HT could act directly on smooth muscle through 5-HT4 and on neurons through 5-HT2 receptors (Dénes et al.,2003).

Tachykinins are another type of widely distributed neuroactive mediators in the mammalian ENS. Tachykinin family includes substance P (SP), physalaemin-like peptide, neurokinin A (NKA), and neurokinin B (NKB) (Lazarus and Di Augustine,1980; Kangawa et al.,1983; Nawa et al.,1984). Immunocytochemical studies described SP-IR nerve elements in all parts and layers of the gastrointestinal system (Heitz et al.,1976; Costa et al.,1980,1981; Schultzberg et al.,1980). Application of tachykinins to the intestine induces various effects including excitation of muscle, increase in blood flow and secretion (Holzer and Holzer-Petsche,1997). The diverse effects of these peptides are mediated through three receptors called NK1, NK2, and NK3 (Guard and Watson,1991; Grady et al.,1996).

In rabbit, we have detailed knowledge about mechanisms how 5-HT regulates nutrient absorption, electrolyte transport, or chloride secretion (Donowitz et al.,1980; Sundaram et al.,1991; Salvador et al.,1997,2000). However, regarding the effect of 5-HT on intestinal motility in rabbit, only a few articles have been published. According to Salvador et al. (2000), 5-HT increases contractions and tone of muscle fibers in the duodenum, jejunum, and ileum. Based on our previously published data, this study was undertaken to ascertain whether (i) noncholinergic neuronal components take part in 5-HT-evoked contractions and (ii) how 5-HT2 receptors are situated within the neuronal circuitry. In addition, (iii) we provide data regarding the morphology and distribution of the tachykinin-containing elements in the rabbit ileal myenteric plexus.


Adult white rabbits of both sexes (2–2.5 kg) were used as experimental animals. They were fed and watered ad libitum. Rabbits were killed by a single strong blow to the cervical vertebrae and pithed.

The last 10 cm of small intestine just before the ileocecal junction was discarded, and the rest of the ileum was used for experiments.


To detect fibers taking up 5-HT, pieces of gut were incubated in Krebs-Henseleit solution containing 10−5 M 5-HT at room temperature for 25 min. To fix tissues, the ileum was filled with KREBS solution (mM: NaCl 119, NaHCO3 25, KH2PO4 1.2, CaCl2 2.5, MgSO4 1.5, KCl 2.5, and glucose 11) and placed into 4% paraformaldehyde. For 5-HT and SP immunostaining, the tissues were fixed for 6–7 hr at room temperature. For choline acetyltransferase (ChAT) immunostaining, tissues were fixed in modified Zamboni's fixative (2% paraformaldehyde with saturated picric acid) overnight at 4°C. Tissues were washed for 6 × 10 min in phosphate buffer saline (PBS) and longitudinal muscle layer-myenteric plexus whole mount preparations were made. After pretreatment with 0.3% Triton X-100 solution (6 × 10 min) and preincubation in blocking solution (1% bovine serum albumin, 0.1% Triton-X 100 in PBS) for 20 min, whole mount preparations were incubated in primary antisera overnight. Primary antibodies used for immunolabeling are listed in Table 1. Whole mount preparations were washed in PBS and placed in secondary antisera labeled with fluorophore for 4–5 hr. The secondary antisera used are listed in Table 2. After removing the secondary antisera by washing in PBS for 6 × 10 min, whole mount preparations were mounted in Vectashield (Vector, Burlingame, CA). Immunostained tissues were viewed under Olympus BX 51 epifluorescence microscope connected to a digital camera (Olympus DP50), and digital photos were recorded with the help of an image analyzer program (AnalySIS). The digital images were further processed (background and contrast adjustments, assembling, and labeling of photographic tables) with the Adobe Photoshop 7.0 program. To check the specificity of both primary and secondary antibodies, three controls were made for the immunocytochemical labeling. Omission of primary antibodies in the immunolabeling procedures resulted no staining. Cross-reactivity of the noncorresponding primary and secondary antibodies was also checked and ruled out. To check the specificity of anti-5-HT2A receptor antibody, whole mount preparations were prepared as described earlier. However, they were incubated in the mixture of primary antisera (0.8 μg) and blocking peptide (8 μg). As a result of presence of blocking peptide, immunostaining could not be detected.

Table 1. Primary antibodies used for immunocytochemistry
Primary antibodyHostDilutionSupplier
anti-5-HTRabbit1:1,000Sigma-Aldrich, St. Louis, MO
anti-SP/NKA/NKBRat1:1,000BD PharMingen, Heidelberg, Germany
anti-ChATSheep1:1,000Chemicon, Temecula, CA
anti-5-HT2A receptorGoat1:500Santa Cruz Biotechnology, Heidelberg, Germany
Table 2. Secondary antibodies used for immunocytochemistry
Secondary antibodyHostDilutionSupplier
Texas red-conjugated anti-rabbit IgGDonkey1:200Jackson, West Grove, PA
FITC-conjugated anti-rabbit IgGDonkey1:200Jackson, West Grove, PA
FITC-conjugated anti-rat IgGDonkey1:200Jackson, West Grove, PA
Texas red-conjugated anti-rat IgGDonkey1:200Jackson, West Grove, PA
Texas red-conjugated anti-sheep IgGDonkey1:200Jackson, West Grove, PA
FITC-conjugated anti-goat IgGRabbit1:400Sigma-Aldrich, St. Louis, MO

To quantify the colocalization between tachykininergic/ChAT- and tachykininergic/5-HT2A-immunopositive elements, three ileal flat mounts were drawn from three animals. Cell bodies (n = 100/animal) were examined whether they were IR for either one of the markers or both. The counting of cell bodies was performed directly through the microscope.

Counting and plotting of SP-IR neurons were performed on whole mounts using semiautomatic computer-assisted image analysis software (Neurolucida).

In Vitro Pharmacological Experiments

Pieces of ileum were cut and the longitudinal muscle layer was removed as described earlier by Paton and Vizi (1969). The muscle strips were placed in organ baths filled with Krebs-Henseleit solution, which was oxygenated with mixture of 5% CO2 and 95% O2 and kept at 37°C.

Drugs and their concentrations used in experiments are listed in Table 3. They were applied directly into the organ bath and were washed out as soon as the peak contractions were reached. At the end of the measurements, 80 mM KCl was added into the baths to evoke the maximal contraction of muscle strips. Contractions of longitudinal muscle were measured by force-displacement transducers and were registered by ISOSYS program (Experimetria, Budapest, Hungary).

Table 3. Drugs used in pharmacological experiments
DrugsTargetConcentration (M)Supplier
SerotoninSerotonin receptors10−6Sigma-Aldrich, St. Louis, MO
TetrodotoxinSelective blocker of voltage-gated Na+ channel5 × 10−8Sigma-Aldrich, St. Louis, MO
HexametoniumNicotinic ACh receptor (antagonist)10−6Sigma-Aldrich, St. Louis, MO
AtropineMuscarinic ACh receptor (antagonist)10−6Sigma-Aldrich, St. Louis, MO
α-Methyl-5-hydroxytryptamine5-HT2 receptor (agonist)10−6Sigma-Aldrich, St. Louis, MO
Ketanserin5-HT2A receptor (antagonist)10−6Sigma-Aldrich, St. Louis, MO
N-Acetyl-L-tryptophan- 3,5(trifluoromethyl) benzyl esterNK1 receptor (antagonist)10−7Sigma-Aldrich, St. Louis, MO
SR 48968NK2 receptor (antagonist)5 × 10−7Sanofi-Synthelabo, Montpellier, France
SR 142801NK3 receptor (antagonist)2 × 10−7Sanofi-Synthelabo, Montpellier, France

Data Processing

Data are presented as means ± SEM (n = 6 or 7), statistical comparisons were made by using Wilcoxon's signed-rank test (two samples) and Quade-test for several related samples. A difference of P < 0.05 was considered statistically significant. Any group of results contains at most two data from the same animal. Each group of pretreated preparations was compared to a control group of muscle strips taken from the same animals.


Involvement of Intrinsic Excitatory Pathways in 5-HT-Induced Contractions

Our previous data have shown that blockade of muscarinic receptors (mAChR) does not fully abolish the 5-HT-induced contractions. At the same time, coadministration of atropine (ATR) and tetrodotoxin (TTX) caused a further decrease in the response of longitudinal muscle (Dénes et al.,2003); thus, it is likely that noncholinergic nerve element(s) participate in mediating the effect of 5-HT. To test this hypothesis, we applied an NK1 receptor antagonist (N-acetyl-L-tryptophan-3,5(trifluoromethyl)benzyl ester, 10−7 M), which decreased the 5-HT-induced (10−6 M) contractions significantly (Fig. 1a). Coapplication of NK1 receptor antagonist and ATR (10−6 M) also significantly reduced the contractions. Moreover, combined administration of NK1 receptor antagonist, ATR and TTX (5 × 10−8 M), further diminished the effect (Fig. 1b) showing that other types of neuronal receptors could also be involved in 5-HT-induced contractions besides NK1 and mAChR. The NK2 receptor blocker (SR 48968, 5 × 10−7 M) also strongly inhibited the 5-HT-evoked contractions (Fig. 1c) but the NK3 receptor antagonist (SR 142801, 2 × 10−7 M) was ineffective (not shown). Figure 1d demonstrates the effect of the NK1 receptor antagonist compared with the coapplication of NK1 antagonist and ATR. Figure 1e demonstrates that the effects of ATR and NK1 antagonist were not additive, as the latter did not have any effect on ATR resistant response.

Figure 1.

Tachykininergic and cholinergic components of the 5-HT-induced contractions. Both NK1 (a) and NK2 (c) receptors could mediate the effect. The final pathway involved at least one more nerve cell other than the cholinergic motorneuron [as evidenced with TTX preincubation in (b)]. After blocking NK1 receptors, ATR could further reduce 5-HT-evoked contractions, (d) whereas NK1 antagonist did not affect them having mACh receptors blocked by ATR (e).

The results suggest that 5-HT may act through more than one neuronal pathway. Therefore, we used hexamethonium (HEX, 10−6 M) to provide evidence for the role of cholinergic interneurons in 5-HT-evoked contractions. Figure 2 clearly shows that the contractile effect of 5-HT could be decreased by blocking nicotinic ACh receptors (nACh).

Figure 2.

A ganglionic ACh receptor blocker was able to decrease the 5-HT-induced contractions.

We reported previously that 5-HT2 and 5-HT4 receptors took part in the 5-HT-triggered contractions and that the response to alpha-methyl-5-hydroxytryptamine (α-m-5-HT, 10−6 M) was decreased after preincubation with TTX (Dénes et al.,2003). We aimed at a more refined pharmacological dissection of these pathways. Therefore, antagonists to block four receptors (ATR, HEX, NK1, and NK2) were examined on contractions stimulated by α-m-5-HT. Pretreatment with ATR and NK2 antagonists caused a significant reduction of contractions evoked by α-m-5-HT (Fig. 3a,b). Application of HEX and NK1 receptor antagonists also inhibited the effect of α-m-5-HT (Fig. 3c,d), although less potently than the above drugs.

Figure 3.

Role of 5-HT2 receptors. Blockade of both ACh (a, c) and tachykinin (b, d) receptors reduced contractions induced by 5-HT2 receptor agonist.

We tested ketanserin, which is widely used as a 5-HT2 receptor antagonist, to confirm the presence of 5-HT2 receptors. Ketanserin was able to prevent the 5-HT2 receptor agonist-induced contractions (not shown).

Tachykinin-Immunopositive Nerve Elements and Their Relationships With Other Neural Components of the Myenteric Plexus in the Rabbit Ileum

The pharmacological relationship we found between 5-HT-stimulated contractions and tachykinins prompted further search for morphological evidence of the possible connections. It is important to point out here that our antiserum for SP does not differentiate SP from NKA and NKB. For this reason, it is more proper to use the terms tachykinin(s) or tachykininergic for immunostained structures. With our antiserum, an extremely extensive meshwork was detected in all compartments of the myenteric plexus. Immunopositive fibers were seen on the surface of muscle cells and in the primary and secondary strands of the plexus (Fig. 4a). Numerous tachykinin-immunopositive neurons were found in most of the myenteric ganglia (Fig. 4b) as well as in the internodal strands (Fig. 4a). The immunopositive fibers often formed pericellular baskets around unlabelled cell bodies (Fig. 4c). Tachykinin-IR cells possessed angular or stellate forms and bore numerous, broad dendrites. Soon after emerging from the cell body, dendrites formed flat expansions, which often ended in short, fine fibers (Fig. 4d). Density of tachykinin-IR cells was calculated to be about 1,304 cells/cm2. The map of IR neurons shows that cells are evenly distributed throughout the myenteric plexus (Fig. 5).

Figure 4.

Tachykinin-immunoreactive elements in the myenteric plexus of rabbit ileum. A rich tachykinin-positive meshwork was found in the main and secondary strands (arrows) of the plexus as well as on the muscle layer (arrowheads) (a). Numerous tachykinin-IR neurons were seen (arrows) within the myenteric ganglia (b). Pericellular baskets (arrowheads) around unlabelled cell bodies indicate integrative function of tachykinin-containing neurons (c). An example is shown, tachykinin-IR neuron (arrow in d) could posses lamellar appendages (arrowheads) on its process (d). Scale bar: 60 μm in a, 25 μm in b, and 40 μm in c, d.

Figure 5.

Distribution of tachykinin-positive neurons within the myenteric plexus. Each black dot represents one tachykinin-IR neuron. Tachykinin-IR neurons were evenly distributed within the myenteric plexus. Scale bar: 800 μm.

Colocalization of tachykinins and ChAT was examined with double-label immunocytochemistry (Fig. 6a,b). Results showed that 77% ± 5% of ChAT-IR neurons were single-labeled and 22% ± 4% of ChAT-IR cell bodies also contained tachykinins in the myenteric plexus of rabbit ileum. We found that 50% ± 2% of tachykinin-immunopositive neurons were also IR for ChAT, whereas the other 50% of tachykinin-IR cells lacked ChAT-immunoreactivity.

Figure 6.

Double labeling of tachykinins and ChAT. Double-labeled cell was seen in the myenteric ganglia (arrows in a, b) but ChAT cells not labeled for tachykinins were also found (arrowhead in a, b). Also, tachykinin-IR cells were observed lacking ChAT-immunoreactivity (asterisk in a, b). Scale bar: 15 μm relating to all pictures.

Double-label immunocytochemistry revealed that 5-HT- and tachykinin-immunopositive fibers ran in the same nerve trunks but they never colocalized. Seeking for possible morphological relationship between the two systems, we checked if 5-HT-IR baskets could be found around SP-IR somata. All, what we found were a few tachykinin-IR neurons apposed by 5-HT-immunopositive varicosities (Fig. 7a,b). Finally, results of colabeling with 5-HT and ChAT antibodies also showed only a few 5-HT-IR varicosities running closely to ChAT-immunopositive nerve cells (Fig. 7c,d).

Figure 7.

Double labeling of 5-HT/tachykinin and 5-HT/ACh. One example of those few cases when fibers with strong 5-HT uptake capacity (single arrowheads in (a) surrounded tachykinin-positive cell (asterisk in b). Arrows and double arrowheads point to similarly positioned, but not identical fibers. The lower panel shows that the varicose 5-HT-immunopositive fibers (arrowheads in c) run close to ChAT-immunoreactive (asterisks in d) cell bodies. Scale bar: 20 μm relating to all pictures.

Immunostaining with anti-5-HT2A receptor antibody revealed a tremendous amount of varicose fibers in the primary, secondary strands of myenteric plexus as well as on the surface of smooth muscle layer (Fig. 8a,c,e). Many cell bodies were also labeled in the myenteric ganglia; however, many of them were hardly visible because they were covered by immunopositive fibers. The shape and size of the immunopositive cell bodies varied in a wide range. We could detect large, round Dogiel type II-like neurons as well as small ones (Fig. 8f,g). 5-HT2A-IR fibers located in the secondary and tertiary plexus bore numerous varicosities (Fig. 8d), whereas processes emerging from the cell bodies were thick and smooth lacking lamellar expansions or varicosities (Fig. 8g). Most importantly, our 5-HT2A receptor- and tachykinin-IR colocalization study proved that tachykinin-containing fibers and neurons did express 5-HT2A receptors. The overlap was clearly observed in the secondary (Fig. 8a,b) and tertiary plexus (Fig. 8c,d). In case of fibers running in the ganglia and primary strands, the colocalization was difficult to establish. Quantitative data showed that 13% ± 4% of tachykinin-IR neurons were also labeled with 5-HT2A receptor antibody and 19% ± 4% of 5-HT2A receptor-IR cell bodies also contained tachykinins in the myenteric plexus of rabbit ileum (Fig. 8e,f).

Figure 8.

Expression of 5-HT2A receptors in rabbit myenteric plexus. In rabbit myenteric plexus, a varicose fiber network was immunopositive for anti-5-HT2A and tachykinin antibody (arrows in a and b). Immunoreactive processes emerging from large cell bodies or running in the strands were smooth (arrowheads in a and g). Varicose fibers innervating the smooth muscle layer were immunoreactive for both tachykinins and 5-HT2A receptors (c and d). Some enteric neurons were double-labeled (arrows in e and f) but not all the tachykinin-immunopositive neurons possessed 5-HT2A receptors (single asterisk) and neurons expressing 5-HT2A receptors lacked tachykinin-immunoreactivity (double asterisks). Scale bar: 25 μm relating to all pictures.


The major goals of our experiments were to dissect the relationship between the serotoninergic and tachykininergic systems in the ileum and to assess the effect of 5-HT, a common transmitter in ileal myenteric system. We integrated morphological and pharmacological data by performing immunostainings and in vitro pharmacological experiments in rabbit ileal muscle strip model.

A particular aim of our research was to study the tachykininergic system, which is also involved in the organization of gut motility. The main source of tachykinins is the intrinsic enteric neurons that supply the ganglia and the smooth muscle (Shuttleworth and Keef,1995). In mammalian species, tachykinin-immunoreactivity was detected in Dogiel type I neurons with either short filamentous processes or many lamellar dendrites and in Dogiel II cells (Ekblad et al.,1987; Sang and Young,1996,1998; Brookes,2001; Shimizu et al.,2008). Functionally, tachykinins are released from enteric excitatory motorneurons innervating the smooth muscle of gut, intrinsic primary afferent neurons, and last but not least, orally and anally directed interneurons (Ekblad et al.,1987; Brookes et al.,1991,1992,1997; Johnson et al.,1996,1998). Tachykinins utilized as transmitters from enteric neurons target numerous enteric elements through NK1, NK2, or NK3 receptors. NK2 receptors occur primarily on muscle, whereas NK3 receptors are involved in neuroneuronal transmission (Johnson et al.,1996,1998; Portbury et al.,1996a; Jenkinson et al.,1999,2000; Lecci et al.,2002). NK1 receptors have been shown acting on both muscle and enteric neurons. More precisely, functional classes of neurons that expressed NK1 receptors were NOS-IR inhibitory motorneurons, ACh/SP-IR excitatory motorneurons, secretomotor neurons, and ChAT/calbindin-IR sensory neurons (Portbury et al.,1996b; Johnson et al.,1998; Lomax et al.,1998; Lecci et al.,2002; Thornton and Bornstein,2002).

Data that indicate functional relationship between 5-HT and tachykininergic system are rare and to date none has been obtained in rabbit. Work of Bucheit et al. (1985) on guinea pig ileum provided data that SP antagonist reduced the effect of 5-HT; however, they did not specify which tachykinin receptor was involved. Also, Briejer and Schuurkes (1996) found that 5-HT due to 5-HT3 and 5-HT4 receptors had stimulatory effects in the guinea pig colon and this action was mediated through activation of the cholinergic and tachykininergic neurons.

A Possible Enteric Circuit Underlying Serotonin–Tachykinin Interaction in Rabbit Ileum

In our previous article, data had been published about the serotoninergic system in rabbit (Dénes et al.,2003). Briefly, the 5-HT synthesizing enzyme could be detected in the myenteric plexus but before revealing a rich varicose 5-HT-IR fiber meshwork, we had to incubate the ileum with 5-HT. After preincubation with 5-HT, an extensive fiber meshwork was seen in the myenteric plexus forming numerous pericellular baskets as well as running to the muscle layer. No 5-HT-IR neurons were observed in the rabbit myenteric plexus; however, 5-HT stimulated movements of the longitudinal muscle layer by direct and indirect ways in the rabbit ileum. The indirect effect of 5-HT was partially elicited through cholinergic pathway via 5-HT2 receptors, whereas the direct action of 5-HT was mediated by 5-HT4 receptors (Dénes et al.,2003). In this study, a more refined analysis of the effect of 5-HT was performed.

Tachykinins are thought to be major excitatoric transmitters in enteric neurons (Costa et al.,1985; Grider,1989). Therefore, we used NK1, NK2, and NK3 receptor antagonists to analyze contractions evoked by 5-HT and 5-HT2 receptor agonist. Considering that the natural ligand of NK1 receptors is SP whereas NK2 receptors are preferred by NKA (Holzer and Holzer-Petsche,1997), we assume that 5-HT stimulates both SP and NKA release through 5-HT2 receptors. From the consecutive and reversed administration of drugs, it is clear that ACh has a prominent role in the residual contraction after application of NK1 antagonist. At the same time, treatment with ATR could prevent the effect of the 5-HT-evoked tachykinin release. It is reasonable to assume that 5-HT could release SP from interneurons rather than from motor neurons. Our result of with HEX treatments indicates that cholinergic interneurons could be also involved in this circuit.

To supplement the data obtained by in vitro pharmacological experiments, immunostainings were also performed. First, we found a massive tachykinin-IR system in the myenteric plexus. We show a strong immunopositive network in the smooth muscle layer, neurons, and pericellular baskets around both unlabelled and tachykinin-immunopositive somata. The finding suggests that tachykinins could participate in motor as well as integrative functions in the rabbit ileum. By mapping SP-IR neurons, we established that the cells were distributed evenly. The proportion of tachykinin-IR neurons in rabbit ileum is surprisingly high compared with rodents. The total neuron density of ileal myenteric plexus in rabbit was established to be ∼2,500 cells/cm2 (Gábriel et al.,1998). We calculated that ∼52% of all myenteric neurons synthesize tachykinins indicating that these peptides must have dominant roles in organizing ileal functions. In guinea pig, the SP-IR neuron population amounted to 23% of all neurons of the ileal myenteric plexus (Costa et al.,1996). In mice, 29% of the total neuron population in the myenteric plexus was found SP-immunopositive (Sang and Young,1996).

Seeking for anatomical evidence for interconnections among serotoninergic, cholinergic, and tachykinin-containing system, double-label studies were performed. In most mammalian species, both SP and NKA appeared to be involved in neuronal transmission in the enteric nervous system and they coexist with ACh in the same excitatory neurons (Grider,1989; Lippi et al.,1998; Maggi,2000). The proportion of coexistence of these markers, however, varies among species. In guinea pig, all SP-IR neurons were also IR for ChAT, whereas 25–45% of all ChAT-IR cells were SP-immunopositive (Steele et al.,1991). In mice, about 50% of ChAT-IR neuron population contained SP, and 80% of SP-IR cells contained also ChAT. Our double-label experiments in rabbit revealed 20% of ChAT-IR cells containing also tachykinins and ∼50% tachykinin-immunopositive cells were IR for ChAT.

We did not find 5-HT-IR pericellular baskets around either ChAT or tachykinin-IR cell bodies. Those few ChAT or tachykinin-IR cells that were apposed by 5-HT-IR boutons were not prominent. Synaptology and thus the function of serotoninergic fibers emerging from serotoninergic descending interneurons has not been clarified yet since the reports that investigated the problem published different results (Erde et al.,1985; Young and Furness,1995; Neal and Bornstein,2007). Erde et al. (1985) reported that mainly type II/AH neurons considered sensory neurons were contacted by serotoninergic terminals. According to Young and Furness (1995), Dogiel type II neurons received a few serotonin inputs and none of them formed synapses, whereas the highest number of serotonin terminals were related to Dogiel type I neurons. A more recent works provided precisely quantified data claiming that anatomical connection between 5-HT-IR varicosities and the longitudinal muscle motor neurons was not significant. In contrast, all excitatory circular muscle motor neurons and some ascending interneurons, which are tachykininergic, were encircled by 5-HT-IR varicosities (Brookes at al.,1997; Neal and Bornstein,2007). However, it is important to note that having close contacts with 5-HT-IR varicosities do not necessarily mean synaptic specializations (Young and Furness,1995). However, this is not a question in case of rabbit considering the small number of cholinergic and tachykinin-containing neurons, which were apposed by 5-HT-IR fibers. Therefore, evidence of morphological basis for 5-HT/tachykinin and 5-HT/ACh interaction could not be found in rabbit ileal myenteric plexus. Clearly, simple light microscopy used in our study is not an efficient tool to reveal synaptic inputs between two structures. This would require ultrastructural analysis. Besides synaptic transmission, intercellular communication can also occur by volume transmission, which makes communication possible between nerve elements lacking synaptic specializations (Zoli et al.,1999; Sykova,2004; Vizi et al.,2004). 5-HT may be released by nonjunctional varicosities and can diffuse to 5-HT receptors of both neurons and smooth muscle cells. 5-HT might also be secreted by endocrine cells and act as a hormone or through activation of local neural pathways. Numerous studies revealed that 5-HT was released in response to pressure, mucosal stimuli from the intestinal mucosa and evoked peristaltic or secretory reflexes (Kirchgessner et al.,1992; Grider et al.,1996; Zhu et al.,2001; Bertrand,2007; Patel et al.,2007).

The pharmacological experiments showing indirect effects of 5-HT through 5-HT2 receptors were also confirmed by immunocytochemistry. We could detect numerous neurons as well as a varicose fiber network that expressed 5-HT2A receptors in the myenteric plexus. Our colocalization studies provided evidence that a small population of tachykininergic neurons possesses 5-HT2A receptors that could be either excitatory motorneurons or interneurons. In addition to this population, there were tachykinin-IR neurons lacking 5-HT2A receptors and many 5-HT2A receptor expressing cells, in fact, did not show tachykinin-immunopositivity.

By comparing our observations with results obtained in other species, we can state that the action of 5-HT on ileal motility in rabbit shares both similarities and dissimilarities to other species like guinea pig, rat, and mouse that are close to rabbits in terms of diet and evolution.

Because the first description that 5-HT potently affects intestinal movements (Gaddum,1953), actions of 5-HT have been investigated in a wide range of animals and from different aspects. Most of the data have been provided from guinea pig small intestine. Two different 5-HT receptors can mediate contractile responses named 5-HT3 and 5-HT4 receptors located on neuronal components (Craig and Clarke,1990; Craig et al.,1990; Eglen et al.,1990; Taniyama et al.,1991). 5-HT acting at 5-HT3 receptors could evoke ascending as well as descending excitatory reflexes (Yuan et al.,1994) through activation of cholinergic motor neurons (Fox and Morton,1990; Monro et al.,2002). 5-HT4 receptors can also mediate contractile effect, increase peristaltic reflex sensitivity (Craig and Clarke,1990,1991; Buchheit and Buhl,1991; Kilbinger and Wolf,1992; Kilbinger et al.,1995; Galligan et al.,2003) like 5-HT3 receptors do. Besides the cholinergic system, evidence was reported that tachykininergic neurons also had roles in contractions induced by 5-HT (Buchheit et al.,1985; Ramirez et al.,1994). In mouse ileum, effect of 5-HT was mediated through neuronally located 5-HT3 receptors (Tuladhar et al.,2000; Chetty et al.,2006). Data obtained in rat ileum are rather controversial. Some studies claimed that 5-HT2A and 5-HT1 receptors located on the smooth muscle were involved in 5-HT-induced contractions, and neither TTX nor atropin caused inhibition suggesting a direct effect on rat ileal muscle (Briejer et al.,1997; Yamano et al.,1997). Meanwhile, the role of 5-HT3 receptors acting through neuronal activation in rat ileum was also proved (Chetty et al.,2006). The latter study also suggested that tachykinins were not mediating 5-HT3-induced contractions.

The main conclusions of our study, therefore, were as follows: (i) in rabbit, 5-HT could stimulate contractions on both indirect and direct ways, whereas in guinea pig, rat, and mouse 5-HT has only one way to act, either direct or indirect. (ii) Cholinergic and tachykinin-containing motorneurons and also interneurons are participating in muscle stimulation in rabbit (Fox and Morton,1990; Briejer and Schuurkes,1996). (iii) In rabbit, movements of longitudinal muscle strips could be evoked potentially via two kinds of 5-HT receptor, 5-HT2 and 5-HT4, whereas 5-HT3, 5-HT4 receptors take parts in mediation of 5-HT-evoked contractions in guinea pig, 5-HT2 receptors in rat and 5-HT3 receptors in mouse, respectively.