INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE ENTERIC NERVOUS SYSTEM (ENS) is different from the autonomic innervation of other organs, because it can mediate coordinated behaviors of the gut without central nervous system input (28, 32, 53). The presence in the bowel of intrinsic primary afferent neurons (IPANs) enables the ENS to respond to luminal stimuli. Because no nerve fibers enter the lumen, Bülbring and Crema (7) proposed that sensory transduction is transepithelial, involving the pressure-induced secretion of 5-HT from enterochromaffin (EC) cells to stimulate the mucosal processes of the submucosal IPANs that initiate reflexes. This hypothesis has since been confirmed, and EC cell-derived 5-HT is now thought to activate both peristaltic (21, 36, 37, 47, 50, 63) and secretory reflexes (12, 73).

In addition to its role in the initiation of enteric reflexes, 5-hydroxytryptamine (5-HT) also functions in ganglionic neurotransmission within the ENS. A subset of myenteric interneurons is serotonergic (8, 13, 22, 29, 31, 76, 77); therefore, 5-HT antagonists can block peristaltic reflexes by inhibiting enteric serotonergic neurotransmission (46, 60, 82) as well as by interfering with the paracrine stimulation of IPANs. 5-HT from EC cells also plays a role in extrinsic sensation by stimulating extrinsic primary afferent nerves (2, 39, 40).

The bowel contains an abundance of 5-HT receptor subtypes located on neurons, smooth muscle, and epithelial cells (24, 25, 30). Enteric neuronal 5-HT receptors include 5-HT_1A (26, 48, 49, 62), 5-HT_1P (6, 59, 64), 5-HT_2B (18), 5-HT₃ (15, 25, 42, 59), and 5-HT₄ (36, 37, 62, 80). Enteric members of the 5-HT₂ family are associated with smooth muscle, which they stimulate to contract (17, 27, 51, 52, 67), and epithelial cells, which they stimulate to secrete (3, 38, 41, 43, 74). The 5-HT_2B receptor, which was originally known as the rat fundus receptor because of its location on the smooth muscle of gastric rumen (10, 11, 20, 54, 56), has now been shown also to be expressed on intestinal neurons, to be developmentally regulated, and to promote the development of enteric neurons (18). The 5-HT_2C receptor is not expressed in the gut (18).

There have been suggestions that a 5-HT₂ receptor might be involved in the modulation of enteric neuronal activity (19, 35, 43, 68). The current study was thus undertaken to test the hypothesis that 5-HT_2A receptors are present on enteric neurons as well as on smooth muscle. To do so, we used RT-PCR to detect mRNA encoding the 5-HT_2A receptor in the developing and adult mouse intestine; moreover, light and electron microscopic immunoreactivity were employed to locate 5-HT_2A receptors intracellularly and on the surfaces of enteric neurons and smooth muscle cells. We also examined the intestines of mice carrying a targeted deletion in the 5-HT_2A promoter region (5-HT_2A /). Contractile responses to exogenous 5-HT, the microscopic structure of muscle, nerve, and epithelium, gastrointestinal transit, and colonic motility were compared in 5-HT_2A / mice and their +/+ littermates. Observations suggest that 5-HT_2A receptor may play roles in the ENS in the maintenance of the targets of enteric innervation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals and tissue collection. Adult Sprague-Dawley rats (Charles River Laboratories) were anaesthetized with methoxyflurane (Pitman Moore) and decapitated. Guinea pigs (Kingstar Laboratories) were stunned and exsanguinated. Transgenic mice with a targeted disruption of the promoter region of the 5-HT_2A receptor (5-HT_2A /) and their 5-HT_2A +/+ littermates were bred on a C57/B6 background (33). All pups were genotyped at birth by PCR analyses of tissue samples from the tails of the animals. All mice used in the current study were genotyped again at the time of use, once more by PCR analyses of tissue obtained from a toe. 5-HT_2A / mice failed to exhibit the activation of cortical layer V pyramidal cells or the behavioral head twitches and ear scratches that characterize responses to the systemic administration of 5-HT_2A agonists (34). Female mice [age ~6 mo; C57/B6 (5-HT_2A / and 5-HT_2A +/+ littermate controls) and CD-1 (Charles River Laboratories)] were killed by cervical dislocation after narcotization with CO₂. Fetuses, obtained from timed pregnant CD-1 mice, were anaesthetized by cooling and exsanguinated before dissection. The Animal Care and Use Committee of Columbia University approved all procedures. The gut was dissected from the animals and cleaned with Krebs solution. Stomach, small intestine, and colon were each analyzed. Whole fetuses were also fixed and frozen for cryostat sectioning (see below). In a subset of experiments, the longitudinal muscle with adherent myenteric plexus (LM-MP) was dissected from the wall of the gut at ages ranging from embryonic (E) day 16 (mouse) to adult (rat and mouse). These LM-MP preparations were stretched and pinned out as flat as possible on dishes coated with a silicone elastomer (Sylgard) and maintained in Krebs solution until fixed (see below). Primary cultures of dissociated cells from the E14 fetal gut were prepared as previously described (18). Briefly, after digestion of the fetal gut with collagenase, the crest-derived cells were immunoselected with the specific antibody for p75^NTR. Cells were grown in defined media on laminin substrate for 4 days and then fixed for immunocytochemistry.

RT-PCR. RNA was extracted from segments of mature or fetal bowel using the guanidinium thiocyanate method (9). RT-PCR was employed to determine whether mRNA encoding any of the members of 5-HT₂ receptor family could be detected in the stomach, small intestine, or colon of either 5-HT_2A +/+ or 5-HT_2A / mice. For first strand cDNA synthesis, 1 µg of RNA was incubated for 1 h at 42°C with 200 units of Moloney murine leukemia virus (M-MLV) reverse transcriptase, using random primers at a concentration of 1 µM. This reaction and subsequent amplification with Taq polymerase was carried out with a commercial kit (GeneAmp; Perkin-Elmer, Foster City, CA) according to the manufacturer's instructions. The sets of PCR primers used for the analyses are listed in Table 1. -Actin was used as an internal control for experiments involving comparative RT-PCR. The PCR profile for each set of primers (listed in Table 1) was programmed into a model PTC-150 programmable thermal cycler (MJ Research, Watertown, MS). PCR reaction products were resolved on 1.2% agarose/40 mM Tris-acetate, 1 mM EDTA gels, and their sizes were determined by using a 123-bp standard ladder.

Immunocytochemistry. Both fresh-frozen and fixed preparations were examined. Segments of adult gut and E16 fetuses were fixed with 4% formaldehyde (from paraformaldehyde) in PBS for 1 and 4 h, respectively, at room temperature. Both preparations were then washed extensively with PBS (6 × 10 min), infiltrated with sucrose (30%; 4°C for 24-36 h), embedded with optimum cutting temperature medium (Lipshaw), frozen in liquid N₂, sectioned (at 10 µm) with a cryostat-microtome, and collected on gelatin-coated glass slides. Sections of fresh-frozen tissue were fixed on slides (1% formaldehyde, 10 min, 4°C) and washed (2 × 10 min) with PBS containing 0.1% Triton X-100 (PBS-T). Endogenous peroxidase activity was inhibited by treating preparations for 30 min with H₂O₂ (0.3%) in PBS-T. Preparations were washed again with PBS-T and blocked for 30 min with 4% goat serum in PBS containing 0.3% Triton X-100. Primary antibodies (Table 2) were then applied to the sections for 72 h at 4°C. Sites of antibody binding were detected with secondary antibodies and, if necessary, visualizing reagents (Table 3). Double label fluorescence immunocytochemistry was used to identify the immunoreactivities of the neuronal marker, ubiquitin hydrolase protein gene product 9.5 (PGP 9.5) (85) together with that of 5-HT_2A receptors (Table 2). For studies of receptor development in vitro, fixed cultures were either permeabilized with PBS-T supplemented with 4% goat serum or examined without prior permeabilization to demonstrate the immunoreactivity of receptors inserted into the plasma membrane. LM-MP preparations were fixed (as above) while pinned flat. The fixed material was then washed with PBS, cut into small rectangles (~0.5 cm × 1.0 cm) and processed as free-floating whole mounts. The LM-MP preparations were washed (3 × 10 min) with PBS-T, blocked for 30 min with 4% goat serum in PBS containing 0.3% Triton X-100 (blocking solution), incubated overnight (4°C) with primary antibodies (Table 2), and finally visualized with appropriate secondary antibodies (Table 3). Preparations were mounted flat on slides and coverslipped with Vectashield media (Vector Laboratories) to prevent fading. Fluorescence microscopy was carried out with a Leitz DMRD microscope equipped for vertical excitation. The filter/mirror cube used to visualize the fluorescence of FITC did not reveal the emission of cyanine 3 and that employed for the visualization of cyanine 3 fluorescence did not pass the FITC emission. Alternatively, specimens were examined with a Zeiss confocal microscope.

Electron microscopy. For routine electron microscopy (EM) analysis, ~1-cm segments from the proximal and distal portions of the small intestine and the middle of the colon of 5-HT_2A +/+ and / mice were pinned flat on Sylgard and fixed for 3 h at room temperature in a solution containing 4% formaldehyde and 3.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4). The specimens were then postfixed for 1 h in a solution containing 1% OsO₄, washed extensively with sodium maleate buffer (pH 6.2), stained en bloc with maleate-buffered uranyl acetate (2%), and dehydrated in an ethanol gradient before embedding in an epoxy resin (Epon 812; Electron Microscopy Science). Thin sections were then cut, picked up on copper grids, and stained with uranyl acetate and lead citrate. Sections were examined and photographed with a JEOL 1200 EX microscope.

EM immunocytochemistry. Segments of bowel were cut along the mesenteric border and pinned flat in tissue culture medium supplemented with nicardipine (1 µM) to relax the smooth muscle cells. Specimens were then fixed by immersion for 3 h in a solution containing 4% formaldehyde, 0.01 M sodium periodate, and 0.075 M lysine · HCl in 0.1 M sodium phosphate buffer (pH 7.4; to maximally preserve membranes without destroying 5-HT_2A antigenicity). After fixation, tissues were washed with 50% ethanol (4 × 10 min) and incubated overnight at 4°C in phosphate buffer (0.1 M). The LM-MP was then dissected from the preparations and treated for 1 h with 1% sodium borohydride and, subsequently, for 30 min with 10% normal goat serum (in PBS). Antibodies to the 5-HT_2A receptor (Table 2) were applied overnight at 4°C in PBS containing 4% normal goat serum. After washing with PBS, immunoreactivity was detected with biotinylated goat anti-mouse secondary antibodies and visualized with streptavidin-horseradish peroxidase (Table 3). The tissue was then prepared for EM examination as described above.

Quantifying the branching of the myenteric plexus. Neurons and their processes in LM-MP preparations from 5-HT_2A +/+ and 5-HT_2A / mice were labeled with the neuronal marker PGP 9.5. Random fields were photographed and examined at a uniform magnification of ×86. The percentage of the total area of each micrograph occupied by PGP 9.5-immunoreactive ganglia or secondary rami was then determined (tertiary branches of the plexus were not considered for this analysis). To estimate this area, a grid of 114 evenly spaced dots was superimposed on each picture and the proportion of the dots falling on the PGP 9.5-immunoreactive neural structures was determined. Data were analyzed by ANOVA using the StatView 4.0 program for the Macintosh.

Gastric emptying and small intestinal transit. Transit through the stomach and small intestine was measured by administering a nonabsorbed marker (10% charcoal suspension in 5% gum Arabic) to 5-HT_2A +/+ and 5-HT_2A / mice (45, 58). The mice were given 0.2 ml of the suspension by gavage through a straight blunt-ended feeding needle. Twenty min after the charcoal was administered, the animals were killed and the entire gastrointestinal tract was removed. The distances from the pylorus to the front of the charcoal bolus and to the ileocecal junction were measured. The rate of transit was determined from the relationship: [distance to charcoal front]  [length of small intestine] × 100 and expressed as a percent. Transit was measured in 12 5-HT_2A +/+ and 12 5-HT_2A / mice, and means were compared by Student's t-test.

Longitudinal muscle contraction. Contraction of the colon from 5-HT_2A +/+ and 5-HT_2A / mice was measured in response to 5-HT (10 µM) and acetylcholine (10 µM) (16). Briefly, colons were removed from 5-HT_2A +/+ and 5-HT_2A / mice with their attached mesentery, cleaned, mounted in a vertical organ bath, and continuously superfused with Krebs solution (equilibrated with 95% O₂-5% CO₂) at 37°C. The aboral end of the colon was fixed and the oral end was attached to a linear motion transducer (model ST-2; Phipps and Bird). Movements of the colon were displayed on a potentiometric pen recorder. The gut was allowed to equilibrate for 30-40 min before the start of each experiment. Compounds were applied for 5 min, during which the superfusion of fresh buffer was halted. At least 15 min were allowed to elapse before another compound was applied to the bath. In no instance was a second drug tested before the response to the prior application of a drug had dissipated and the preparation had resumed its resting length. Segments of colon were removed from three mice of each type for these analyses.

Colonic transit. The motility of the colon was evaluated in separate sets of 5-HT_2A +/+ and 5-HT_2A / littermates. Animals were lightly anesthetized with ether, and a glass microbead (3 mm in diameter) was inserted through the anus and pushed, with a polished glass rod, into the colon for a distance of 2 cm (61). The time from the completion of insertion to the expulsion of the bead was measured to the nearest 0.1 min to provide an estimate of the rate of motility of the colon.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Enteric 5-HT_2A receptors were located by immunocytochemistry. Immunocytochemistry with 5-HT_2A-selective antibodies was used to locate sites of 5-HT_2A receptor expression in mature small and large intestines of the mouse, rat, and guinea pig. Fixed-frozen sections and whole mount preparations from each species were examined. The patterns of expression observed in the adult bowel of the mouse, rat and guinea pig were identical. 5-HT_2A receptor immunoreactivity was detected in the mucosal epithelium (Fig. 1, A and C), longitudinal and circular layers of the muscularis externa (Fig. 1, A, C, and F), as well as in both the submucosal (Fig. 1, A, C, and D) and myenteric plexuses (Fig. 1, A, C, and E). The epithelial 5-HT_2A immunoreactivity was moderate in enterocytes and crypt cells of the small intestine but intense in Paneth cells (Fig. 1A). The 5-HT_2A receptor immunoreactivity of crypt cells and enterocytes was more apparent in the colon, where it was highly concentrated at basolateral cell surfaces, than in the small intestine (compare Fig. 1, A with C). Many neurons in each of the enteric plexuses (~40%) of both the small and large intestines were 5-HT_2A immunoreactive (Fig. 1, A, and C-E). The ganglionic neuropil and the interganglionic connectives also displayed 5-HT_2A receptor immunoreactivity (Fig. 1, D and E). In whole mount preparations, a series of 5-HT_2A immunoreactive "hot" spots were visible in an irregular distribution along the long axis of the muscle fibers (Fig. 1F). 5-HT_2A-immunoreactive nerve fibers coursed through the circular muscle layer; these nerve fibers were particularly striking in the colon (Fig. 1C). No 5-HT_2A-immunoreactive nerve fibers were observed in the longitudinal muscle of either the small intestine or the colon.

View larger version (182K): [in this window] [in a new window]   Fig. 1.   5-Hydroxytryptamine (5-HT)2A immunoreactivity is found in enterocytes, Paneth cells, the enteric nervous system (ENS; submucosal and myenteric neurons), and the smooth muscle of the muscularis externa in the rat small intestine. A: full thickness of the wall of the small intestine is cut in cross section. Paneth cells (arrowheads) are densely 5-HT2A immunoreactive, whereas the neighboring enterocytes of the crypt epithelium are lightly immunostained. Submucosal and myenteric ganglia (arrows) appear to be more densely immunoreactive than the muscle cells of either the circular muscle (cm) or longitudinal muscle (lm) layers. Smooth muscle 5-HT2A immunoreactivity outlines cells and is thus probably concentrated in plasma membranes. The small dark cells in the lamina propria bind avidin and are stained nonspecifically; they can also be seen as tissue not exposed to primary antibodies (B). B: control; there is no staining of enterocytes, Paneth cells, the ENS, or smooth muscle in the absence of antibodies to the 5-HT2A receptor. C: cross section of the rat colon. The basolateral surfaces of crypt epithelial cells (arrowheads), as well as submucosal and myenteric ganglia (arrows), express intense 5-HT2A immunoreactivity. Smooth muscle coats of both cm and lm layers are outlined by 5-HT2A immunoreactivity. D: dissected whole mount of the submucosal plexus. Many neuronal cell bodies are densely 5-HT2A immunoreactive (arrows). The neuropil and interganglionic connectives are moderately immunostained. E: dissected whole mount of the lm with adherent myenteric plexus (LM-MP). The muscle is out of focus. Many neuronal cell bodies in ganglia are densely 5-HT2A immunoreactive (arrows). The neuropil and interganglionic connectives are moderately immunostained. F: dissected whole mount of the lm layer, which is in focus. Muscle fibers are 5-HT2A immunoreactive, and there are "hot" spots (arrowheads) that stand out along the length of the fibers. Bars = 50 µm.

View larger version (135K): [in this window] [in a new window]   Fig. 2.   5-HT2A immunoreactivity is found on the plasma membranes of axonal varicosities and in caveolae of smooth muscle. A: ganglion of the myenteric plexus of the rat small intestine. 5-HT2A immunoreactivity outlines the plasma membranes of several varicosities (arrows). B: interganglionic connectives of the myenteric plexus of the rat small intestine. 5-HT2A immunoreactivity appears on the plasma membranes of axonal varicosities (arrows) that contain synaptic vesicles. Some irregularly shaped vesicles within these varicosities are immunostained. C: control section through a ganglion of the myenteric plexus of the rat small intestine. There is no immunostaining in the absence of primary antibodies to the 5-HT2A receptor. D: control section through the smooth muscle of the muscularis externa of the rat small intestine. No immunostaining is seen in the absence of primary antibodies to the 5-HT2A receptor. E and F: smooth muscle of the muscularis externa of the rat small intestine. 5-HT2A immunoreactivity is found in caveolae (arrowheads) associated with the plasma membranes of smooth muscle cells. Bars = 1 µm.

View larger version (167K): [in this window] [in a new window]   Fig. 3.   Intracellular 5-HT2A immunoreactivity is found in a subset of enteric neurons. A: control. Neither varicosities nor cell bodies of enteric neurons display a 3,3'-diaminobenzidene reaction product in tissues not exposed to antibodies to the 5-HT2A receptor. B: 5-HT2A immunoreactivity is found in the cytoplasm of a myenteric neuron. The immunoreactivity is most dense in ribosome-rich patches near elements of endoplasmic reticulum that appear to have been damaged by the horseradish peroxidase reaction; n, nucleus. C: in another myenteric neuron, 5-HT2A immunoreactivity can be seen in vesicles near the Golgi apparatus and in association with the dilated ends of Golgi cisternae. D: 5-HT2A immunoreactivity is found in subsets of neuritic processes (arrows) in an interganglionic connectives. Inset: reaction product can also be seen in association with synaptic membranes where varicosities containing small, clear synaptic vesicles appose dendrites. E: vesicles containing 5-HT2A immunoreactivity are found in subsets of labeled neurites and varicosities (arrows). Bars = 1 µm.

The 5-HT_2A receptor is expressed in the E16 fetal bowel. 5-HT_2A receptor immunoreactivity was demonstrated in whole mounts of preparations of the E16 fetal bowel from which the mucosa had been removed by dissection under microscopic control. The tissues were then examined by laser scanning confocal microscopy. The immunoreactivity was found to be strikingly localized in two distinct planes of the preparations. One corresponded to the developing myenteric plexus (Fig. 4A). Within this plane, the immunoreactivity of the 5-HT_2A receptor was found in an anastomosing network of relatively thick rami comprised mainly of the cellular elements of the primordial plexus. Because immunofluorescent cells adjoined one another with little or no intervening space, the borders of individual cells could not be distinguished. Distinct secondary and tertiary branches of the plexus, such as are seen in the mature myenteric plexus, were not apparent, although the presumptive interganglionic connectives were thinner and more fibrous than the presumptive ganglia. The second plane, at the border between the circular muscle and the submucosa, consisted of a dense anastomosing mesh of thin fibers with a predominant orientation parallel to the direction of the circular muscle (Fig. 4B). The immunofluorescent fibers were not varicose. Depth-coding the images confirmed that the 5-HT_2A-immunoreactive structures were indeed located in separate planes of the gut wall, that there was little overlap of the thin fibers and the myenteric plexus, and that the predominant orientation of the thin 5-HT_2A-immunoreactive fibers was perpendicular to that of developing myenteric ganglia (Fig. 4C).

View larger version (78K): [in this window] [in a new window]   Fig. 4.   Networks of 5-HT2A-immunoreactive processes are found in the embryonic (E) day 16 small intestine. These networks lie in 2 distinct planes and differ in appearance. A: in the plane of the developing myenteric plexus, 5-HT2A immunoreactivity is found in ganglia and developing interganglionic connectives. B: second plexus of much finer 5-HT2A-immunoreactive processes is seen deep to the myenteric plexus. C: depth-coded illustration. General orientation of the deep 5-HT2A-immunoractive plexus is roughly perpendicular to that of the primordial myenteric plexus and ~20 µm below it. Pseudocolors: 20 µm, red  blue = superficial  deep. Bar = 50 µm.

View larger version (99K): [in this window] [in a new window]   Fig. 5.   5-HT2A-immunoreactive neurons and processes are abundant in the E16 small intestine. A-C: cross-sectioned preparations. D-F: whole mount preparations. Specimens were processed to demonstrate protein gene product 9.5 (PGP 9.5) and 5-HT2A immunoreactivities simultaneously. A: PGP 9.5 immunoreactivity; this panneuronal marker shows the locations in the intestinal wall of ganglia of the myenteric (arrow) and submucosal (smp) plexuses. Varicose axons extend into the cores of villi (v). The lm contains no nerve fibers. B: 5-HT2A immunoreactivity; myenteric but not submucosal neuronal cell bodies are immunoreactive. C: Superimposed images. Note that the varicose neurites in the villus core are not 5-HT2A immunoreactive. D: PGP 9.5 immunoreactivity. A thick nerve trunk (arrow) extends into the core of a villus from the smp. E: 5-HT2A immunoreactivity. Neurites in the smp and the thick nerve trunk (arrow) are immunoreactive. F: superimposed images. 5-HT2A-immunoreactive neurites are a subset of the PGP 9.5-immunoreactive population. Bar = 25 µm.

5-HT_2A immunoreactivity is present on cell bodies and varicosities of neurons that develop in vitro. 5-HT promotes the development of neurons in vitro, and this effect is mediated by 5-HT_2B receptors (18). The observations discussed above, however, suggest that 5-HT_2A receptors are also expressed in the developing ENS. It is difficult to expose the surfaces of neurons in situ for investigations of the distribution of receptors. Individual processes and especially the growth cones of extending neurites, furthermore, are not easily distinguished. To facilitate an analysis of the location of 5-HT receptors in developing neurons, therefore, we took advantage of superior accessibility of enteric neurons developing in dissociated cell cultures from neural crest-derived precursors (E16). The distribution on enteric neurons of the 5-HT_2A receptor was compared with that of 5-HT_2B receptors. Antibodies that react with extracellular domains were used for the immunocytochemical detection of the 5-HT_2A and 5-HT_2B receptors. Cultures to be examined were fixed and investigated as whole mounts with or without prior permeabilization. The nonpermeabilized preparations were studied to minimize intracellular immunoreactivity and thus to facilitate the localization of receptors on cell surfaces.

View larger version (81K): [in this window] [in a new window]   Fig. 6.   5-HT2A immunoreactivity is present on the cell bodies and perikarya of neurons that develop in vitro. A and B: enteric neurons and varicosities processed as whole mounts and viewed to demonstrate PGP 9.5 (A) and 5-HT2A immunoreactivities (B) simultaneously. Note that 5-HT2A immunoreactivity is concentrated on the cell bodies of the neurons and the varicosities of their axons. C: enteric neuron developing in vitro has been doubly immunostained with antibodies to PGP 9.5 (green fluorescence) and 5-HT2A receptors (red fluorescence). The 5-HT2A immunoreactivity in this cell is confined to "hot spots" at varicose expansions and branch point of extending neurites (arrows). D and E: enteric neurons and varicosities processed as whole mounts and viewed to demonstrate PGP 9.5 (D) and 5-HT2B immunoreactivities (E) simultaneously. One neuron is 5-HT2B immunoreactive, whereas the other is not. F-H: extending neurites and growth cones illuminated to demonstrate PGP 9.5 (F) and 5-HT2B (G) immunoreactivities. Superimposed images are seen in H. Note that 5-HT2B immunoreactivity is strikingly concentrated at growth cones. Bars = 25 µm.

Relative size of the myenteric plexus in 5-HT_2A / mice is not significantly different from that of their 5-HT_2A +/+ littermates. To determine whether 5-HT_2A receptors affect the development of enteric neurons, the relative size of the myenteric plexus was measured in 5-HT_2A / mice, which lacked 5-HT_2A receptors, and in their 5-HT_2A +/+ littermates, which served as controls. Neurons were demonstrated in dissected whole mounts of colon. This portion of the bowel was investigated, because a preliminary study carried out by RT-PCR did not reveal mRNA encoding the 5-HT_2A receptor in the colons or stomachs of the 5-HT_2A / animals (Fig. 7A). In contrast, the 5-HT_2A / small intestine was found to contain small amounts of residual mRNA encoding the 5-HT_2A receptor. Immunocytochemistry was also employed in 5-HT_2A / animals, but the 5-HT_2A immunoreactivity could not be distinguished from background. Exogenous 5-HT (10 µM) was applied to the isolated colon of 5-HT_2A +/+ and / mice to determine whether the 5-HT_2A-mediated contractile effect, which is a direct response of the longitudinal smooth muscle to 5-HT (16), was retained or absent in the animals lacking 5-HT_2A receptors (Fig. 7B). Whereas 5-HT evoked a contraction of the 5-HT_2A +/+ colon, no response was elicited by 5-HT when it was applied to colon of the 5-HT_2A / mice. The 5-HT_2A / colon, however, contracted in response to ACh (10 µM) and thus was able to respond to agents for which receptors were present in the tissue.

View larger version (36K): [in this window] [in a new window]   Fig. 7.   Expression of the 5-HT2A receptor is lacking in the colons of 5-HT2A / mice. A: RT-PCR; mRNA encoding the 5-HT2A receptor is present in the stomach, small intestine, and colon of 5-HT2A +/+ mice (wt). 5-HT2A transcripts are absent in the stomach and colon of 3 separate 5-HT2A / mice (ko), but a minimal amount of residual 5-HT2A expression can be detected in each of their small intestines. B: 5-HT (10 µM) induces a contraction of the lm of the 5-HT2A +/+ colon. In response to 5-HT (10 µM), lm of the 5-HT2A / colon does not contract, but it does contract in response to ACh (10 µM).

View larger version (77K): [in this window] [in a new window]   Fig. 8.   The size and pattern of the myenteric plexus in 5-HT2A +/+ and / mice do not differ significantly from one another. A: percentage of serosal area occupied by PGP 9.5-immunoreactive elements of the myenteric plexus in wt and ko mice. B: PGP 9.5 immunoreactivity in a dissected preparation of LM-MP from a 5-HT2A / mouse. C: PGP 9.5 immunoreactivity in the LM-MP from a 5-HT2A +/+ mouse. Bars = 200 µm.

View larger version (86K): [in this window] [in a new window]   Fig. 9.   Nitric oxide synthase-1 (NOS-1)- and calcitonin gene-related peptide (CGRP)-immunoreactive cells and neurites cannot be distinguished in 5-HT2A +/+ and / mice. A, C, and E: images from bowel of 5-HT2A +/+ mice. B, D, and F: images from bowel of 5-HT2A / mice. A-D: NOS-1 immunoreactivity in the small intestine (A and B) and colon (C and D). E and F: CGRP immunoreactivity in the colon. Bars = 50 µm. G: RT-PCR; mRNA encoding NOS-1 is found in the colon (Col), small intestine (SI), and stomach (St) of 5-HT2A +/+ (left) and 5-HT2A / (right) mice. Relative amounts can be compared semiquantitatively relative to simultaneously analyzed mRNA encoding -actin (Act).

5-HT_2B expression in 5-HT_2A +/+ mice is not distinguishable from their 5-HT_2A / littermates. The possibility that other members of the 5-HT₂ family of receptors might be upregulated in the bowel to compensate for the knockout of the 5-HT_2A receptor was investigated. As noted previously (18), mRNA encoding the 5-HT_2C receptor could not be detected by RT-PCR anywhere in the bowel of wild-type (5-HT_2A +/+; 5-HT_2B +/+) mice. Transcripts encoding the 5-HT_2C receptor did not appear in the 5-HT_2A / gut (data not illustrated). The 5-HT_2B receptor, however, is detectable in wild-type mice, both in the stomach, the small intestine, and the colon (Fig. 10). The level of expression of mRNA encoding the 5-HT_2B receptor in 5-HT_2A +/+ mice was compared with that in their 5-HT_2A/ littermates by semiquantitative RT-PCR (Fig. 10). No difference was apparent in the stomach, the small intestine, or the colon; therefore, if 5-HT_2B expression changes as a result of the knockout of the 5-HT_2A receptor, it does so too subtly to be detectable by semiquantitative RT-PCR.

View larger version (18K): [in this window] [in a new window]   Fig. 10.   Expression of 5-HT2B receptors is not substantially changed in 5-HT2A / mice. mRNA encoding the 5-HT2B receptor and -actin (-Act) was assessed by RT-PCR in the Col, SI, and St of 5-HT2A +/+ and / mice.

Smooth muscle, Paneth cells, and enterocytes are abnormal in 5-HT_2A / mice. The bowel wall appeared to be thinner in 5-HT_2A / mice than in their 5-HT_2A +/+ littermates. Measurements were thus made to quantify the thickness of the muscle layers and the epithelium to identify the components of the bowel responsible for this apparent difference among animals. These studies concentrated on cells (enterocytes, Paneth cells, smooth muscle) found in the immunocytochemical studies (above) to express 5-HT_2A immunoreactivity. The height of enterocytes was measured from the tip of the microvillus border to the basement membrane (Fig. 11A). The region selected for measurement was on the sides of villi, where mature cells are found. The cell extrusion zones at the tips of villi and cells in crypts were excluded. The plane of section was such that the nuclei of the measured cells formed a single row parallel to the basement membrane. Measurements were obtained in both the proximal (duodenum) and distal (ileum) small intestine. The height of the enterocytes of 5-HT_2A +/+ mice in both the proximal (37.9 ± 0.8 µm) and distal (33.9 ± 0.9 µm) small intestine were significantly greater (P < 0.0001) than those of their 5-HT_2A / littermates (proximal = 28.5 ± 1.0 µm; distal = 19.8 ± 1.3 µm). The normal proximodistal decrease in enterocyte height was preserved, although the cells were smaller both proximally and distally, in the 5-HT_2A / animals. In contrast to villus height, the number of Paneth cells per crypt increased proximodistally in 5-HT_2A +/+ mice (Fig. 11B). The density of Paneth cells, counted as the number per crypt profile, was significantly less in 5-HT_2A / mice than in their 5-HT_2A +/+ littermates, both in the proximal (1.3 ± 0.2 vs. 3.3 ± 0.4, respectively; P < 0.0001) and distal (2.6 ± 0.4 vs. 5.3 ± 0.6 respectively; P < 0.004) small intestine. In addition to the epithelial abnormalities, the thickness of both the circular (Fig. 11C) and longitudinal (Fig. 11D) muscle layers was significantly less (P < 0.0001) in both the proximal and distal small bowel of 5-HT_2A / mice than in their 5-HT_2A +/+ littermates. The thickness of the two external muscle layers tended to increase proximodistally, a pattern seen in both 5-HT_2A +/+ and / animals. The difference in the thickness of the layers appeared to be due to the presence of thinner muscle cells in the 5-HT_2A / mice rather than to fewer cells. The longitudinal muscle was about four to five cells thick in the small intestine and three cells thick in the colon of both 5-HT_2A +/+ and / mice. The thickness of the circular muscle was composed of ~6 cells in the small intestine and 17 cells in the colon in both types of mouse.

View larger version (28K): [in this window] [in a new window]   Fig. 11.   The height of enterocytes, numbers of Paneth cells, and thickness of the layers of the muscularis externa were all significantly less in 5-HT2A / mice than in their 5-HT2A +/+ littermates. Comparisons were made with toluidine blue-stained plastic sections (1.0 µm thick) of the proximal and distal small intestine. Three mice per group were analyzed (the n values listed below are the numbers of measurements made for each parameter). Differences between 5-HT2A +/+ mice and their / littermates were statistically significant (*). A: enterocyte height (proximal: n = 21 +/+ and 25 /; distal: n = 8 +/+ and 12 /). B: numbers of Paneth cells per crypt. (proximal: n = 17 +/+ and 31 /; distal: n = 18 +/+ and 10 /). C: thickness of the cm (proximal: n = 22 +/+ and 31 /; distal: n = 8 +/+ and 16 /). D: thickness of the lm. (n = 11 +/+ and 25 /; distal: n = 10 +/+ and 9 /).

View larger version (166K): [in this window] [in a new window]   Fig. 12.   Degenerative changes are found in a subset of muscle fibers in the muscularis externa of 5-HT2A / mice. Three mice of each phenotype were investigated. A and B: illustrated sections pass through the muscularis externa of the small intestines of two 5-HT2A +/+ mice. Sections were taken either from the center of the circular layer (A) or the border of the cm with the submucosa (B). A ganglion of the smp appears at the upper right in B. C-F: illustrated sections pass through the muscularis externa of two 5-HT2A / littermates. Sections were taken from the circular layer at its border with submucosa (C), center of cm, or from lm (D) of small intestine and from the circular layer of colon (F). Cells outlined by arrowheads show degenerative change. Note relative electron lucent cytoplasm and relative paucity and disorientation of thin filaments of these cells. In cross sections of abnormal muscle cells, particularly in colon, zones free of myofilaments appear so that remaining myofilaments are form islands in cytoplasm. Bars = 1.0 µm.

Gastrointestinal transit and colorectal motility are similar in 5-HT_2A +/+ and / mice. Gastrointestinal transit was measured to determine whether the loss of the 5-HT_2A receptor affects the timing of gastric emptying and subsequent propulsion in the small intestine. The rate of gastrointestinal transit was measured in 5-HT_2A / mice and compared with that in their 5-HT_2A +/+ littermates. Charcoal (in a suspension in gum Arabic) was administered by gavage, and proportion of the length of the small intestine traversed by the charcoal was determined. The rate of gastrointestinal transit in 5-HT_2A / animals did not differ significantly from that in their 5-HT_2A +/+ littermates (Fig. 13A). The motility of the colon was also examined. In this case, the speed of colonic transit was estimated from the time required to expel a glass bead inserted into the colon a fixed distance of 2 cm from the anal verge. The rate of colonic transit of 5-HT_2A / mice did not differ significantly from that of their 5-HT_2A +/+ littermates (Fig. 13B).

View larger version (12K): [in this window] [in a new window]   Fig. 13.   Neither gastrointestinal transit time nor colorectal motility differ significantly in 5-HT2A / mice from those of their 5-HT2A +/+ littermates. Measurements of each type were made on 10 paired littermates of each phenotype. A: gastrointestinal transit time estimated as the percentage of length of small intestine traversed within 20 min by a bolus of charcoal instilled into the stomach by gavage. B: motility in the distal colon and rectum estimated as the time to expulsion through the anus of a glass bead pushed into the colon for a distance of 2 cm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Since Gaddum and Picarelli (23) first distinguished between enteric M and D receptors, muscular and neuronal 5-HT receptors of the bowel have generally been thought to be distinct and nonoverlapping (17, 69). When the enteric 5-HT₃ and 5-HT_2A receptors were later identified, the 5-HT₃ receptor was equated with the M receptor, because it was neuronal, whereas the 5-HT_2A receptor was equated with the D, because it acts directly on intestinal smooth muscle (4). It is now apparent, however, that enteric 5-HT receptors do not fall neatly into nonoverlapping neuronal (M) and muscular (D) categories. The 5-HT₄ receptor is expressed both by enteric neurons (14, 25, 62) and muscle cells (51, 52, 70). We have previously demonstrated that the 5-HT_2B receptor is also expressed by enteric nerve and muscle (18), and we now find that the prototypic D receptor, the 5-HT_2A (4, 17, 69), is even more widespread in its enteric distribution and is expressed not just by smooth muscle but also by epithelial cells and neurons.

5-HT_2A immunoreactivity was observed in mucosal enterocytes and Paneth cells, neurons in both the submucosal and myenteric plexuses, and in both circular and longitudinal muscle cells. The location of 5-HT_2A receptors on enterocytes probably accounts for the ketanserin-sensitive ability of 5-HT to act directly on the epithelium to induce Cl secretion, increase inositine trisphophate, and for the specific binding of ³H-ketanserin (3, 38, 41, 43, 74). 5-HT has also been reported to increase the rate at which enterocyte precursors proliferate (81). Because 5-HT₂ receptors stimulate proliferation when expressed in transfected fibroblasts (44, 55), the enhancement of enterocyte proliferation by 5-HT may be mediated by a 5-HT₂ receptor.

In contrast to epithelial cells, no physiological response in enteric neurons has yet been demonstrated to be 5-HT_2A-mediated. It is conceivable that the response evoked in enteric neurons by 5-HT_2A stimulation has not yet been revealed by the electrophysiological or biochemical methods thus far employed. Species differences in the enteric innervation may also be significant, and the electrophysiological investigations that have been carried out have been heavily biased toward the guinea pig. Certainly, the location of 5-HT_2A receptors on the plasma membranes of axon terminals and their concentration at synaptic junctions (Figs. 2, A and B, and 3, D and E) is consistent with the possibility that 5-HT_2A receptors evoke pre- and/or postsynaptic effects; moreover, indirect pharmacological studies have suggested that a receptor in the 5-HT₂ family modifies the release of ACh (35, 68) and other transmitters (43) from enteric nerves. The distribution of 5-HT_2A receptors on neuronal plasma membranes is thus consistent with a 5-HT_2A mediation of these 5-HT₂-like actions. A 5-HT_2A-mediated trophic effect would be equally consistent and might not be detected by measurements of membrane potential or current. The abundance of 5-HT_2A immunoreactivity that is found intracellularly in subsets of neuronal perikarya probably reflects the biosynthesis of the receptors in the rough endoplasmic reticulum (Fig. 3, B and C). The intracellular 5-HT_2A immunoreactivity observed in axons and in irregularly shaped vesicles in varicosities is likely to reflect axonal transport of the receptors from cell bodies to terminals (Figs. 2B and 3E).

The distribution of 5-HT_2A receptors in caveolae on the plasma membranes of smooth muscle cells (Fig. 2, E and F) is shared with certain other receptors involved in cell signaling (1, 72). Because the affinity of the 5-HT_2A receptor for 5-HT is relatively low (65), the concentration of 5-HT in caveolae by potocytosis (1) may facilitate ligand-receptor interactions. The abundance of 5-HT_2A receptors on fibers in the deep muscle plexus (Fig. 4B) is consistent with a role in interactions of enteric neurons with interstitial cells of Cajal, which are concentrated in this region (71, 78, 79), intimately associated with nerve fibers (78, 83), and thought to be intermediaries in the cholinergic neural control of the circular muscle (84).

Transgenic mice with a targeted deletion of the promoter region of the 5-HT_2A receptor were investigated to gain insight into the physiological role(s) played by these receptors by a loss-of-function analysis. 5-HT_2A expression was confirmed to be lacking in the colons of the 5-HT_2A / animals, although minimal 5-HT_2A expression was retained in their small intestine. Conceivably, the 5-HT_2A receptor is expressed to a slight extent in the small intestine, because a limited tissue-specific activation of the gene can occur in this site despite the deletion in its promoter region. Although rendered detectable by the high sensitivity of RT-PCR amplification, the extremely low level of 5-HT_2A mRNA in the small intestine of 5-HT_2A / mice may not be physiologically meaningful. The contraction of colonic smooth muscle evoked in 5-HT_2A +/+ mice by exogenous 5-HT was lacking in 5-HT_2A / animals, although the 5-HT_2A / colon contracted in response to ACh, suggesting that failure to respond to 5-HT is due to the absence of responding receptors. Because the 5-HT_2A receptor is expressed as early as E14, developmental, as well as physiological, abnormalities might be expected to accompany an absence of 5-HT_2A expression. Changes in the development of other members of the 5-HT₂ receptor family to compensate for the lack of 5-HT_2A expression, however, were not observed in 5-HT_2A / mice. 5-HT_2B expression in 5-HT_2A / animals could not be distinguished from that in their 5-HT_2A +/+ littermates, and 5-HT_2C expression was detected in neither. No defects were seen in the morphology or size of the 5-HT_2A / ENS or in the proportion of enteric neurons expressing an early developing (NOS-1) (5) or a late-developing phenotype (CGRP) (66). In addition, the rates of gastrointestinal transit and colorectal motility were not significantly different in 5-HT_2A / mice and their 5-HT_2A +/+ littermates. Although these are all relatively gross indices of enteric structure and function, their normality in the 5-HT_2A / mice suggests that the gut develops adequately and can function without 5-HT_2A receptors.

The fact that the motility of 5-HT_2A / bowel is good enough to support the life of an unchallenged mouse is not surprising. Despite the abundance of this 5-HT receptor subtype in the gut, there are no sources of endogenous 5-HT close to muscular 5-HT_2A receptors. The function of these receptors, therefore, may be more important in emergency or pathophysiological responses than in normal motility. A difference in pathophysiological effects between 5-HT_2A / and +/+ mice would probably require the application of noxious stimuli for detection. In contrast, the 5-HT_2A receptors on the basolateral surfaces of epithelial cells would be expected to be exposed to the 5-HT constitutively secreted by EC cells into the lamina propria. The absence of these receptors is associated with significant decreases in the size of enterocytes and the numbers of Paneth cells, suggesting that 5-HT_2A receptors function in the development and/or maintenance of end-stage epithelial cells generated in intestinal crypts (75). Development of intestinal epithelial cells is a continuous process that persists throughout life.

The 5-HT_2A receptors on smooth muscle, like those of the epithelium, may also have evolved a role in maintenance. Whereas the number of cells did not appear to be different, both longitudinal and circular layers of muscle were thinner in 5-HT_2A / mice than in their 5-HT_2A +/+ littermates; moreover, some smooth muscle cells showed evidence of degenerative change in the 5-HT_2A / animals that were not apparent in the 5-HT_2A +/+ mice. Because 5-HT_2A receptors are present both on nerve and muscle, these observations do not establish that muscular 5-HT_2A receptors are necessary for the long-term maintenance of the cells of the muscularis externa, although it is plausible that they are. An alternative possibility is that the innervation of the muscle coats trophically maintains muscle and functions abnormally when 5-HT_2A receptors are absent in the ENS. Because 5-HT_2A receptors are also located on enteric nerves within villi (Fig. 5, D-F), their absence from the intravillus nerves of 5-HT_2A / mice might also indirectly affect trophic influences, if any, that these nerves exert on their epithelial targets. Further study is needed to identify the function of 5-HT_2A receptors in the enteric plexuses. Whether effects of 5-HT_2A knockout are direct or indirect, the current observations suggest that 5-HT and its 5-HT_2A receptor are important for the maintenance and/or development of smooth muscle and epithelial cells in the bowel. The preservation of gross measures of motility despite the changes in the musculature suggest that there is a safety factor within which abnormalities in the intestinal musculature can be tolerated in animals not pathologically challenged.