Region-specific projections from the subfornical organ to the paraventricular hypothalamic nucleus in the rat.

Neurons in the subfornical organ (SFO) project to the paraventricular hypothalamic nucleus (PVN) and there, in response to osmolar and blood pressure changes, regulate vasopressin neurons in the magnocellular part (mPVN) or neurons in the parvocellular part (pPVN) projecting to the cardiovascular center. The SFO is functionally classified in two parts, the dorsolateral peripheral (pSFO) and ventromedial core parts (cSFO). We investigated the possibility that neurons in each part of the SFO project region-specifically to each part of the PVN, using anterograde and retrograde tracing methods. Following injection of an anterograde tracer, biotinylated dextran amine (BDX) in the SFO, the respective numbers of BDX-uptake neurons in the pSFO and cSFO were counted and the ratio of the former to the latter was obtained. In addition, the respective areas occupied by BDX-labeled axons per unit area of the mPVN and pPVN were measured and the ratio of the former to the latter was obtained. Similarly, following injection of the retrograde tracer in the PVN, the respective areas occupied by tracer per unit area of the mPVN and pPVN were measured and the ratio of the former to the latter was obtained. The respective numbers of retrogradely labeled neurons in the pSFO and cSFO were also counted and the ratio of the former to the latter was obtained. It became clear by statistical analyses that there are strong positive correlations between the ratio of BDX-uptake neuron number in the SFO and the ratio of BDX-axon area in the PVN in anterograde experiment (correlation coefficient: 0.787) and between the ratio of retrograde neuron number in the SFO and the ratio of tracer area in the PVN in retrograde experiment (correlation coefficient: 0.929). The result suggests that the SFO projects region-specifically to the PVN, the pSFO to the mPVN and the cSFO to the pPVN.

Activation of muscarinic receptors in rat subfornical organ neurones.

Cholinergic muscarinic inputs to subfornical organ (SFO) neurones in rats were studied using histochemical, molecular-biological and electrophysiological techniques. Neurones in the medial septum and the diagonal band (MS-DBB) were retrogradely labelled by a tracer wheat germ agglutinin-conjugated horseradish peroxidase-colloidal gold complex injected into the SFO. Some in the MS-DBB were double-labelled by choline acetyltransferase (ChAT) antibody. Many ChAT-immunoreactive fibres were observed in the SFO. M3 muscarinic receptor subtype-like immunoreactivity, detected using a polyclonal antiserum, was observed in the SFO. In slice preparations, muscarine induced inward currents in a dose-related manner. The inward currents were suppressed by the relatively M3 muscarinic receptor selective antagonist 4-diphenylacetoxy-N-methylpiredine methiodide. In the whole-cell current mode, muscarine depolarized the membrane with increased frequency of action potentials. Reverse transcriptase-polymerase chain reaction showed the presence of M2-M5 receptor mRNA in the SFO tissues. These results suggest that the SFO receives cholinergic muscarinic synaptic inputs from the MS-DBB. Acetylcholine postsynaptically activates and depolarizes neurones in the SFO partly through specific muscarinic receptors, including M3 receptor subtypes.

Further evidence for suicide inhibition of semicarbazide-sensitive amine oxidase in guinea pig lung by 2-bromoethylamine.

The inhibitory effects of 2-bromoethylamine (2-BEA), a derivative of ethylamine, on guinea pig lung semicarbazide-sensitive amine oxidase (SAO) have been studied. Preincubation with 2-BEA time-dependently inhibited SSAO activity. The mode of the initial phase of inhibition was competitive, with a Ki value of 52 microM. After preincubation at 37 degrees C for 2 h, the inhibition was noncompetitive and irreversible, as there was no recovery of SSAO activity by dilution of the inhibited samples. Kinetic analyses confirmed previous results with rat lung SSAO that 2-BEA is a suicide SSAO inactivator with a dissociation constant of 42 microM. This latter value is similar to that of the Ki value (52 microM) for the reversible phase of inhibition by 2-BEA. Addition of the nucleophilic compound 2-mercaptoethanol could not reduce the SSAO inhibition, indicating that inactivation could not be prevented by trapping the enzymatic reaction product from 2-BEA. This finding clearly indicates that the reaction product should not diffuse away from its site of genesis and agrees with one of the characteristics of suicide inhibitors. This conclusively excludes the possibility of an affinity-labeling mechanism.

Organ-specific autoimmunity in mice whose T cell repertoire is shaped by a single antigenic peptide.

Organ-specific autoimmune diseases have been postulated to be the result of T cell response against organ-specific self-peptides bound to MHC molecules. Contrary to this paradigm, we report here that transgenic mice lacking MHC class I expression and expressing an MHC class II I-A(b) molecule that presents only a single peptide (E alpha 52-68) spontaneously develops peripheral nervous system-specific autoimmune disease with many of the histopathological features found in experimental allergic neuritis. Reciprocal bone marrow chimeras produced using susceptible and resistant lines revealed that bone marrow-derived cells determined disease susceptibility. While the expression of the I-A(b)-E alpha 52-68 complex in the periphery was readily detectable in both lines, its expression on thymic dendritic cells responsible for tolerance induction was markedly lower in the susceptible line than in the resistant line. Consistent with this, CD4(+) T cells that can be activated by the I-A(b)-E alpha 52-68 complex were found in the susceptible line, but not in the resistant line. Such CD4(+) T cells conferred the disease to the resistant line by adoptive transfer, and administration of Ab specific for the I-A(b)-E alpha 52-68 complex inhibited disease manifestation in the susceptible line. These results indicate that disease development involves systemic T cell reactivity to I-A(b)-E alpha 52-68 complex, probably caused by incomplete negative thymocyte selection.

Target specific organization and neuron types of the dog pelvic ganglia: a retrograde-tracing and immunohistochemical study.

The major pelvic ganglion in both the rat and guinea pig has been extensively studied because of its anatomical simplicity. To clarify the target specific neural pathway in the diffusely distributed pelvic ganglia of larger animals, the pelvic plexus of the female dog was investigated by retrograde tracing and immunohistochemistry. The whole mount staining of the pelvic plexus with acetylcholinesterase histochemistry revealed 70-100 ganglia of varying sizes. Neurons retrogradely labeled from the rectum were mainly found in ganglia located in the dorso-caudal part of the plexus. The majority of these were non-catecholaminergic, immunoreactive for either calbindin (Calb) or neuropeptide Y (NPY), and characteristically associated with baskets of enkephalin (ENK)-immunoreactive varicose fibers. Neurons projecting to the utero-vaginal walls were distributed in ganglia located in the ventro-caudal part of the plexus. These mainly consisted of two major neuron groups: catecholaminergic Calb-immunoreactive neurons, and non-catecholaminergic neurons containing nitric oxide synthase (NOS) and/or vasoactive intestinal peptide (VIP), which were preferentially associated with a network of ENK-immunoreactive varicose fibers. Neurons retrogradely labeled from the urinary bladder mainly occurred in ganglia located around the junction between the ureter and the bladder. These consisted of catecholaminergic Calb neurons and noncatecholaminergic neurons containing Calb or NOS. Only a few ENK-immunoreactive fibers were found within the clusters of catecholaminergic neurons. These results indicate that organ specific neurons are located in separate ganglia and have both a distinctive composition of neuron types as well as different innervation by preganglionic fibers.

Substrate affects the initial attachment and subsequent behavior of human osteoblastic cells (Saos-2).

Interaction between implant materials and bone cells contributes to the clinical success of dental implants. The object of this study was to investigate the initial attachment and subsequent behavior of human osteoblastic cells (Saos-2) to pure titanium (Ti), hydroxyapatite (HA), and glass. We, therefore, performed a time-course study for examining the area, attachment rate, distribution of focal adhesion kinase (FAK) vinculin, and actin, and the motility of Saos-2 cells on the materials. On Ti, cell area increased gradually, whereas on HA, cells spread quickly, but quitted spreading at 12 h after cell seeding. The number of cells on HA was greater than on the other materials. On Ti, the numbers of FAK- and vinculin-positive focal adhesions increased continuously. On HA, although the number of FAK-positive focal adhesions also increased continuously, the number of vinculin-positive focal adhesions decreased. Furthermore, actin staining showed that the cells on HA poorly formed stress fibers with weak polarity, whereas the cell on Ti possessed well-defined polarized stress fibers. On HA, cells started extension earlier than on Ti, motility was inactive, and the cells settled on the materials. These results suggest that the earlier settling of osteoblasts on HA might result in earlier osteogenesis on HA than other materials.

Effects of water deprivation on neurokinin B production by the arginine-vasopressin neurons of hypothalamic paraventricular and supraoptic nuclei.

In order to clarify the role of neurokinin B (NKB), the dynamic changes in NKB expression and synthesis following water deprivation were examined in the arginine-vasopressin (AVP) neurons of hypothalamic paraventricular (PVN) and supraoptic (SON) nuclei. In intact rats, NKB and AVP showed almost the same high level of immunohistochemical reactivity in the magnocellular neurons of the PVN and SON, as well as in the varicose fibers in the median eminence (ME). In contrast, NKB precursor peptide (NKBp) immunoreactivity in the SON and PVN were relatively weak. Five days after water deprivation, AVP and NKB immunoreactivity decreased drastically, while NKBp-immunoreactivity increased in both the PVN and SON magnocellular neurons. Reverse transcription-polymerase chain reaction analysis of control animals revealed high levels of AVP mRNA and substantial amounts of NKB mRNA in the SON. This was contrast to the relatively low levels of AVP mRNA and undetectable levels of NKB mRNA in the PVN. After five days of water deprivation, AVP mRNA in the PVN and NKB mRNA in both the PVN and the SON increased considerably. These results indicate that synthesis and release of NKB, which colocalizes to AVP neurons, are enhanced by water deprivation in the same manner as AVP in the PVN and SON. Thus, NKB seems to be involved in the central control of body fluid levels. The results also suggest that the production rate of NKB under normal conditions in SON dominant.

Tyrosine hydroxylase-immunoreactive projections from the caudal ventrolateral medulla to the subfornical organ in the rat.

The subfornical organ (SFO) is known to be innervated by noradrenergic fibers. One possible origin of these fibers, which carry peripheral baroreceptor information to enhance the activity of SFO neurons, is the nucleus tractus solitarius (NTS). To investigate possible sites of origin of the catecholaminergic projections to the SFO, a retrograde tracing method was combined with immunohistochemistry in the rat. Stereotaxical injection of a retrograde tracer, wheat germ agglutinin-conjugated horseradish peroxidase--colloidal gold complex, into the SFO from the dorsal aspect revealed retrogradely labeled neurons in several catecholaminergic cell groups. A substantial number of retrogradely labeled neurons showing tyrosine hydroxylase (TH) immunoreactivity were found in the NTS and ventrolateral medulla (VLM) at levels caudal to the obex and in the locus coeruleus, while retrogradely labeled neurons without TH immunoreactivity were found in the VLM at levels rostral to the obex and in the nucleus prepositus hypoglossi. When the tracer was injected into the structures dorsal to the SFO, including the triangular septal nucleus, the frequency of retrogradely labeled neurons in the NTS and VLM at the caudal level was very low. These findings indicate the existence of catecholaminergic projections from the VLM (probably A1) to the SFO, in addition to the noradrenergic projections from the NTS previously reported.

Arginine-vasopressin neurons in the rat hypothalamus produce neurokinin B and co-express the tachykinin NK-3 receptor and angiotensin II type 1 receptor.

Secretion of arginine-vasopressin (AVP) from the hypothalamic paraventricular (PVN) and supraoptic (SON) nuclei is induced by neurokinin B (NKB) and angiotensin. To characterize the mechanisms by which this occurs, we used immunohistochemical techniques to assess the ability of AVP-producing neurons to express NKB, NKB receptor (NK-3 receptor) and angiotensin II type 1 receptor (AT-1 receptor). Double fluorescence immunohistochemistry indicated that AVP-immunoreactive cell bodies in the PVN and SON, as well as their axon varicosities in the posterior pituitary, co-express NKB. Almost all AVP-neuron perikarya also expressed both the NK-3 receptor and AT-1 receptor. Thus, AVP-producing neurons in the PVN and SON, which are regulated by NKB, are themselves a source of NKB. Furthermore, the regulation of AVP release by these neurons by NKB and angiotensin II is mediated by the NK-3 receptor and the AT-1 receptor, respectively.

Beta-endorphin-, adrenocorticotrophic hormone- and neuropeptide y-containing projection fibers from the arcuate hypothalamic nucleus make synaptic contacts on to nucleus preopticus medianus neurons projecting to the paraventricular hypothalamic nucleus in the rat.

The nucleus preopticus medianus is known to be situated in a key site in pathways regulating the paraventricular hypothalamic nucleus. To investigate the innervation pattern to nucleus preopticus medianus neurons by afferent fibers containing beta-endorphin, adrenocorticotrophic hormone and neuropeptide Y, a retrograde tracing method was combined with immunohistochemistry for these peptides in the rat. In the first experiment with injection of a retrograde tracer in the nucleus preopticus medianus, retrogradely labeled neurons were found in many regions throughout the brain. Among these, the arcuate hypothalamic nucleus contained a number of retrogradely labeled neurons showing immunoreactivity to the neuropeptides examined. About 20%, 20% and 40% of retrogradely labeled arcuate hypothalamic nucleus neurons showed beta-endorphin, adrenocorticotrophic hormone and neuropeptide Y immunoreactivity, respectively. About 18% and 57% of retrogradely labeled neurons in the nucleus tractus solitarius and ventrolateral medulla, respectively, were immunoreactive to neuropeptide Y. There were many more neuropeptide Y-immunoreactive projections to the nucleus preopticus medianus from the arcuate hypothalamic nucleus than those from the medulla. None of the retrogradely labeled neurons in the medulla showed immunoreactivity to beta-endorphin or adrenocorticotrophic hormone. In the second experiment with injection of a retrograde tracer in the paraventricular hypothalamic nucleus, electron microscopic observation revealed that retrogradely labeled neurons in the nucleus preopticus medianus were in synaptic contact with beta-endorphin-, adrenocorticotrophic hormone- and neuropeptide Y-immunoreactive axon terminals. The present finding indicates that nucleus preopticus medianus neurons projecting to the paraventricular hypothalamic nucleus are innervated by beta-endorphin-, adrenocorticotrophic hormone- and neuropeptide Y-containing arcuate hypothalamic nucleus neurons in addition to being innervated by neuropeptide Y-containing catecholaminergic medullary neurons which have been reported in our previous study.

Pharyngeal branch of the vagus nerve carries intraepithelial afferent fibers in the cat pharynx: an elucidation of the origin and central and peripheral distribution of these components.

The presence of a sensory component in the pharyngeal branch of the vagus nerve (PhB), including its peripheral distribution and central projection, was studied by denervation and tracer experiments in the cat. The distribution of nerve fibers immunoreactive to protein gene product 9.5, a sensitive neuronal marker; calcitonin gene-related peptide; and substance P in the pharyngeal epithelium was analyzed in both intact animals and animals subjected to partial denervation by means of sectioning two of the three nerve trunks, the glossopharyngeal nerve, the superior laryngeal nerve, and the PhB, while leaving one intact. The results of this study show that the glossopharyngeal nerve and superior laryngeal nerve carry nerve fibers to the pharyngeal epithelium rostral and caudal to the middle level of the epiglottis, respectively, whereas the PhB carries nerve fibers to the mesopharyngeal epithelium. Tracer experiments, by applying wheat germ agglutinin-horseradish peroxidase to the PhB, demonstrated retrogradely labeled primary sensory neurons in the jugular ganglion and transganglionic labeling of terminals in the interstitial subnucleus of the nucleus of the tractus solitarius. These results indicate that the PhB contains a sensory component that originates from the jugular ganglion, innervates the mesopharyngeal epithelium, and projects to the nucleus of the tractus solitarius.

Catecholaminergic and neuropeptide Y-immunoreactive synaptic inputs onto median preoptic nucleus neurons projecting to the ventrolateral medullary catecholaminergic area in the rat.

Median preoptic nucleus (POMe) neurons are innervated by catecholaminergic and neuropeptide Y (NPY)-immunoreactive nerve terminals originating from the catecholamine area of the ventrolateral medulla (VLM). The possibility that such POMe neurons project to the VLM catecholamine area was investigated in the rat. Immunoelectron microscopy revealed synaptic contacts of tyrosine hydroxylase- and NPY-immunoreactive axon terminals onto POMe neurons retrogradely labeled from the VLM catecholamine area, suggesting the existence of bidirectional connections between these two regions.

Synaptic contacts between nerve terminals originating from the ventrolateral medullary catecholaminergic area and median preoptic neurons projecting to the paraventricular hypothalamic nucleus.

A significant role of catecholaminergic projection to the median preoptic nucleus (POMe) which activates vasopressin-producing cells of the paraventricular hypothalamic nucleus (PVN) has been suggested. We investigated the existence of the synaptic contacts between catecholaminergic fibers from the ventrolateral medulla and the POMe neurons projecting to the PVN. Rats received a retrograde tracer in the PVN and subsequently an anterograde tracer into the catecholaminergic area of the ventrolateral medulla at the level of the area postrema. In the POMe, anterogradely labeled nerve terminals were found to make axo-somatic and axo-dendritic synaptic contacts onto retrogradely labeled neurons. Additional studies in which a retrograde tracer was injected into the POMe revealed that almost all retrogradely labeled neurons in the ventrolateral medulla at the level of the area postrema were immunoreactive to tyrosine hydroxylase, suggesting that projection to the POMe from the ventrolateral medulla is largely limited to catecholamine neurons. These results provide, for the first time, direct evidence that catecholaminergic inputs from the ventrolateral medulla affect POMe neurons projecting to the PVN by way of direct synaptic contact.

Phenotypic characterization of septal neurons in culture: immunohistochemistry of GABA, calbindin D-28k and choline acetyltransferase, and histochemistry of acetylcholinesterase.

Phenotypes of septal neurons, dissociated from 19-day-old fetal rat brains and then cultured in a medium containing nerve growth factor for 4 weeks, were examined using gamma-aminobutyric acid (GABA), calbindin D-28k, parvalbumin and choline acetyltransferase immunohistochemistry, and acetylcholinesterase histochemistry. There were primarily four groups of neurons identified in this septal culture: the first group (12.7% of 212 neurons examined) displayed a cholinergic, but not GABAergic, phenotype and had an average diameter of 13.6 +/- 2.7 microm (mean +/- S.D.); the second group (31.6%) displayed both cholinergic and GABAergic phenotypes and had a diameter of 12.2 +/- 2.8 microm; the third group (31.0%) displayed only a GABAergic phenotype and had a diameter of 10.4 +/- 2.3 microm; and the fourth group (24.7%) displayed neither a GABAergic nor cholinergic phenotype and had a diameter of 10.4 +/- 2.1 microm. Neurons in the first two groups described were significantly larger than those in the second two groups; neurons in the third and fourth groups were the same size. Calbindin D-28k was expressed in some neurons of each group (31.3%, 18.8%, 9.6% and 15.7%, respectively). These results demonstrate that septal neurons have the ability to express a variety of phenotypes when grown in vitro. This culture will be a useful tool for studying mechanisms of phenotype expression in septal neurons.

Phenotypic characterization of septal neurons in culture: immunohistochemistry of GABA, calbindin D-28k and choline acetyltransferase, and histochemistry of acetylcholinesterase

Phenotypes of septal neurons, dissociated from 19-day-old fetal rat brains and then cultured in a medium containing nerve growth factor for 4 weeks, were examined using gamma-aminobutyric acid (GABA), calbindin D-28k, parvalbumin and choline acetyltransferase immunohistochemistry, and acetylcholinesterase histochemistry. There were primarily four groups of neurons identified in this septal culture: the first group (12.7% of 212 neurons examined) displayed a cholinergic, but not GABAergic, phenotype and had an average diameter of 13.6±2.7 μm (mean±S.D.); the second group (31.6%) displayed both cholinergic and GABAergic phenotypes and had a diameter of 12.2±2.8 μm; the third group (31.0%) displayed only a GABAergic phenotype and had a diameter of 10.4±2.3 μm; and the fourth group (24.7%) displayed neither a GABAergic nor cholinergic phenotype and had a diameter of 10.4±2.1 μm. Neurons in the first two groups described were significantly larger than those in the second two groups; neurons in the third and fourth groups were the same size. Calbindin D-28k was expressed in some neurons of each group (31.3%, 18.8%, 9.6% and 15.7%, respectively). These results demonstrate that septal neurons have the ability to express a variety of phenotypes when grown in vitro. This culture will be a useful tool for studying mechanisms of phenotype expression in septal neurons.

Neuronal circuitry between the inferior mesenteric ganglion and lower intestine of the dog.

Neuronal circuitry between the inferior mesenteric ganglion (IMG) and the distal colon as well as the rectum, forming the intestino-intestinal reflex pathway, was investigated in the dog using immunohistochemistry combined with retrograde tract tracing and denervation experiments. Virtually all IMG neurons were tyrosine hydroxylase (TH)-immunoreactive. Of these ganglionic neurons, about 64% were also immunoreactive for calbindin (Calb), some 35% for neuropeptide Y (NPY), and 2% for vasoactive intestinal peptide (VIP). The retrograde tracer experiments revealed that both Calb/TH neurons and NPY/TH neurons projected to the distal colon and the rectum. In these intestinal walls, Calb/TH positive varicose fibers were found in the myenteric and submucous ganglia as well as in the longitudinal muscle layer, while NPY/TH positive fibers were mainly distributed around the vascular walls. Around Calb/TH neurons of the IMG, abundant varicose nerve fibers immunoreactive for VIP, dynorphin (DYN), calcitonin gene-related peptide (CGRP), enkephalin (ENK), substance P (SP) and bombesin (BOM) were distributed. These immunoreactive fibers disappeared after the total denervation of the IMG. After the application of Fast Blue into the IMG or distal stumps of transected lumbar colonic and hypogastric nerves, retrogradely labeled neurons occurred in the myenteric plexus with increasing density along the distal colon and rectum, and were immunoreactive for VIP, DYN, CGRP, ENK, SP or BOM. Double immunostaining of nerve fibers in the distal stumps of the ligated colonic and hypogastric nerves revealed the presence of viscerofugal fibers containing VIP with DYN and/or CGRP and those containing ENK with SP and/or BOM. These results demonstrate for the first time that the efferent limb of the canine intestino-intestinal reflex arch via the IMG consists of Calb-immunoreactive ganglion neurons projecting to the longitudinal muscles in addition to the enteric plexus of the lower intestine and also of NPY-immunoreactive ganglion neurons projecting to the intestinal blood vessels, and that the afferent limb is composed of at least two discrete groups with different peptide contents, i.e., myenteric neurons containing VIP with DYN and/or CGRP and those containing ENK with SP and/or BOM.

Synaptic contacts of substance P-immunoreactive axon terminals in the nucleus tractus solitarius onto neurons projecting to the caudal ventrolateral medulla oblongata in the rat.

The possibility that substance P (SP)-immunoreactive axon terminals in the nucleus tractus solitarius (NTS) make synaptic contacts onto NTS neurons projecting to the catecholaminergic cell region in the caudal ventrolateral medulla oblongata (CVLM) was examined in the rat using a retrograde tract-tracing method combined with immunohistochemistry. After injection of a retrograde tracer, wheat germ agglutinin-conjugated horseradish peroxidase-colloidal gold complex (WGA-HRP-gold), into the CVLM region where tyrosine hydroxylase-immunoreactive neurons were situated, many retrogradely labeled neurons were detected in the dorsal parts of the NTS, especially at levels between 1.0 mm caudal and 0.5 mm rostral to the obex. Immunoelectron microscopy revealed synaptic contacts between SP-immunoreactive axon terminals and WGA-HRP-gold-labeled neurons in the NTS. These findings indicated that SP regulates NTS neurons which project to the catecholaminergic cell region of the CVLM.

Distribution and origins of nitric oxide-producing nerve fibers in the dog tongue: correlated NADPH-diaphorase histochemistry and immunohistochemistry for calcitonin gene-related peptide using light and electron microscopy.

The distribution and origins of nitric oxide (NO)-producing nerves in the dog tongue with reference to calcitonin gene-related peptide (CGRP)-containing sensory fibers were investigated using NADPH-diaphorase (NADPH-d) histochemistry and immunohistochemistry for CGRP and NO synthase combined with retrograde axonal tracing and denervation experiments. The ultrastructural relationships between NADPH-d-positive and CGRP-immunoreactive neuronal elements were also examined electron microscopically. NADPH-d-positive and CGRP-immunoreactive varicose fibers were found within the taste buds and surrounding the epithelia of the fungiform papillae, and they disappeared completely after severance of the lingual nerve. Following injection of fast blue into the subepithelial layer of the anterior two thirds of the tongue, retrogradely labeled neurons possessing NO synthase and/or CGRP immunoreactivities were mainly detected in the trigeminal ganglion. Some of the retrogradely labeled trigeminal cells showed the coexistence of NADPH-d reactivity and CGRP immunoreactivity, but in the geniculate ganglion neither NADPH-d reactivity nor NO synthase immunoreactivity was found instead of retrogradely labeled CGRP-immunoreactive neurons. The lingual artery and its branches, including the arteriovenous anastomoses, showed dense distributions of NADPH-d-positive fibers, most of which were unaffected by the denervation experiments. There were many small ganglia in the tongue, and virtually all ganglionic neurons were NADPH-d reactive. CGRP-immuno-reactive varicose fibers were also found around the vascular walls and within the intralingual ganglia. Ultrastructural analysis revealed a close distribution of NADPH-d-positive and CGRP-immunoreactive varicose fibers within the arterial walls, and synaptic contacts between CGRP-immunoreactive terminals and NADPH-d-positive intralingual ganglionic neurons. These results indicated that the taste buds of epithelia of fungiform papillae in the anterior two thirds of the dog tongue receive NADPH-d-positive and CGRP-immunoreactive sensory fibers from the trigeminal ganglion, and that perivascular NADPH-d-positive fibers mainly originate from intrinsic ganglia in the tongue. The ultrastructural findings suggest an intrinsic peripheral nerve-reflex mechanism in the regulation of the lingual vascular function by NO-producing postganglionic parasympathetic neurons and CGRP-containing sensory fibers.