5-HT3A receptor subunits in the rat medial nucleus of the solitary tract: subcellular distribution and relation to the serotonin transporter

5-HT3A receptor subunits in the rat medial nucleus of the solitary tract: subcellular distribution and relation to the serotonin transporter

Brain Research 1028 (2004) 156 – 169 www.elsevier.com/locate/brainres Research report 5-HT3A receptor subunits in the rat medial nucleus of the soli...

5MB Sizes 0 Downloads 20 Views

Brain Research 1028 (2004) 156 – 169 www.elsevier.com/locate/brainres

Research report

5-HT3A receptor subunits in the rat medial nucleus of the solitary tract: subcellular distribution and relation to the serotonin transporter Jie Huanga, Avron D. Spierb, Virginia M. Pickela,* a

Division of Neurobiology, Department of Neurology and Neuroscience, Weill Medical College of Cornell University, 411 East 69th Street, New York, NY, 10021, United States b Division of Neurobiology, Laboratory of Molecular Biology, MRC Centre, Cambridge, UK Accepted 8 September 2004 Available online 13 October 2004

Abstract The 5-hydroxytryptamine 3 (5HT3) receptor is a serotonin-gated ion channel implicated in reflex regulation of autonomic functions within the nucleus of the solitary tract (NTS). To determine the relevant sites for 5-HT3 receptor mediated transmission in this region, we used electron microscopic immunocytochemistry to examine the subcellular distribution of the 5HT3 receptor subunit A (5HT3A) in relation to the serotonin transporter (SERT) in the intermediate medial NTS (mNTS) of rat brain. The 5HT3A immunolabeling was detected in many axonal as well as somatodendritic and glial profiles. The axonal profiles included small axons and axon terminals in which the 5HT3A immunoreactivity was localized to membranes of synaptic vesicles and extrasynaptic plasma membranes. In dendrites and glia, the 5HT3A immunoreactivity was located on the plasma membranes or in association with membranous cytoplasmic organelles. The dendritic plasmalemmal 5HT3A labeling was prominent within and near excitatory-type synapses from terminals including those that resemble vagal afferents. The 5HT3A-labeled glial processes apposed 5HT3A-immunoreactive axonal and dendritic profiles, some of which also contained SERT. Terminals containing 5-HT3A and/or SERT were among those providing synaptic input to 5HT3A-labeled dendrites. Thus, 5HT3A has a subcellular distribution consistent with the involvement of 5-HT3 receptors in modulation of both presynaptic release and postsynaptic responses of mNTS neurons, some of which are serotonergic. The results further suggest that the neuronal as well as glial 5HT3 receptors can be activated by release of serotonin from presynaptic terminals or by diffusion facilitated by SERT distribution at a distant from the synapse. D 2004 Elsevier B.V. All rights reserved. Theme: Neurotransmitters, modulators, transporters, and receptors Topic: Serotonin receptors Keywords: Dorsal vagal complex; Cardiovascular regulation; Emesis; Autonomic function; Plasmalemmal uptake

1. Introduction The cardiovascular actions of serotonin (5-hydroxytryptamine, 5HT) are attributed, in part, to activation of multiple 5HT receptor subtypes within the medial nucleus of the solitary tract (mNTS) at the area postrema level [26,65]. These include not only G-protein coupled receptors, but also 5HT3 receptors that are ligand-gated ion channels [33].

* Corresponding author. Tel.: +1 212 570 2900; fax: +1 212 988 3672. E-mail address: [email protected] (V.M. Pickel). 0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2004.09.009

Activation of the 5HT3 receptor in this region increases blood pressure [9,34,40,60]. The 5HT3 receptors in the dorsal vagal complex, including the NTS and area postrema, also are highly implicated in emesis, which is consistent with the termination of abdominal vagal afferents in these regions [21]. The beneficial effects of 5HT3 antagonists in treating chemotherapy induced emetic reactions further support this conclusion [15–17,54]. Binding sites for 5HT3 receptors are abundant in the NTS [27,28] where lesions of sensory vagal afferents dramatically reduce the receptor binding, suggesting that 5HT3 receptors have a presynaptic distribution in these

J. Huang et al. / Brain Research 1028 (2004) 156–169

afferents [30,47], the majority of which are glutamatergic [57]. Involvement of 5HT3 receptors in presynaptic modulation of the release of glutamate, as well as other neurotransmitters is suggested in several brain regions [6,38,41,64–66]. A major presynaptic localization of 5HT3 receptors is consistent with low levels of the receptor mRNA in the NTS [59]. In addition, however, pharmacological studies suggest that 5HT3 receptors may also mediate postsynaptic excitation [42] in the NTS [65]. There is no ultrastructural evidence that 5HT3 receptors are located either pre- or postsynaptically in the mNTS, although other subtypes such as 5HT2A receptors are seen in both axon terminals and dendrites in this region [23]. Serotonin needed for 5HT3 receptor activation in the mNTS is derived principally from inputs from the raphe nuclei, containing serotonergic neurons [4,52]. In addition, serotonin is present in certain mNTS neurons [10,23] and in some of the visceral sensory afferents to this region [54]. Thus, there are multiple potential sources of 5HT in the mNTS. Since serotonergic neurons uniquely express the plasmalemmal serotonin transporter (SERT) [58], this transporter is a reliable marker for identity of the location of neurons having the capacity for serotonin uptake in the NTS and other brain regions. In addition, we have shown previously by electron microscopic immunocytochemistry that SERT is located on the plasma membranes and within the cytoplasm of many axon terminals in the mNTS [23]. In the present study, we examined in rat medial NTS the subcellular distribution of SERT and 5HT3A, a 5HT3 receptor subunit whose expression can produce functional 5-HT3 receptors even in the absence of the 5HT3 B subunit [38]. The results provide the first ultrastructural evidence that 5HT3 receptors are targeted to neuronal and glial membranes having potential access to the endogenous ligand.

2. Materials and methods 2.1. Animals and tissue preparation Experimental animal protocols were approved by the Weill Cornell Research Animal Resource Center according to NIH guideline. Male adult (250–300 g) Sprague–Dawley rats (Taconic Farms, Germantown, NY) were anesthetized by injection (i.p.) of sodium pentobarbital (100 mg/kg). The anesthetized animals were perfused via the ascending aorta first with 50 ml of 3.8% acrolein, followed by 200 ml of 2% paraformaldehyde dissolved in 0.1 M phosphate buffer (PB), pH 7.4. The rat brainstems were removed and postfixed in 2% paraformaldehyde at room temperature (RT) for 30 min. The blocks of brainstems were cut on a Vibratome (Leica, Rockleigh, NJ) into sections of 40 Am through the NTS, and these sections were collected into 0.1 M PB. An incubation of the sections in 1% sodium borohydride was performed to neutralize the aldehydes.

157

The sections were rinsed and transferred into a cryprotection buffer containing 25% sucrose and 2.5% glycerol for 15 min. To increase the permeability, this was followed by a rapid freeze–thaw procedure that includes sequential immersion of sections in liquid freon, liquid nitrogen, and room temperature PB. A polyclonal antiserum was raised in rabbit [53] against a 13 amino acid peptide sequence contained within the extracellular domain of the cloned 5HT3A receptor subunit [31,53]. In the initial characterization, Spier et al. [53] showed that this antiserum recognizes both recombinantly expressed 5-HT3A receptors in HEK293 cells and native 5HT3A receptors in N1E-115 cells, and labels a 54 kDa band corresponding to the expected molecular weight of 5-HT3A receptors in Western blots of N1E-115 cell membranes. The specificity of the antiserum was further demonstrated by 5HT3A receptor binding being eliminated following preincubation of the serum with the 13 amino acid 5-HT3A derived peptide against which the antibody was generated [53]. A polyclonal SERT antiserum against 20 amino acids in the C-terminal of the human SERT was raised in goat (Santa Cruz Biotechnology; SC-1458). This antiserum was affinity-purified and characterized by Western blot in brain tissue (Santa Cruz Biotechnology). The commercial SERT antiserum shows the same labeling pattern as seen with two other anti-SERT antisera in the rat forebrain [44] and in the NTS [23]. 2.2. Immunocytochemical labeling A detailed description of the procedures for electron microscopic immunocytochemical labeling was published previously [11]. In short, the vibratome sections were incubated in 0.1 M Tris-buffered saline (TBS), pH 7.6, containing 0.1% bovine serum albumin (BSA). After 30 min incubation, these sections were transferred into a solution containing primary 5HT3A antiserum for single labeling or into a buffer containing both 5HT3A and SERT antisera for double labeling. The antisera concentrations were 1:5000 for 5HT3A receptor immunoperoxidase labeling and 1:1000 for 5HT3A or SERT immunogold labeling. Following incubation in the primary antisera for 48 h at 4 8C, the sections used for peroxidase detection of 5HT3A in single or dual labeling experiments were rinsed, and subsequently, incubated in a biotin conjugated secondary antiserum, donkey anti-rabbit IgG (1:400, Vector Laboratories, Burlingame, CA) for 30 min. These sections were rinsed again and transferred to a solution containing avidin biotin peroxidase complex (ABC) (1:100, Vectastain Elite Kit, Vector Laboratories) [22] for 30 min. The bound peroxidase was visualized by reaction in a Tris-buffered solution containing 0.022% 3,3V diaminobenzidine and 0.003% hydrogen peroxide for 6 min. For immunogold detection of SERT in dual labeling experiments, the sections were blocked in 0.01 M phosphate-buffered saline (PBS), pH 7.4, containing 0.8% BSA and 0.1% gelatin for

158

J. Huang et al. / Brain Research 1028 (2004) 156–169

10 min. These sections were then transferred to a solution containing one nm gold conjugated rabbit anti-goat IgG (1:50, AuroProbeOne; Amersham, Arlington Height, IL) for 2 h. All incubations were at room temperature with constant agitation provided by a VSOS-4 shaker (Shelton Scientific, Shelton, CT). The bound gold was further attached to the tissue by fixation in 2% glutaraldehyde for 10 min. The particles were enlarged by silver intensification using a 1:1 mixture of a silver solution from an intensification kit (IntenSE-M kit, Amersham) for 6–8 min. The same protocol was used for single immunogold labeling of 5HT3A except for omission of the SERT antiserum and replacement of the secondary IgG with one nm gold conjugated goat anti-rabbit IgG (Amersham).

Adobe PhotoShop (version 6.0, Adobe Systems Inc, Mountain View, CA) and Microsoft Powerpoint 2001 were used to adjust contrast and to assemble and label the composite figures.

3. Results The 5HT3A immunolabeling was present in axons and axon terminals as well as somatodendritic and glial profiles within the mNTS. A few of the 5HT3A-labeled axonal and somatodendritic profiles contained SERT, although SERT was principally seen in axon terminals without 5-HT3 immunoreactivity. The SERT-labeled terminals formed synapses with unlabeled or, less commonly with 5-HT3 labeled dendrites.

2.3. Electron microscopic processing and data analysis 3.1. Axonal 5HT3A distribution and relation to SERT For electron microscopy, the sections were processed by incubation in 0.1 M PB containing 2% osmium tetroxide for 60 min. After dehydration in serial dilutions of ethanol and three changes of propylene oxide, these sections were incubated in a solution containing a 1:1 mixture of propylene oxide and EPON (EM bed-812; Electron Microscopy Science, Fort Washington, PA) on a rotator for 24 h. These sections then were flat embedded in EPON between two sheets of Aclar plastic film, and placed in an oven (68 8C) for 48 h to harden the EPON. The flat embedded sections were examined with a light microscope to locate the mNTS region to be further examined by electron microscopy. For this, the mNTS region at the area postrema level was selected using the atlas of Paxinos and Watson (1986). The tissue was cut on an ultramicrotome (Ultratome, NOVA, LKB-Productor, Bromma, Sweden), and thin sections (50 nm) from the surface of the vibratome sections were collected on 400 mesh copper grids. Thin sections on the grids were further stained with uranyl acetate and Reynolds’ lead citrate. After air-drying, the thin sections were systematically examined and photographed using a Advanced Microscopy Techniques (AMT) digital camera installed on a Philips CM10 Electron Microscope (Mahwah, NJ). Dendrites, axons, axon terminals and glial profiles were classified according to Peters et al. [43]. In large, multisynaptic terminals, such as those of vagal afferents, we examined serial sections whenever necessary to distinguish the ultrastructural features of axon terminals and their targets. Profiles were considered as immunoperoxidase labeled when they contained a granular precipitate imparting an electron density that was considerably greater than that seen in comparable structures in the neuropil. Those profiles containing at least two immunogold particles were considered as immunolabeled, although those immunogold-labeled for SERT usually showed considerably more than two particles. Images acquired with the digital camera were saved onto a pocket drive.

Intense immunoperoxidase 5HT3A labeling was seen in small unmyelinated axons and axon terminals. In these axons, the 5HT3A immunoreactivity (IR) was prominently associated with large (100–150 nm) dense core vesicles (Fig. 1A). Diffuse labeling was also seen on the plasma membrane and throughout the axoplasm (Fig. 1B). The 5HT3A-labeled axons were apposed by unlabeled unmyelinated axons and by glial processes, some of which were 5HT3A immunoreactive (Fig. 1A). Terminals containing 5HT3A IR showed considerable variation in their size, vesicle content, and synaptic specializations. These terminals ranged from 0.5 to 2.0 Am in cross-sectional diameter and contained either all small (40–50 nm) clear vesicles or small clear vesicles and varying numbers of large dense core vesicles (Figs. 1C, 2, and 3A–C). Those containing abundant dense core vesicles often lacked recognizable synapses (Fig. 1C). The 5HT3Aimmunoreactive terminals with few dense core vesicles formed asymmetric or symmetric synapses with dendrites. Even the largest terminals having ultrastructural features of vagal afferents [1] often showed only one 5HT3Aimmunolabeled vesicle within a single plane of section (Fig. 2A). Within axonal profiles, 5HT3A labeling in association with large dense core vesicles or segments of plasma membrane was most readily identified using the highly sensitive immunoperoxidase method (Figs. 1C and 2A,B). The 5HT3A-labeled large dense vesicles and aggregates of smaller vesicles (Fig. 2C) were located around the perimeter or toward the interior of axon terminals. The immunoreactive terminals frequently were apposed by either 5HT3A labeled (Fig. 2A) or unlabeled (Fig. 2B) glial processes. There was no apparent relationship between the location of the 5-HT3 immunoreactive vesicles and the apposing glial processes. Immunolabeled large dense core vesicles were observed, however, at points of contact with other labeled axon terminals (Fig.

J. Huang et al. / Brain Research 1028 (2004) 156–169

159

Fig. 1. Electron micrographs illustrating 5HT3A immunoperoxidase distribution in small axons and a non-synaptic axon terminal or glia in the intermediate medial nucleus of the solitary tract (mNTS). (A) A small unmyelinated axon (5HT3-a) shows 5HT3A peroxidase reaction product in association with a dense core vesicle (straight arrow). The labeled axonal profile is apposed by a small glial profile (5HT3-g) that also contains immunoperoxidase for 5HT3A. Another nearby axon (ax) is unlabeled. (B) Immunoperoxidase 5HT3A labeling is seen within a small unmyelinated axon (5HT3-a) that is apposed by unlabeled axons (ax) and unlabeled glial processes indicated by asterisks. (C) The 5HT3A is localized to one of the dense core vesicles in an axon terminal that is without a recognizable synaptic junction and contains many other unlabeled dense core vesicles (dcv). The peroxidase reactions associated with the axonal vesicle and also seen on the membrane of an adjacent astrocytic process are indicated by arrows. Asterisks indicate the labeled and another unlabeled glial process joined by a gap junction (GJ). Scale bar=0.5 Am.

2B). In contrast with the dense core vesicles, segments of plasma membrane immunolabeled for 5HT3A were apposed by glia (Fig. 2B). The subcellular distribution of 5HT3A receptors in both axonal and glial profiles was confirmed by the less sensitive, but non-diffusible immunogold silver labeling method (Fig. 3). The gold particles were located on the plasma membrane or near synaptic vesicles in these terminals, although the gold–silver deposits were relatively large and may have obscured many of the underlying vesicles. In glial processes, the immunogold for 5HT3A also appeared to be localized either on the plasma membrane or in association with endomembranes (Fig. 3D). These glial processes were seen near 5HT3A-labeled or unlabeled terminals. In one favorable section, the unlabeled terminal formed an asymmetric synapse with a dendrite showing a perisynaptic distribution of 5-HT3A immunogold (Fig. 3D). The 5HT3A-immunoreactivity also was seen in many of the dendrites that received input from 5HT3A-labeled terminals (Fig. 3A–C). SERT immunoreactivity was mainly localized on the plasma membranes of axonal profiles in the mNTS, as

has been previously described in this brain region [23]. Many of the SERT-labeled axonal profiles contained large dense core vesicles (Fig. 4A), some of which showed 5HT3A IR (Fig. 4B). Co-expression with 5HT3A was seen in 21% (20/96) of the SERT-labeled axonal profiles. The SERT-immunolabeled axon terminals were apposed or near other neuronal and glial profiles that contained 5HT3A IR, but usually were without recognizable synaptic junctions. Approximately 15% (7/49) of the SERT-labeled terminals, however, formed asymmetric synapses on dendritic profiles containing 5HT3A IR that could be seen in a single plane of section (Fig. 4C). Other labeled terminals formed junctions that appeared symmetric, having equally dense pre- and post-synaptic membrane specializations. 3.2. Somatodendritic localization of 5HT3A and relation to SERT The 5HT3A IR was preferentially detected in small dendritic profiles including spines and more rarely in larger somatodendritic profiles as seen by using either

160

J. Huang et al. / Brain Research 1028 (2004) 156–169

Fig. 2. Vesicular 5HT3A immunoperoxidase labeling (straight arrows) in synaptic terminals. (A) 5HT3A is localized to a dense core vesicle (dcv) in a large terminal (5HT3-t), making multiple contacts (square block arrows) that are not clearly distinguishable with respect to their synaptic specialization. Apposed structures include an unlabeled dendrite (ud) and two unlabeled terminals (ut1–2), as well as a glial profile (5HT3-g) also containing 5HT3A immunolabeling. The glial immunoreactivity appears to be associated with the plasma membrane and an endomembrane (em). (B) 5HT3A immunoreactivity is localized to a dense core vesicle (arrow) in an axon terminal (5HT3-t1) contacting a second terminal (5HT3-t2) in which the peroxidase labeling (arrow) is distributed along the plasma membrane and nearby small synaptic vesicles. Unlabeled dense core vesicle (dcv) is also seen in 5HT3-t1, and the plasmalemmal labeling in 5HT3t2 is located near an appositional contact by an unlabeled glial process (*). (C) 5HT3A immunolabeling is associated with several small vesicles in a terminal (5HT3-t) forming an asymmetric synapse (curved arrow) with an unlabeled dendrite (ud). Scale bar=0.5 Am.

immunoperoxidase (Fig. 5) or immunogold–silver (Fig. 6) labeling methods. In smaller dendrites, the 5HT3A IR was localized within and near postsynaptic membranes (Fig. 5A–C) or in association with endomembranes resembling smooth endoplasmic (Fig. 5D). These synapses appeared largely asymmetric (Fig. 6A). The synaptic inputs were largely from unlabeled terminals containing

contained either almost all small clear vesicles (Fig. 6A,B) or several large dense core vesicles (Fig. 6C). In addition to the synaptic plasmalemmal and endomembrane distributions, the 5-HT3A immunolabeling was sometimes localized to non-synaptic portions of the dendritic plasma membrane contacted by unlabeled glial process (Fig. 6B).

J. Huang et al. / Brain Research 1028 (2004) 156–169

161

Fig. 3. Pre- and postsynaptic or glial immunogold labeling (block arrows) for 5HT3A in mNTS. (A) Immunogold 5HT3A is localized near the plasma membrane in a large vagal-like axon terminal (5HT3-t) forming an asymmetric synapse with a dendrite (5-HT3-d). The dendrite has a mainly non-synaptic plasmalemmal labeling for 5HT3A. (B) 5HT3A immunogold labeling is distributed in presynaptic portions of a large axon terminal (5HT3-t1) that forms asymmetric synapses with a 5HT3A labeled (5HT3-d) and unlabeled (ud) dendrites. The terminal apposes a second terminal (5HT3-t2) also containing immunogold for 5HT3A. (C) Immunogold 5HT3A labeling is distributed near a dense core vesicle (dcv) in an axon terminal that forms an asymmetric synapse with a dendrite (5HT3-d). The postsynaptic dendrite also shows a 5HT3A immunogold–silver deposit located near an endomembrane (em) beneath the postsynaptic density. (D) Immunogold 5HT3A is localized to the plasma membrane and endomembrane (em) within a glial profile (5HT3-g) covering portions of an unlabeled axon terminal (ut). The terminal forms an asymmetric synapse with a dendrite (5HT3-d) showing perisynaptic 5HT3A immunogold. Curved arrows=asymmetric synapses. Scale bar=0.5 Am.

162

J. Huang et al. / Brain Research 1028 (2004) 156–169

Fig. 4. Differential distribution and co-expression of 5HT3A and SERT in axonal profiles in the mNTS. (A) Immunoperoxidase 5HT3A labeling (thin arrow) is seen in association with a dense core vesicle (dcv) in an axon terminal (5HT3-t), while SERT immunogold labeling (block arrows) is located mainly on the plasma membranes of nearby terminals (SERT-t1 and SERT-t2). SERT-t1 is apposed by a glial process (*) that is also immunolabeled for 5HT3A. (B) Both 5HT3A immunoperoxidase labeling (thin arrow) and SERT gold particles (block arrows) are co-localized in an axonal profile (5HT3 and SERT-t). The 5HT3A immunoperoxidase labeling is prominently associated with a large dense core vesicle (dcv), while the immunogold particles for SERT (block arrows) are distributed along the plasma membranes. (C) Immunogold labeling (block arrows) for SERT is located in a terminal (SERT-t) that contains a large dense core vesicle (dcv) and forms a synapse with a spine (5HT3-s) showing diffuse 5HT3A immunoperoxidase labeling. The peroxidase reaction product (thin arrows) is slightly more accumulated on postsynaptic and non-synaptic plasma membranes of the dendritic spine. Scale bar=0.5 Am.

In somata, 5HT3A IR was associated with endomembranes, smooth endoplasmic reticulum or Golgi lamellae, which in some examples was flanked by several large dense core vesicles (Fig. 6D). Somata with cytoplasmic labeling of 5HT3A often also showed nuclear labeling as seen by either immunoperoxidase (not shown) or immunogold (Fig. 6D) methods. In other somata, however, only

cytoplasmic labeling of 5HT3A was seen (Fig. 6E). In large proximal dendrites, the 5HT3A immunoreactivity was principally associated with endomembranes or smooth endoplasmic reticulum and Golgi lamellae (Fig. 7A). A similar distribution of 5HT3A was seen in a few dendritic profiles that co-expressed 5HT3A and SERT (Fig. 7B).

J. Huang et al. / Brain Research 1028 (2004) 156–169

163

Fig. 5. Dendritic 5HT3A immunoperoxidase labeling of postsynaptic plasma membranes and endomembranes within the mNTS. (A and B) In small dendritic profiles, the peroxidase reaction products for the 5HT3A (arrows) are seen on the dendritic plasma membranes contacted by medium-size unlabeled terminals (ut). The terminals contain a mixture of large dense core vesicles (dcv) and small clear vesicles (scv) (A) or all small clear vesicles (B). Peroxidase 5HT3A immunoreactivity (straight arrows) is located on endomembranes (em) in the cytoplasm near the postsynaptic density and nearby mitochondrion (m) in B. (C) An unlabeled terminal (ut) is presynaptic to a 5HT3A immunolabeled dendrite (5HT3-d) and apposed by a 5HT3A-labeled glial profile (5HT3-g). The glial labeling appears to be associated with endomembranes (em), whereas the dendritic labeling (arrow) is within and near the postsynaptic density. (D) The immunoperoxidase reaction product (straight arrow) representing the 5HT3A labeling is seen in a dendrite (5HT3-d) as well as in the presynaptic terminal (5HT3-t). The dendritic 5HT3A labeling is associated with endomembranes (em), while the presynaptic labeling is localized to a perisynaptic segment of the axonal plasma membrane. The 5HT3A immunoreactivity in the terminal is also associated with small clear vesicles (scv). Scale bar=0.5 Am.

164

J. Huang et al. / Brain Research 1028 (2004) 156–169

Fig. 6. Dendritic and somatic or glial immunogold labeling (block arrows) for 5HT3A in mNTS. (A) The 5HT3A immunoreactivity is located beneath an asymmetric synapse (curved arrow) in a dendrite (5HT3-d) contacted by an unlabeled astrocytic process (*). (B) Immunogold 5HT3A is localized to non-synaptic portions of the plasma membrane of a dendrite (5HT3-d) in contact with an unlabeled astrocytic process (*). The plasma membrane in a nearby glial profile (5HT3-g) is immunolabeled for 5HT3A. (C) Immunogold 5HT3A localization to an endomembranes (em) in a dendrite (5HT3-d1) that is contacted by an axon terminal containing an unlabeled dense core vesicle (dcv). Labeling also is localized to non-synaptic portions of the plasma membrane in another dendrite (5HT3-d2). (D) Immunogold 5HT3A particles are distributed near Golgi lamellae and a dense core vesicle (dcv) in the cytoplasmic region just outside the outer nuclear membrane of a soma (5HT3-som) in which gold particles are also seen within the nucleus (nuc). (E) Localization of 5HT3A immunogold near an endomembrane (em) in the cytoplasm of a soma (5HT3-som), whose nucleus is without labeling. The somatic plasma membrane is apposed by an unlabeled glial process (*). Scale bar=0.5 Am.

J. Huang et al. / Brain Research 1028 (2004) 156–169

4. Discussion Our results provide the first ultrastructural evidence that 5HT3A is present in unmyelinated axons and morphologically diverse axon terminals including those that express

165

SERT in the rat mNTS. These results are consistent with the involvement of 5HT3 receptors in the presynaptic release of multiple neurotransmitters including serotonin in this brain region. Moreover, we show that 5HT3A subunits are localized on the plasma membrane of both axonal and

Fig. 7. Somatodendritic immunoperoxidase 5HT3A and immunogold SERT distributions in the mNTS. (A) In a somatic profile (5HT3-som), immunoperoxidase labeling for 5HT3A (thin arrows) is associated with the Golgi lamellae (G) and membranes near a mitochondrion (m) and endomembranes resembling smooth endoplasmic reticulum (ser). 5HT3 peroxidase immunoreactivity is also seen within a small dendrite (5HT3-d) within the neuropil. (B) 5HT3A immunoperoxidase (thin arrows) and SERT immunogold (block arrows) are co-localized in the same dendritic profile (5HT3 and SERT-d). 5HT3A and SERT immunoreactivities are seen in the central cytoplasm and near the plasma membrane. Scale bar=0.5 Am.

166

J. Huang et al. / Brain Research 1028 (2004) 156–169

dendritic profiles, but also prominently associated with axonal dense core vesicles and dendritic endomembranes that may be involved in trafficking of functional subunits to receptive sites in the NTS. The afferent terminals to the 5HT3A-labeled dendrites included those that expressed SERT suggesting that postsynaptic 5HT3 receptors in the NTS can be activated by locally released serotonin. Other presynaptic terminals, however, were unlabeled or also contained 5HT3A supporting the notion that 5HT3 receptors play a role in modulating the presynaptic release as well as the postsynaptic responses in non-serotonergic neurons. These distributions, together with the selective localization of 5HT3A in perisynaptic glial profiles suggest that activation of 5HT3 receptors in the NTS can have a major impact on synaptic transmission and neuroglial signaling directly affecting autonomic function. 4.1. Methodological consideration Immunolabeling of 5HT3A in the mNTS is consistent with the 5HT3 receptor distribution seen in ligand binding studies, suggesting detection of functional sites for receptor activation. The levels of 5HT3 mRNA in the NTS are relatively low, however, which may indicate low levels of synthesis or a localization of the receptor mainly in afferent terminals derived from neurons in other brain regions [59]. The low levels of synthesis may account for sparse light microscopic distribution of immunoreactivity [36]. Both the peroxidase and gold–silver labeling methods showed a sparse membrane distribution of 5HT3A immunoreactivity, which was often associated with vesicles or plasmalemmal segments below the resolution of the light microscope. The restricted subcellular distribution of 5HT3A immunoreactivity may have contributed to an underestimation of the relative abundance of profiles containing 5HT3A in the mNTS. This limitation also may account, in part, for the infrequencies of detecting 5HT3A and SERT immunoreactivities within the same or synaptically linked neurons. We did, however, optimize the recognition of both antigens by using the highly sensitive avidin–biotin peroxidase for localization of 5-HT3A and immunogold for identity of the more prevalent SERT in the dual labeling experiments. Although less sensitive, the lack of diffusion of the immunogold–silver products enabled a more clearly defined subcellular distribution of 5-HT3A within neurons and glia of the NTS. 4.2. Axonal localization of 5HT3A The present localization of 5HT3A immunoreactivity in small axons and axon terminals supports earlier binding and lesion studies indicting a presynaptic receptor distribution in the mNTS [30,47]. In addition, we observed that the 5HT3A-labeled terminals are morphologically heterogeneous in synaptic specializations. The heterogeneity is con-

sistent with the finding that activation of 5HT3 receptors modulates the release of a variety of neurotransmitters, including 5HT [6], glutamate [3], dopamine [38,41], and acetylcholine [13]. The labeled terminals formed mainly asymmetric, but also symmetric synapses that are typical of those containing excitatory or inhibitory neurotransmitters, respectively [1,61]. In both axon and axon terminals, the 5HT3A labeling was frequently associated with dense core vesicles. These vesicles are known to store neuropeptides and also to express the vesicular monoamine transporter (VMAT2) [35,39,45]. The prominent location of 5HT3A immunoreactivity in association with large dense core vesicles and less frequently on the plasma membranes suggests that the availability of the 5HT3 receptor for binding to extracellular ligands may depend, in part, on the exocytotic release of a peptide or modulator (serotonin, for example) from the vesicles as has been suggested for certain other receptors, particularly the delta opioid receptors [68]. Thus, 5HT3A on membranes of dense core vesicles may be delivered to the cell surface during the exocytosis of neuropeptide or monoamine transmitters. Fusion of the vesicles membrane would permit activation of the receptor by extracellular 5HT, thus allowing Ca2+ influx and ensuring changes in transmitter release [19,46,63]. Activation of 5HT3 receptors on the axonal plasma membrane may also directly affect membrane excitability [14,31]. The detection of the SERT in terminals containing 5HT3A suggests that 5HT3 receptors play a role in the modulation of the release of serotonin or a co-existing transmitter. Both glutamate and substance P are present in visceral afferent terminals and are also co-expressed with 5HT in central neurons [8,18]. The 5HT3 receptors, even though present in serotonergic neurons, may act largely as a heteroreceptors affecting the release of other neurotransmitters [7]. Both the morphology of the 5HT3Alabeled terminals and reduced receptor binding in the NTS following vagal lesions [30,47], suggest that at least some of these terminals containing 5HT3A along with serotonin and/or glutamate are derived from vagal sensory neurons. Others may come from raphe serotonergic neurons [24]. 4.3. Somatodendritic localization of the 5HT3A In smaller dendrites and spines, the 5HT3A immunolabeling was associated with the postsynaptic specialization at asymmetric synapses, which is consistent with involvement of 5HT3 receptors in excitatory postsynaptic responses [42]. Others, however, appeared symmetric, which are typical of inhibitory terminals including GABAergic terminals in the NTS [29,61]. The apparent association of the 5HT3A with inhibitory-type synapses is not consistent with known facilitation mediated by the 5HT3 receptor. The apparent symmetry may reflect a preferential distribution of 5HT3 receptors near the lateral edge of asymmetric synapses,

J. Huang et al. / Brain Research 1028 (2004) 156–169

which resemble symmetric junctions when viewed in transverse sections. The 5HT3 receptor also may be coupled with a different secondary effector mediating inhibitory actions at certain synapses [2]. In somata and large dendrites, 5HT3A was localized prominently on endomembranes and Golgi lamellae, both of which are implicated in the synthesis and trafficking of surface receptors [56]. The labeled endomembranes may identify 5HT3A subunits in transit to the plasma membrane or portions of the endosomal pathway of internalized receptors [62]. In addition, both endomembranes and neuronal nuclei, which sometimes expressed 5HT3A immunoreactivity, are 5HT3-sensitive calcium storage sites [51]. The 5HT3A-immunoreactive somatodendritic profiles received synaptic input from SERT-labeled terminals, suggesting that the 5HT3 receptor can be activated by synaptic release of serotonin [14]. In addition, however, many of the presynaptic terminals were without SERT immunoreactivity, indicating that the extracellular serotonin required for receptor activation may be derived from more distantly located axons (i.e. volume transmission) [20]. Conceivably, the limited detection of SERT near profiles containing 5HT3A may reflect local regulation of SERT by released serotonin or by retrograde signaling molecules such as nitric oxide present in the NTS neurons and their afferents [5,67]. 4.4. Glial 5HT3A distribution The localization of 5HT3A in glial cells is consistent with the known involvement of glial 5HT3 receptors in ion channels mediating K+ and Na+ currents [32,48,49]. These ion channels on glia are critical in maintaining extracellular K+ homeostasis [25]. In addition, we demonstrated the presence of the 5HT3A on plasma membranes of perisynaptic glia. Such receptors may be important in determining the extracellular availability of neurotransmitters, maintaining local synaptic environment, and modulating neuronal transmission [55]. This is in line with the existence of K+, Na+ and Ca2+ channels in perisynaptic glia [12,50]. The 5HT3A-labeled perisynaptic glial processes often apposed similarly labeled dendritic or axonal profiles, suggesting dual involvement of the receptor in coordinating the neuronal and glial activities [55]. Further studies are needed to investigate these novel-signaling mechanisms. 4.5. Conclusion The localization of the 5HT3A in vagal-like afferents and in somatodendritic profiles in the NTS is comparable to the distribution of the 5HT3 receptors pre- and postsynaptically in the dorsal horn of the spinal cord [37]. Thus, 5HT3 receptors have subcellular distributions supporting a major involvement in modulation of synaptic transmission at the

167

first central synapse in both visceral and somatic sensory pathways. Moreover, in the mNTS, the glial localizing of 5HT3A suggests a role for non-neuronal 5HT3 receptors in mechanisms affecting synaptic transmission. The finding also indicates that 5HT3A is expressed in dendrites postsynaptic to terminals with or without detectable SERT. The latter finding is consistent with either volume serotonin transmission [20] or local regulation of the plasmalemmal expression of SERT in this brain region [5]. The present localization of 5HT3A provides insight into the cellular substrates mediating cardiovascular and emetic responses produced by activation of 5HT3 receptors in the dorsal vagal complex [21,60].

Acknowledgments This work was supported by grants from NIH (HL18974) and NIMH (MH48776) to VMP.

References [1] S. Aicher, S. Sharma, V. Pickel, N-methyl-d-aspartate receptors are present in vagal afferents and their dendritic targets in the nucleus tractus solitarius, Neuroscience 91 (1999) 119 – 132. [2] C. Ashby, Y. Minabe, E. Edwards, R. Wang, 5-HT3-like receptors in the rat medial prefrontal cortex: an electrophysiological study, Brain Res. 550 (1991) 181 – 191. [3] M. Ashworth-Preece, B. Jarrott, A. Lawrence, 5-Hydroxytryptamine3 receptor modulation of excitatory amino acid release in the rat nucleus tractus solitarius, Neurosci. Lett. 191 (1995) 75 – 78. [4] E. Azmitia, M. Segal, An autoradiographic analysis of the differential ascending projections of the dorsal and median raphe nuclei in the rat, J. Comp. Neurol. 179 (1978) 641 – 667. [5] R. Blakely, S. Ramamoorthy, S. Schroeter, Y. Qian, S. Apparsundaram, A. Galli, F.L. De, Regulated phosphorylation and trafficking of antidepressant-sensitive serotonin transporter proteins, Biol. Psychiatry 44 (1998) 169 – 178. [6] P. Blier, C. Bouchard, Functional characterization of a 5-HT3 receptor which modulates the release of 5-HT in the guinea-pig brain, Br. J. Pharmacol. 108 (1993) 13 – 22. [7] P. Blier, P. Monroe, C. Bouchard, D. Smith, D. Smith, 5-HT3 receptors which modulate [3H]5-HT release in the guinea pig hypothalamus are not autoreceptors, Synapse 15 (1993) 143 – 148. [8] V. Bouryi, D. Lewis, The modulation by 5-HT of glutamatergic inputs from the raphe pallidus to rat hypoglossal motoneurones, in vitro, J. Physiol. 553 (2003) 1019 – 1031. [9] J. Callera, C. Sevoz, R. Laguzzi, B. Machado, Microinjection of a serotonin3 receptor agonist into the NTS of unanesthetized rats inhibits the bradycardia evoked by activation of the baro- and chemoreflexes, J. Auton Nerv. Syst. 63 (1997) 127 – 136. [10] L. Calza, L. Giardino, R. Grimaldi, M. Rigoli, H. Steinbusch, M. Tiengo, Presence of 5-HT-positive neurons in the medial nuclei of the solitary tract, Brain Res. 347 (1985) 135 – 139. [11] J. Chan, C. Aoki, V.M. Pickel, Optimization of differential immunogold–silver and peroxidase labeling with maintenance of ultrastructure in brain sections before plastic embedding, J. Neurosci. Methods 33 (1990) 113 – 127. [12] N. Cholet, L. Pellerin, P. Magistretti, E. Hamel, Similar perisynaptic glial localization for the Na+, K+-ATPase alpha 2 subunit and the glutamate transporters GLAST and GLT-1 in the rat somatosensory cortex, Cereb. Cortex 12 (2002) 515 – 525.

168

J. Huang et al. / Brain Research 1028 (2004) 156–169

[13] D. Crespi, M. Gobbi, T. Mennini, 5-HT3 serotonin hetero-receptors inhibit [3H]acethylcholine release in rat cortical synaptosomes, Pharmacol. Res. 35 (1997) 351 – 354. [14] V. Derkach, A. Surprenant, R. North, 5-HT3 receptors are membrane ion channels, Nature 339 (1989) 706 – 709. [15] R. de Wit, Current position of 5HT3 antagonists and the additional value of NK1 antagonists; a new class of antiemetics, Br. J. Cancer 88 (2003) 1823 – 1827. [16] V. Dieras, J. Extra, B.M. Le, I. Madelaine, M. Espie, M. Marty, 5HT3antagonist receptors: their role in drug-induced emesis, Nouv. Rev. Fr. Hematol 32 (1990) 449 – 452. [17] D. Gandara, F. Roila, D. Warr, M. Edelman, E. Perez, R. Gralla, Consensus proposal for 5HT3 antagonists in the prevention of acute emesis related to highly emetogenic chemotherapy. Dose, schedule, and route of administration, Support Care Cancer 6 (1998) 237 – 243. [18] M. Hancock, Substance P-immunoreactive processes on 5HT/SPimmunoreactive medullary cells., Brain Res. Bull. 13 (1984) 559 – 563. [19] C. Heinemann, R.L. von, R. Chow, E. Neher, A two-step model of secretion control in neuroendocrine cells, Pflugers Arch. 424 (1993) 105 – 112. [20] M. Herkenham, Extrasynaptic receptors and parasynaptic communication in the brain, Brain Res. Bull. 50 (1999) 351 – 352. [21] P. Hornby, Central neurocircuitry associated with emesis, Am. J. Med. 111 (Suppl. 8A) (2001) 106S – 112S. [22] S. Hsu, L. Raine, H. Fanger, Use of avidin–biotin–peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures, J. Histochem. Cytochem. 29 (1981) 577 – 580. [23] J. Huang, V. Pickel, Serotonin transporters (SERTs) within the rat nucleus of the solitary tract: Subcellular distribution and relation to 5HT2A receptors, J. Neurocytol. 31 (2002) 667 – 679. [24] Z. Jackson, B. Stringer, G. Foster, Identification of 5-HT receptor subtypes in a homogeneous population of presumptive serotoninergic neurones, Neuropharmacology 36 (1997) 543 – 548. [25] D. Janigro, S. Gasparini, R. D’Ambrosio, G. McKhann, F.D. Di, Reduction of K+ uptake in glia prevents long-term depression maintenance and causes epileptiform activity, J. Neurosci. 17 (1997) 2813 – 2824. [26] M. Kalia, M. Mesulam, Brain stem projections of sensory and motor components of the vagus complex in the cat: I. The cervical vagus and nodose ganglion, J. Comp. Neurol. 193 (1980) 435 – 465. [27] G. Kilpatrick, B. Jones, M. Tyers, Identification and distribution of 5HT3 receptors in rat brain using radioligand binding, Nature 330 (1987) 746 – 748. [28] A. Laporte, T. Koscielniak, M. Ponchant, D. Verge, M. Hamon, H. Gozlan, Quantitative autoradiographic mapping of 5-HT3 receptors in the rat CNS using [125I]iodo-zacopride and [3H]zacopride as radioligands, Synapse 10 (1992) 271 – 281. [29] W. Len, J. Chan, GABAergic neurotransmission at the nucleus tractus solitarii in the suppression of reflex bradycardia by parabrachial nucleus, Synapse 42 (2001) 27 – 39. [30] R. Leslie, D. Reynolds, P. Andrews, D. Grahame-Smith, C. Davis, J. Harvey, Evidence for presynaptic 5-hydroxytryptamine3 recognition sites on vagal afferent terminals in the brainstem of the ferret, Neuroscience 38 (1990) 667 – 673. [31] A. Maricq, A. Peterson, A. Brake, R. Myers, D. Julius, Primary structure and functional expression of the 5HT3 receptor, a serotoningated ion channel, Science 254 (1991) 432 – 437. [32] G. McKhann, R. D’Ambrosio, D. Janigro, Heterogeneity of astrocyte resting membrane potentials and intercellular coupling revealed by whole-cell and gramicidin-perforated patch recordings from cultured neocortical and hippocampal slice astrocytes, J. Neurosci. 17 (1997) 6850 – 6863. [33] N. Merahi, R. Laguzzi, Cardiovascular effects of 5HT2 and 5HT3 receptor stimulation in the nucleus tractus solitarius of spontaneously hypertensive rats, Brain Res. 669 (1995) 130 – 134.

[34] N. Merahi, H. Orer, A. Laporte, H. Gozlan, M. Hamon, R. Laguzzi, Baroreceptor reflex inhibition induced by the stimulation of serotonin3 receptors in the nucleus tractus solitarius of the rat, Neuroscience 46 (1992) 91 – 100. [35] M. Morales, F. Bloom, The 5-HT3 receptor is present in different subpopulations of GABAergic neurons in the rat telencephalon, J. Neurosci. 17 (1997) 3157 – 3167. [36] M. Morales, E. Battenberg, L.L. de, P. Sanna, F. Bloom, Cellular and subcellular immunolocalization of the type 3 serotonin receptor in the rat central nervous system, Brain Res. Mol. Brain Res. 36 (1996) 251 – 260. [37] M. Morales, N. McCollum, E. Kirkness, 5-HT(3)-receptor subunits A and B are co-expressed in neurons of the dorsal root ganglion., J. Comp. Neurol. 438 (2001) 163 – 172. [38] H. Neijt, H. Vijverberg, d.B.J. Van, The dopamine response in mouse neuroblastoma cells is mediated by serotonin 5HT3 receptors, Eur. J. Pharmacol. 127 (1986) 271 – 274. [39] M. Nirenberg, Y. Liu, D. Peter, R. Edwards, V. Pickel, The vesicular monoamine transporter 2 is present in small synaptic vesicles and preferentially localizes to large dense core vesicles in rat solitary tract nuclei, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 8773 – 8777. [40] A. Nosjean, B. Franc, R. Laguzzi, Increased sympathetic nerve discharge without alteration in the sympathetic baroreflex response by serotonin3 receptor stimulation in the nucleus tractus solitarius of the rat, Neurosci. Lett. 186 (1995) 41 – 44. [41] Q. Pei, T. Zetterstrom, R. Leslie, D. Grahame-Smith, 5-HT3 receptor antagonists inhibit morphine-induced stimulation of mesolimbic dopamine release and function in the rat, Eur. J. Pharmacol. 230 (1993) 63 – 68. [42] J. Peters, J. Lambert, Electrophysiology of 5-HT3 receptors in neuronal cell lines, Trends Pharmacol. Sci. 10 (1989) 172 – 175. [43] A. Peters, S.I. Palay, H.F. Webster (Eds.), The Fine Structure of the Nervous System, Oxford University Press, New York, 1991, 494 pp. [44] V. Pickel, J. Chan, Ultrastructural localization of the serotonin transporter in limbic and motor compartments of the nucleus accumbens, J. Neurosci. 19 (1999) 7356 – 7366. [45] V. Pickel, T. Joh, J. Chan, A. Beaudet, Serotoninergic terminals: ultrastructure and synaptic interaction with catecholamine-containing neurons in the medial nuclei of the solitary tracts, J. Comp. Neurol. 225 (1984) 291 – 301. [46] H. Plattner, A. Artalejo, E. Neher, Ultrastructural organization of bovine chromaffin cell cortex—analysis by cryofixation and morphometry of aspects pertinent to exocytosis, J. Cell Biol. 139 (1997) 1709 – 1717. [47] G. Pratt, N. Bowery, The 5-HT3 receptor ligand, [3H]BRL 43694, binds to presynaptic sites in the nucleus tractus solitarius of the rat, Neuropharmacology 28 (1989) 1367 – 1376. [48] K. Preussat, C. Beetz, M. Schrey, R. Kraft, S. Wolfl, R. Kalff, S. Patt, Expression of voltage-gated potassium channels Kv1.3 and Kv1.5 in human gliomas, Neurosci. Lett. 346 (2003) 33 – 36. [49] C. Ransom, H. Sontheimer, D. Janigro, Astrocytic inwardly rectifying potassium currents are dependent on external sodium ions, J. Neurophysiol. 76 (1996) 626 – 630. [50] R. Robitaille, M. Bourque, S. Vandaele, Localization of L-type Ca2+ channels at perisynaptic glial cells of the frog neuromuscular junction, J. Neurosci. 16 (1996) 148 – 158. [51] P. Ronde, R.A. Nichols, 5-HT3 receptors induce rises in cytosolic and nuclear calcium in NG108-15 cells via calcium-induced calcium release, Cell Calcium 22 (1999) 357 – 365. [52] N. Schaffar, J.P. Kessler, O. Bosler, A. Jean, Central serotonergic projections to the nucleus tractus solitarii: evidence from a double labeling study in the rat, Neuroscience 26 (1988) 951 – 958. [53] A. Spier, G. Wotherspoon, S. Nayak, R. Nichols, J. Priestley, S. Lummis, Antibodies against the extracellular domain of the 5-HT3 receptor label both native and recombinant receptors, Brain Res. Mol. Brain Res. 67 (1999) 221 – 230.

J. Huang et al. / Brain Research 1028 (2004) 156–169 [54] R.M. Sykes, K.M. Spyer, P.N. Izzo, Central distribution of substance P, calcitonin gene-related peptide and 5-hydroxytryptamine in vagal sensory afferents in the rat dorsal medulla, Neuroscience 59 (1994) 195 – 210. [55] T. Tabata, A. Aiba, M. Kano, Extracellular calcium controls the dynamic range of neuronal metabotropic glutamate receptor responses, Mol. Cell. Neurosci. 20 (2002) 56 – 68. [56] J. Tabb, B. Molyneaux, D. Cohen, S. Kuznetsov, G. Langford, Transport of ER vesicles on actin filaments in neurons by myosin V, J. Cell. Sci. 111 (1998) 3221 – 3234. [57] W.T. Talman, M.H. Perrone, D.J. Reis, Evidence for l-glutamate as the neurotransmitter of baroreceptor afferent nerve fibers, Science 209 (1980) 813 – 815. [58] J. Tao-Cheng, F. Zhou, Differential polarization of serotonin transporters in axons versus soma-dendrites: an immunogold electron microscopy study, Neuroscience 94 (1999) 821 – 830. [59] L. Tecott, S. Shtrom, D. Julius, Expression of a serotonin-gated ion channel in embryonic neural and nonneural tissues, Mol. Cell. Neurosci. 6 (1995) 43 – 55. [60] K. Tsukamoto, T. Kurihara, N. Nakayama, O. Isogai, S. Ito, K. Komatsu, K. Kanmatsuse, Pressor responses to serotonin injected into the nucleus tractus solitarius of Sprague–Dawley rats and spontaneously hypertensive rats, Clin. Exp. Hypertens. 22 (2000) 63 – 73. [61] E. Van Bockstaele, J. Chan, V. Pickel, Pre- and postsynaptic sites for serotonin modulation of GABA-containing neurons in the shell region of the rat nucleus acumbens, J. Comp. Neurol. 371 (1996) 116 – 128.

169

[62] L. Volpicelli, J. Lah, A. Levey, Rab5-dependent trafficking of the m4 muscarinic acetylcholine receptor to the plasma membrane, early endosomes, and multivesicular bodies, J. Biol. Chem. 276 (2001) 47590 – 47598. [63] L. von Ruden, E. Neher, A Ca-dependent early step in the release of catecholamines from adrenal chromaffin cells, Science 262 (1993) 1061 – 1065. [64] Y. Wang, A. Ramage, D. Jordan, Mediation by 5-HT3 receptors of an excitatory effect of 5-HT on dorsal vagal preganglionic neurones in anaesthetized rats: an ionophoretic study, Br. J. Pharmacol. 118 (1996) 1697 – 1704. [65] Y. Wang, A. Ramage, D. Jordan, In vivo effects of 5-hydroxytryptamine receptor activation on rat nucleus tractus solitarius neurones excited by vagal C-fibre afferents, Neuropharmacology 36 (1997) 489 – 498. [66] Y. Wang, A. Ramage, D. Jordan, Presynaptic 5-HT3 receptors evoke an excitatory response in dorsal vagal preganglionic neurones in anaesthetized rats, J. Physiol. 509 (1998) 683 – 694. [67] N. Zahniser, S. Doolen, Chronic and acute regulation of Na+/Cl– dependent neurotransmitter transporters: drugs, substrates, presynaptic receptors, and signaling systems, Pharmacol. Ther. 92 (2001) 21 – 55. [68] X. Zhang, L. Bao, U. Arvidsson, R. Elde, T. Hokfelt, Localization and regulation of the delta-opioid receptor in dorsal root ganglia and spinal cord of the rat and monkey: evidence for association with the membrane of large dense-core vesicles, Neuroscience 82 (1998) 1225 – 1242.