SEARCH

SEARCH BY CITATION

Keywords:

  • 5-hydroxytryptamine;
  • enteric co-innervation;
  • mast cell;
  • myenteric neuron;
  • neuromuscular junction;
  • nucleus ambiguus

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

Background  Serotonin is a major transmitter in the gastrointestinal tract, but little is known about the serotonergic system in the esophagus.

Methods  The aim of this study was to use multilabel immunofluorescence to characterize serotonin-positive nerve cell bodies and fibers and their relationship with other neuronal and non-neuronal elements in the mouse esophagus. Antibodies against serotonin, vesicular acetylcholine transporter (VAChT), choline acetyltransferase (ChAT), protein gene product 9.5 (PGP 9.5), and α-bungarotoxin (α-BT), were used.

Key Results  Serotonin-containing perikarya represented ∼10% of all PGP 9.5-positive myenteric neurons. Serotonin-positive varicose nerve fibers were found in the lamina muscularis mucosae and present on ∼13% of α-BT–labeled motor endplates in addition to VAChT-immunoreactive motor terminals. As ChAT-positive neurons of the compact formation of the nucleus ambiguus were negative for serotonin, serotonin-positive varicosities on motor endplates are presumed to be of enteric origin. On the other hand, cholinergic ambiguus neurons were densely supplied with serotonin-positive varicosities. The tela submucosa and tunica adventitia contained large numbers of serotonin-positive mast cells, a few of which were in close association with serotonin-positive nerve fibers.

Conclusions & Inferences  The mouse esophagus is endowed with a rich serotonin-positive intrinsic innervation, including enteric co-innervation of striated muscles. Serotonin may modulate vagal motor innervation of esophageal-striated muscles not only at the central level via projections of the raphe nuclei to the nucleus ambiguus but also at the peripheral level via enteric co-innervation. In addition, mast cells represent a non-neuronal source of serotonin, being involved in neuroimmune processes.


Abbreviations:
BSA

bovine serum albumin

α-BT

α-bungarotoxin

ChAT

choline acetyltransferase

5-HT

5-hydroxytryptamine, serotonin

Lmm

lamina muscularis mucosae

PGP 9.5

protein gene product 9.5

RT

room temperature

Ta

tunica adventitia

TBS

Tris-buffered saline

Tm

tunica muscularis

TpH

tryptophan hydroxylase

Ts

tela submucosa

VAChT

vesicular acetylcholine transporter.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

Serotonin (5-hydroxytryptamine, 5-HT) is used for signaling in diverse systems both in the brain and in the periphery. About 5% of the serotonin in the body is found in the central nervous system, and about 95% occurs in the gastrointestinal tract. The major store in the gastrointestinal tract, about 90%, is localized in the enterochromaffin subtype of enteroendocrine cells of the mucosal layer, and the remaining 10% is present in serotonergic neurons of the myenteric plexus (for review see Refs.1–7). Serotonin synthesis in mammals is initiated by two distinct tryptophan hydroxylases (TpH-1 and -2). TpH-1 is responsible for serotonin production in enterochromaffin cells, and TpH-2 operates in enteric neurons and the central nervous system.8–10 The serotonin reuptake transporter is the primary molecule responsible for inactivating serotonin in the central nervous system and gut (for review see Refs.2, 11). Serotonin exerts its effects by binding to cell surface receptors that can be classified into seven subtypes, 5-HT1 through 5-HT7 (for review see Refs.2, 12–17). The existence of several 5-HT receptor subtypes has enabled the development of selective drugs for treatment of a number of gastrointestinal motility disorders. Current examples include the 5-HT3 antagonists alosetron and granisetron, the 5-HT4 partial agonist tegaserod, and the 5-HT4 agonist and 5-HT3 antagonist mosapride, which has been approved for treatment of, for example, diarrhea- and constipation-predominant irritable bowel syndromes, chemotherapy-induced nausea and vomiting, and functional dyspepsia.2,12,18,19 Pharmacological clinical investigations with agents such as the 5-HT1 agonist sumatriptan, mosapride, and the 5-HT4 agonist cisapride have further indicated that serotonin is involved in the control of esophageal motility.20–22

As the identification of serotonin as a transmitter in the mouse gut,23 the presence of serotonin has been demonstrated radioautographically and immunohistochemically in the gastrointestinal tract of various vertebrates including humans.24–33 However, morphological investigations of serotonin distribution in the gastrointestinal tract have neglected the esophagus almost completely. In contrast to the adequate pharmacological and functional studies, little is known about the morphological organization of serotonergic neurons and other serotonin-containing elements in the esophagus.

Therefore, the aim of this study was to immunohistochemically identify the serotonin-containing structures in the esophagus and to facilitate the understanding of functional roles of serotonin in this organ. As esophageal-striated muscle receives dual innervation both from vagal motor efferents originating in the brain stem and varicose enteric nerve fibers originating in myenteric neurons (for review see Ref. 34), the question of whether serotonin-positive nerve fibers might also be involved in this so-called enteric co-innervation is much interesting.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

Tissue sampling, fixation, and processing

Thirty-five mice (C57/Bl6) of either sex, aged 12–15 weeks were obtained from Charles River (Sulzfeld, Germany) and euthanized with an overdose of sodium thiopental (250 mg kg−1 i.p.). The European Communities Council Directive and animal welfare protocols approved by the local government were followed.

For cryosections of normally treated tissue, mice were perfused through the ascending aorta with a Ringer prewash, followed by 4% phosphate-buffered formaldehyde (pH 7.4). The cervical, upper and lower thoracic, and abdominal portions of the esophagi, pieces of jejunum, and brain stems were dissected, postfixed in 4% phosphate-buffered formaldehyde for 5 h at 4 °C, and rinsed in phosphate buffer overnight. They were subsequently immersed in 12–14% phosphate-buffered sucrose for cryoprotection overnight, mounted in tissue-embedding medium (Slee medical GmbH, Mainz, Germany), and frozen in liquid nitrogen-cooled isopentane.

To enhance serotonin immunoreactivity on cryosections, the length of esophageal portions of some mice were measured after opening of the thorax and abdomen for subsequent calculation of the enteric co-innervation rate, and the entire esophagi, jejuna, and brain stems were dissected and incubated in modified oxygenated (95% O2, 5% CO2) Krebs solution at 37 °C for 30 min according to Krebs and Henseleit 35 and Costa and colleagues.25 The modified Krebs solution consisted of NaCl 117 mmol L−1, KCl 5 mmol L−1, CaCl2 2.5 mmol L−1, MgSO4 1.2 mmol L−1, NaHCO3 25 mmol L−1, NaH2PO4 1.2 mmol L−1, and glucose 10 mmol L−1. After immersion fixation for 5 h in 4% phosphate-buffered formaldehyde (pH 7.4) at 4 °C and washing in phosphate buffer overnight, tissue samples were rinsed in 12–14% phosphate-buffered sucrose at 4 °C for cryoprotection overnight. Esophagi were divided into cervical, upper and lower thoracic, and abdominal portions and jejuna into pieces of about 1 cm in length. The samples were mounted on tissue-embedding medium (Slee medical GmbH, Mainz, Germany) and frozen in liquid nitrogen-cooled isopentane.

To enhance serotonin immunoreactivity on whole mount preparations in some mice, each esophagus was distended by insertion of plastic tubing with an outer diameter of 2 mm, and dissected and additionally longitudinally stretched by pinning of the cranial and caudal ends onto a cork plate. Subsequently, esophagi were incubated in modified oxygenated Krebs solution at 37 °C for 30 min, fixed by immersion for 5 h in 4% phosphate-buffered formaldehyde (pH 7.4) at 4 °C, and washed in phosphate buffer overnight. Esophagi were opened longitudinally, and the mucosa was peeled off.

Immunocytochemistry

Double labeling of serotonin and α-bungarotoxin  For qualitative analysis of serotonin immunoreactivity in the esophagus, six whole mount preparations were preincubated for 2 h with a solution mixed with 1% bovine serum albumin (BSA), 5% normal donkey serum, 0.5% Triton-X-100, and 0.05% thimerosal at room temperature (RT). After a rinse in Tris-buffered saline (TBS; 0.05 mol L−1, pH 7.4) for 10 min, the whole mounts were incubated with the primary antiserum against serotonin (Table 1) for 3 days at 4 °C, washed in TBS for 1 day at 4 °C, and subsequently incubated with the secondary antiserum (Table 1) for 4 h at RT. After being rinsed in TBS for 1 day, whole mounts were incubated with α-bungarotoxin (α-BT; 1 : 4000; Table 1) for 1 h at RT, washed in TBS for 1 h, and coverslipped with Vectashield (Linaris, Wertheim, Germany).

Table 1.   Overview and characterization of antisera and toxins used for double and triple staining
Primary antiseraHost speciesDilutionSource (catalog number)
SerotoninRabbit1 : 1000Immunostar, Hudson, WI, USA (20080)
VAChTGoat1 : 200Biotrend, Köln, Germany (0030-5009)
PGP 9.5Guinea pig1 : 500Millipore, Billerica, MA, USA (AB5898)
ChATGoat1 : 40–1 : 60Millipore, Billerica, MA, USA (AB144P)
Secondary antisera and toxinsDilutionSource (catalog number)
Donkey anti-rabbit Alexa-4881 : 1000Molecular Probes, Eugene, OR, USA (A-21206)
Donkey anti-goat Alexa-5551 : 1000Molecular Probes, Eugene, OR, USA (A-21432)
Donkey anti-goat Alexa-6471 : 1000Molecular Probes, Eugene, OR, USA (A-21447)
Donkey anti-guinea pig Cy31 : 1000Dianova, Hamburg, Germany (706-165-148)
α-bungarotoxin Alexa-5941 : 1000 or 1 : 4000Molecular Probes, Eugene, OR, USA (B-13423)

For qualitative analysis of serotonin immunoreactivity and quantitative analysis of co-innervation rates, six esophagi were longitudinally cut at 12–14 μm on a cryostat (Leica CM1900, Wetzlar, Germany), mounted on poly-L-lysine-coated slides, and air dried for 1 h. Sections were preincubated with a solution mixed with 1% BSA, 5% normal donkey serum, and 0.5% Triton-X-100 for 1 h at RT. After a rinse in TBS for 5 min, samples were incubated with the primary antiserum against serotonin (Table 1) overnight at RT, washed in TBS for 15 min at 4 °C, and subsequently incubated with the secondary antiserum (Table 1) for 1 h at RT. After being rinsed in TBS, sections were incubated with α-BT (1 : 1000; Table 1) for 20 min at RT, subsequently rinsed in TBS for 15 min, and coverslipped with TBS–glycerol (1 : 1, pH 8.6).

Co-innervation rates were estimated by evaluating 250 α-BT–positive motor endplates in each cervical, upper and lower thoracic, and abdominal portion of six esophagi (i.e., n = 1000 for each esophagus). Mean values (percentages) including standard error of the mean were calculated for the cervical, thoracic and abdominal portions, and for the entire esophagus.

Triple labeling of serotonin, α-bungarotoxin, and the vesicular acetylcholine transporter  For determination of the spatial relationships of vagal and enteric nerve terminals on motor endplates, four whole mount preparations were incubated for double immunofluorescence of serotonin and the vesicular acetylcholine transporter (VAChT) after pre-incubation, as described above. Binding sites were visualized using Alexa-488–conjugated donkey anti-rabbit IgG and Alexa-647–conjugated donkey anti-goat IgG. For primary and secondary antibody antisera, see Table 1. After being rinsed in TBS, the whole mount preparations were incubated with α-BT (1 : 1000) for 1 h at RT, and coverslipped with Vectashield (Linaris, Wertheim, Germany).

Double labeling of serotonin and protein gene product 9.5  For qualitative and quantitative analysis of enteric neurons in the esophagus, six whole mount preparations were immunostained for serotonin and the pan-neuronal marker protein gene product 9.5 (PGP 9.5), as described above. For primary and secondary antisera, see Table 1. After being washed in TBS, the whole mount preparations were coverslipped with Vectashield (Linaris, Wertheim, Germany).

For quantitative analysis, PGP 9.5- and serotonin-positive neurons were counted by focusing through the full thickness of the myenteric plexus in 25 visual fields in each of the cervical, upper and lower thoracic, and abdominal portions (100 visual fields in each of six esophagi; 20 × objective lens). Myenteric neurons were considered to be PGP 9.5- and serotonin-positive if presenting homogeneous staining of the cytoplasm and no staining of the nucleus. Mean values (percentages) including standard error of the mean were calculated for the cervical, thoracic and abdominal portions, and for the entire esophagus.

Double labeling of serotonin and toluidine blue  Mast cells are secretory cells defined by their toluidine blue-stained metachromatic granules.36,37 We used this method as a counterstain to identify serotonin-positive cells as mast cells. Cryosections of all portions of six esophagi were stained for serotonin as described above. Presumptive serotonin-positive mast cells were photographed, coverslips were removed, and the tissue was incubated in 0.1% toluidine blue (VWR International GmbH, Darmstadt, Germany) dissolved with 60% ethanol for 15 min.37 Sections were washed in distilled water, coverslipped with Kaiser’s glycerol gelatin (Merck, Darmstadt, Germany), and photographed again.

Double labeling of serotonin and choline acetyltransferase  Brain stems of eight mice (four without and four with pre-incubation in modified Krebs solution) were cut serially at 12–14 μm on a cryostat (Leica CM1900, Wetzlar, Germany), mounted on poly-L-lysine–coated slides, and air dried for 1 h. Sections were stained for serotonin and choline acetyltransferase (ChAT), as described above. Binding sites were visualized using Alexa-488–conjugated donkey anti-rabbit IgG and Alexa-555–conjugated donkey anti-goat IgG. For primary and secondary antisera, see Table 1. Sections were washed in TBS and coverslipped with TBS–glycerol (1 : 1, pH 8.6).

Control experiments  The specificity of the immunocytochemical reactions was assessed by replacing the primary antibody with TBS or rabbit normal serum as negative controls or by pre-absorbing the antibody against serotonin, VAChT, and ChAT with its respective antigen (serotonin: Immunostar, Acris Antibodies GmbH, Herford, Germany; 2 μg antigen 100 μL−1 antibody solution; VAChT: Phoenix, Mountain View, CA, USA; 10 μg antigen 100 μL−1antibody solution; ChAT: Millipore, Billerica, MA, USA; 0.5 μg antigen 100 μL−1 antibody solution). In addition, as positive controls, cryosections of the jejunum and brain stem were incubated with antibodies against serotonin as described above.

Microscopy

For conventional light and fluorescence microscopy, a Leica Aristoplan microscope (Leica, Wetzlar, Germany) equipped with appropriate filter settings and a CCD camera (Visitron Systems, Puchheim, Germany) was used. Confocal images were acquired using a confocal laser scanning system (Nikon Eclipse E1000-M; Nikon Digital Eclipse C1 with software EZ-C1 3.91; Tokyo, Japan). Laser configuration was as follows: 488 nm Argon Laser, 543 nm Helium-Neon Laser (both from Melles Griot Inc., Carlsbad, CA, USA), and 638 nm Diode-Laser (Coherent, Santa Clara, CA, USA). A 20 × dry and a 60 × oil immersion objective lenses (numerical aperture 0.75 and 1.40, respectively) were used combined with an electronic zoom factor of up to 2.0. To obtain all-in-focus images, up to 10 optical sections were taken at z-intervals of 1 μm, and were electronically superimposed. Images were processed using Nikon Free Viewer software (EZ-C1 3.60), adjusted for brightness and contrast using Adobe Photoshop CS4 (Adobe Systems, San Jose, CA, USA), and laid out using CorelDraw X4 software (Corel, Ottawa, ON, Canada).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

In initial experiments with normally formaldehyde-fixed tissue, the serotonin immunoreactivity of nerve cell bodies and fibers was weak in the esophagus and jejunum, but strong in the brain stem. Pretreatment of tissue samples with Krebs solution resulted in enhanced intensity of the immunohistochemical reaction in the esophagus and jejunum, but in little change or even a slight decrease in immunoreactivity in the brain stem. Therefore, in all subsequent experiments, the distribution of serotonin-positive structures in peripheral tissues was examined using Krebs solution pretreatment, whereas the brain stem was studied both with and without pretreatment. In the esophagus, serotonin immunoreactivity could be detected in nerve fibers in the lamina propria mucosae and tela submucosa (Fig. 1A), between smooth muscle cells of the lamina muscularis mucosae (Fig. 1B), around many blood vessels throughout the esophageal wall (Fig. 1C), between striated muscle cells of the tunica muscularis (Fig. 1A), and on α-BT–stained motor endplates (Fig. 1D–G). Serotonin immunostaining was additionally visible in myenteric neurons (Fig. 2) and in mast cells located mainly in the tela submucosa and tunica adventitia (Fig. 3). No serotonin staining could be observed in the squamous epithelium of the esophagus. In the brain stem, serotonin-immunoreactive nerve cell bodies were clearly present in selected medullary raphe nuclei, but not detected in the compact formation of the nucleus ambiguus (Fig. 4).

image

Figure 1.  Serotonin-immunoreactive nerve fibers in mouse esophagus. Numerous nerve fibers positively stained for serotonin could be demonstrated in the tela submucosa (Ts; A, B: short arrows), tunica muscularis (Tm; A: long arrow), lamina muscularis mucosae (Lmm; B: long arrow), and surrounding blood vessels (C: long arrow). In some cases, serotonin-positive nerve fibers on blood vessels were opposed to serotonin-immunoreactive mast cells (C: short arrows). Varicose nerve fibers immunoreactive for serotonin (D, E: short arrows) were in close contact with α-BT–positive motor endplates (D, E: arrowheads) or with coarse VAChT-positive vagal nerve terminals (E: long arrow). They were intermingled, but completely separated from vagal nerve terminals on motor endplates (E–G). Confocal images from cryosection (A) and whole mount preparations (B–G) of thoracic (A–C, E–G) and abdominal (D) portions; z shows the superimposed confocal optical sections; E shows a confocal image superimposed with triple staining for serotonin (F), VAChT (G), and α-BT (E). Bars: 50 μm in A, C; 20 μm in B, D; 25 μm in E–G.

Download figure to PowerPoint

image

Figure 2.  Serotonin-immunoreactive nerve cell bodies and nerve fibers in myenteric ganglia in mouse esophagus. Double labeling for serotonin (A, C, E) and PGP 9.5 (B, D, F) on whole mount preparations showed serotonin/PGP 9.5-positive (A–D: long arrows) and only PGP 9.5-positive (E, F: arrows) myenteric neurons. Serotonin staining in majority of nerve cell bodies was weak (A, C: long arrows), but only a few strongly stained serotonin-positive neurons were detectable (C, D: short arrows). Serotonin-positive and -negative nerve cell bodies were mostly encircled by baskets of serotonin-positive varicose nerve fibers (A, C, E). Confocal images of whole mount preparations of the abdominal portion; z shows the superimposed confocal optical sections. Bar: 50 μm in A, also applies to B–F.

Download figure to PowerPoint

image

Figure 3.  Serotonin-immunoreactive mast cells in mouse esophagus. Serotonin-positive cells in the tela submucosa (Ts; A, B) and tunica adventitia (Ta; C, D) on cryostat sections counterstained for toluidine blue (B, D). The violet-blue color of the metachromatic granules after toluidine blue staining allowed serotonin-positive cells to identify as mast cells (A–D: long arrows). Note the close apposition between serotonin-positive nerve fibers (A, C: short arrows) and mast cells. Tm, tunica muscularis. Conventional microscopic images of the thoracic portion. Bar: 50 μm in A.

Download figure to PowerPoint

image

Figure 4.  Serotonin-immunoreactive nerve fibers and nerve cell bodies in mouse medulla. In cryostat sections of the medulla, serotonin-positive nerve cell bodies were present in the nucleus raphe pallidus (A: arrows) and nucleus raphe obscurus (B: arrows), but not in the compact formation of the nucleus ambiguus (C, D). In A–D, the dorsal is the upper side of images. Double labeling revealed that in the nucleus ambiguus, numerous serotonin-immunoreactive nerve fibers were detectable (C), which encircled ChAT-positive neurons (D; boxed area at a higher magnification in E), forming close associations on their cell bodies (E: arrowhead). Confocal images; z shows the superimposed confocal optical sections. Bars: 50 μm in A–D; 10 μm in E.

Download figure to PowerPoint

Motor endplates: enteric co-innervation and co-innervation rate

In both cryosections and whole mount preparations, α-BT–stained motor endplates were distributed dispersedly over both striated muscle layers. Their density increased slightly from cervix to abdomen (for quantitative data see also Ref.38). The area of motor endplates was generally ovoid in shape, and in some cases, it was branched. Each motor endplate was composed of spots of intense α-BT staining intermingled with unstained areas (Fig. 1D,E). Double labeling of serotonin and α-BT revealed serotonin-positive nerve fibers on motor endplates of striated muscle fibers in all portions of the organ (Fig. 1D,E). On average, 13.0 ± 2.1% of all motor endplates (n = 6000 in six animals, age 3 months) calculated from the cervical, thoracic, and abdominal portions were co-innervated by serotonin-positive nerve terminals. Co-innervation rates increased from 7.7 ± 1.4% in cervical to 12.8 ± 2.2% in thoracic to 18.6 ± 3.7% in abdominal esophagus. Triple labeling for serotonin/VAChT/α-BT on esophagus whole mounts showed intermingling of varicose serotonin-positive and cholinergic vagal motor terminals on α-BT-positive motor endplates (Fig. 1E). VAChT-immunoreactive nerve terminals matched almost entirely with the α-BT–stained areas of the postsynaptic muscle fiber membrane. Serotonin-positive boutons were closely located to, but completely separated from VAChT-positive nerve terminals (Fig. 1E–G).

Serotonin-immunoreactive myenteric neurons

To elucidate the origin of serotonin-positive nerve fibers on motor endplates, we focused first on enteric ganglia in whole mount preparations of the esophagus. Serotonin-immunoreactive nerve cell bodies and fibers were observed in ganglia of the myenteric plexus in the entire esophagus (Fig. 2A–D). The majority of reactive cell bodies were faintly stained, and only a few of them were intensely stained, with being small in size (Fig. 2C,D). On average, 10.3 ± 1.9% of all PGP 9.5-positive neurons (n = 2661 in six animals) was stained for serotonin. Proportions of serotonin-positive neurons ranged from 8.7 ± 1.9% in the cervical to 11.1 ± 2.1% in the thoracic to 10.6 ± 1.8% in the abdominal esophagus. Myenteric ganglia frequently contained one or two serotonin-positive neurons (Fig. 2A,C) and occasionally three to four serotonin-positive neurons. Most of the serotonin-positive and -negative nerve cell bodies were encircled by abundant baskets of serotonin-positive varicose fibers (Fig. 2A,C,E).

Mast cells

Numerous round-to-oval cells with a round nucleus and cytoplasmic serotonin immunoreactivity could be detected in the tela submucosa and tunica adventitia of the cervical and upper thoracic esophagus (Fig. 3A,C), but only a few were present in the lower thoracic and abdominal esophagus. These cells were identified as mast cells by their content of metachromatic granules with toluidine blue counterstaining (Fig. 3B,D). All serotonin-immunoreactive cells stained for toluidine blue, but some toluidine blue-stained cells were negative for serotonin. Only occasionally were single serotonin-reactive and toluidine blue mast cells seen between the smooth muscle cells of the lamina muscularis mucosae or in the lamina propria mucosae. However, it was difficult to clearly localize the serotonin-positive cells in the mucosa, as the layer of the lamina propria mucosae was very thin, especially in the upper portions, and was partly continuous with the tela submucosa because of incompletely developed lamina muscularis mucosae. In addition, a few of serotonin-positive mast cells were seen in close association with blood vessels (Fig. 1C) and serotonin-positive nerve fibers in the tela submucosa and tunica adventitia (Fig. 3A,C). In the jejunum, a few serotonin-immunoreactive and toluidine blue-stained mast cells were detected mainly in the telae submucosa and subserosa, but occasionally also in the lamina propria mucosae.

Brain stem

Independent of tissue treatment, all brain stem sections showed distinct serotonin immunoreactivity of neuronal perikarya and processes in the area of the raphe nuclei pallidus and obscurus, representing serotonergic groups B1 and B239–41 (Fig. 4A,B). In the compact formation of the nucleus ambiguus, numerous serotonin-positive nerve terminals were detected, but no serotonin-positive neurons were seen (Fig. 4C). Remarkably, these nerve terminals were in close contact with ChAT-positive nerve cell bodies (Fig. 4C,D).

Control experiments

The specificity of the serotonin immunoreactivity was tested in three ways: by replacing the primary antibody against serotonin with TBS or rabbit normal serum; by pre-absorbing the antibody against serotonin with the corresponding antigen; and by comparing the immunostaining with the used serotonin antibody in the jejunum and brain stem to previously published results. In negative and pre-absorption controls, only background fluorescence could be detected, with no specific immunostaining (data not shown). Cryosections of the jejunum showed serotonin-positive enterochromaffin cells in the epithelium of the mucosal villi and nerve cell bodies in the myenteric plexus. Serotonin-immunoreactive nerves were found in the myenteric plexus and around blood vessels in the jejunum (data not shown). Serotonergic neurons were present in the raphe nuclei pallidus and obscurus of the caudal medulla (Fig. 4A,B). Taken together, the results with the immunohistochemical negative and pre-absorption controls and with the serotonin immunoreactivity in intestinal enterochromaffin cells and neurons and the medullary raphe nuclei demonstrated the specificity of the serotonin antibody.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

This study represents the first description of serotonin-immunoreactive myenteric neurons and nerve fibers in myenteric ganglia, motor endplates, lamina muscularis mucosae, and blood vessels in the mouse esophagus. Numerous mast cells mainly located in the tela submucosa and tunica adventitia were double stained for serotonin and toluidine blue, and a few of them were in close association with serotonin-reactive nerve fibers.

Methodological considerations

In normally formaldehyde-fixed esophagus and small intestine, but not in the brain stem, the serotonin immunoreactivity in nerve fibers and cell bodies was too weak for understandable examination. Pretreatment with Krebs solution enhanced serotonin immunoreactivity in the esophagus and jejunum as the same high level as seen in the un-pretreated brain stem.39–41 Costa and colleagues25 have already reported this mismatch for the gastrointestinal tract, with subsequent studies confirming it,24,26,28,33 but the explanation remains unclear. It has been suggested that during pretreatment, the immunoreactive substance becomes unmasked within the tissue by dissociation from some cellular component that normally binds it.25 Different cellular components bound to serotonin could result in the mismatch in intensity of immunoreactivity between the gastrointestinal tract and brain stem.

Enteric co-innervation

Double and triple immunostaining revealed that serotonin and VAChT immunoreactivities were localized in completely separate nerve terminals on α-BT–tagged motor endplates and that serotonin was absent from nerve cell bodies in the compact formation of the nucleus ambiguus. Thus, serotonin-positive enteric neurons in the esophagus are considered to contribute to enteric co-innervation, as has been shown for nitrergic and peptidergic neurons.34,42 About 13% of all motor endplates were co-innervated by enteric nerve terminals positive for serotonin, suggesting serotonergic modulation of vagally induced contraction of a part of esophageal-striated muscles. The value for co-innervation seems low in comparison to the co-innervation rates of other common enteric markers, such as vasoactive intestinal peptide (37%) and nitric oxide synthase (39%), in the mouse esophagus.38 However, in an in vitro study, galanin does affect esophageal-striated muscles,43 despite a low co-innervation rate of the positive nerve terminals with 15%.38 As previous studies have reported that co-innervation rates varied between species, including humans44 and even between mouse strains,34 it is further necessary to examine the co-innervation rates other species like rats and guinea pigs display.

Myenteric neurons

About 10% of all myenteric neurons were positive for serotonin. Studies on the gastrointestinal tract of the guinea pig have revealed 2% serotonin-positive myenteric neurons in the stomach and small intestine and about 4% in the gastric antrum.25,31,45 These findings significantly differ from our results, suggesting a dominant influence of serotonin in the esophagus. The current results on serotonin innervation of the lamina muscularis mucosae and serotonin-positive enteric co-innervation of striated muscle suggest motor roles for serotonergic myenteric neurons in the esophagus. We demonstrated that serotonin-positive varicose fibers encircled most of nerve cell bodies in the ganglia, which were either positive or negative for serotonin. Thus, this suggests an interneuronal function of serotonergic myenteric neurons, similar to previous reports.46,47 Although an enteric source for serotonin-positive axons in the esophagus is plausible because of the presence of serotonin-immunoreactive myenteric neurons, caution is warranted because sympathetic postganglionic axons could take up serotonin from non-neuronal sources, such as thrombocytes, as shown for adrenergic neurons innervating cerebral arteries.48–50 Thus, some serotonin-positive axons at the various locations described in this study, might in fact represent adrenergic sympathetic nerve fibers containing serotonin. This issue could be settled using dopamine β-hydroxylase/serotonin double staining.

Functional considerations

Several in vitro and in vivo studies have demonstrated effects of serotonin on the isolated lamina muscularis mucosae51–53 and on the whole esophagus.54–56 The immunohistochemical demonstration of 5-HT4 receptors in the lamina muscularis mucosae of the guinea pig, rat, and mouse esophagus57, has suggested an involvement of serotonin in influencing esophageal contractility, the data being supported by in vitro experiments in the rat51–53 and the present study revealing the presence of serotonin-positive nerve fibers in this smooth muscle layer in the mouse. However, the in vitro results were controversial, because serotonin exhibited inhibitory and excitatory effects in the isolated lamina muscularis mucosae.51–53 The reason for this discrepancy remains unclear. Heterogeneous responses of serotonin have also been found in in vitro and in vivo experiments of the whole esophagus from some species including rats,54–56 guinea pigs, rabbits, and dogs.55 Serotonin has a relaxing effect on the rat, a contracting effect on the guinea pig and rabbit, and no effect on the dog esophagus. These differences also remain unexplained. In almost all of these experiments, only an effect of serotonin on smooth muscle was shown, but the possibility that serotonin influences striated muscle contractility via enteric co-innervation as demonstrated in the present study, was not discussed. Therefore, further detailed studies are needed to determine the physiologic roles of serotonin and its receptors at each of the sites identified in immunohistochemical studies.

Brain stem

Strong serotonin-reactive neurons were detectable in the area of the selected raphe nuclei pallidus and obscurus, but not in the compact formation of the nucleus ambiguus. Thus, a central origin of serotonin-positive nerve terminals on motor endplates in the esophagus could be ruled out. Moreover, serotonin-positive terminals encircling ChAT-positive neurons were identified in the nucleus ambiguus, consistent with previous reports on rich serotonergic input to almost all motor nuclei of the cranial nerves in several species.58,59 As serotonin-positive neurons of the brain stem raphe nuclei have extensive projections to virtually all areas of the brain and spinal cord,41,60 some of these projections could reach the compact formation of the nucleus ambiguus. In addition, ChAT-positive ambiguus neurons innervate esophageal-striated muscle fibers via cholinergic efferents terminating on motor endplates.61 Therefore, serotonin may modulate excitation of motor neurons in the compact formation of the nucleus ambiguus. Thus, we speculate that serotonin affects esophageal motility via the brain stem and enteric co-innervation peripheral levels.

Mast cells

A few mast cells were found in close association with serotonin-positive nerve fibers. A bidirectional communication between enteric neurons and mast cells in the intestine has been proposed (for review see Ref.62). Mast cells are considered as key elements in intestinal neuroimmune interactions, located with close proximity to enteric neurons, vagal nerve fibers, and spinal sensory nerves.62,63 Thus, the possible role of mast cells might also be true for the esophagus. Previous studies on the chronic inflammatory disease eosinophilic esophagitis support this view.64–66 Accumulating evidence suggests a strong role of mast cells in the inflammatory infiltrate in the pathophysiology of eosinophilic esophagitis (for review see Ref.66). As the mast cell activation in this disease seems not to be IgE-mediated, alternative mechanisms for the release of mediators by mast cells are discussed, for instance through the enteric nerve system (for review see Ref. 66). Nevertheless, the current results suggest that the serotonin effects on the esophagus occur not only through myenteric neurons or neurons in the raphe nuclei via the nucleus ambiguus but also through non-neuronally via mast cells.

Concluding remarks

The present study suggests the presence of serotonergic motor neurons for esophageal striated and smooth muscles in the tunica muscularis and lamina muscularis mucosae, and serotonergic interneurons. It is tempting to speculate that serotonin modulates vagal motor innervation of esophageal striated muscles both at the central level via projections of raphe nuclei to the compact formation of the nucleus ambiguus and at the peripheral level via enteric co-innervation of motor endplates. Furthermore, this study suggests that mast cells represent an additional source of serotonin, being particularly related to neuroimmune interactions.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

The authors are indebted to Anita Hecht, Andrea Hilpert, Hedwig Symowski, Karin Löschner, and Inge Zimmermann for expert technical assistance. Special thanks go to an anonymous reviewer for his valuable comments and suggestions to improve the quality of the manuscript. This study was supported by DFG NE 534/3-1.

Author contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

CH was involved in acquisition of data, data analysis and interpretation, statistical analysis, and drafting of the manuscript. WN was involved in the study concept and critical revision of the manuscript for important intellectual content. JW was involved in the study concept, study supervision, data analysis and interpretation, and critical revision of the manuscript for important intellectual content.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References