Labeled Nerve Fibers Research Articles

Neuronal pathways and transmission to the lower esophageal sphincter of the guinea pig

Background & Aims: The lower esophageal sphincter (LES) normally controls the opening and closing of the gastroesophageal junction to resist gastric reflux but allow swallowing. Neuronal pathways controlling the guinea pig LES were investigated anatomically and physiologically in isolated preparations. Methods: Intracellular recording from the LES with focal electrical stimulation and retrograde and anterograde neuronal tracing were used. Results: Electrical stimulation on the LES evoked inhibitory junction potentials (IJPs), which were reduced by 60% by 100 μmol/L N-nitro- L-arginine and subsequently blocked by 0.5 μmol/L apamin, unmasking excitatory junction potentials, which were abolished by 1 μmol/L hyoscine. Esophageal or vagal stimulation evoked IJPs, which were blocked by 100 μmol/L hexamethonium. Focal stimulation of the upper stomach evoked IJPs at 5–8 of 20 stimulation sites, which were abolished by cutting between the stimulation site and sphincter. Application of 1,1'-didodecyl-3,3,3',3'-tetramethyl indocarbocyanine perchlorate (DiI) to the gastric sling muscle anterogradely labeled many motor axons in the sling muscle but few in the LES, confirming that the two muscles are separately innervated. DiI on the esophagus labeled nerve fibers, but not cell bodies, in the upper stomach. Conclusions: The inhibitory motor neurons of the LES receive inputs from the vagus nerve, esophagus, and upper stomach. GASTROENTEROLOGY 1998;115:661-671

Identification of motor neurons to the circular muscle of the guinea pig gastric corpus.

The projections of enteric neurons to the circular muscle of the guinea pig gastric corpus were investigated systematically by using the retrogradely transported fluorescent carbocyanine dye 1,1'-didodecyl-3,3,3',3'-tetramethyl indocarbocyanine perchlorate (DiI), applied to the muscle layer or myenteric plexus in vitro. DiI-labeled motor neuron cell bodies were located up to 6.3 mm aboral, 17 mm oral, and up to 20 mm circumferential to the DiI application site. Labeled nerve fibers ran for long distances from the DiI application site toward the greater and lesser curvatures, where they coursed parallel to the bundles of the "gastric sling" muscle. The majority of labeled cells were located toward the lesser curvature of the stomach. Nerve cell bodies that were aboral to the DiI application site were usually small, immunoreactive for choline acetyltransferase, and, thus, were likely to be excitatory motor neurons. Neurons that were located orally were larger, fewer in number, and immunoreactive for nitric oxide synthase and, thus, were likely to be inhibitory motor neurons. Application of DiI directly to the myenteric plexus filled neurons up to 15 mm aborally and up to 21 mm orally but labeled few neurons circumferentially. All nerve cells that were filled from either the circular muscle or the myenteric plexus had Dogiel type I morphological features. These results demonstrate a clear polarity of projection of inhibitory and excitatory motor neurons and a functionally continuous innervation of the circular and gastric sling muscle layers. Nonmotor neurons in the myenteric plexus were demonstrated, but neurons with Dogiel type II morphological features are apparently absent.

Implantation of stimulated homologous macrophages results in partial recovery of paraplegic rats.

Postinjury recovery in most tissues requires an effective dialog with macrophages; however, in the mammalian central nervous system, this dialog may be restricted (possibly due to its immune-privileged status), which probably contributes to its regeneration failure. We circumvented this by implanting macrophages, pre-exposed ex vivo to peripheral nerve segments, into transected rat spinal cord. This stimulated tissue repair and partial recovery of motor function, manifested behaviorally by movement of hind limbs, plantar placement of the paws and weight support, and electrophysiologically by cortically evoked hind-limb muscle response. We substantiated these findings immunohistochemically by demonstrating continuity of labeled nerve fibers across the transected site, and by tracing descending fibers distally to it by anterograde labeling. In recovered rats, retransection of the cord above the primary transection site led to loss of recovery, indicating the involvement of long descending spinal tracts. Injection of macrophages into the site of injury is relatively non-invasive and, as the cells are autologous, it may be developed into a clinical therapy.

Neuropeptide Y-immunoreactive intracardiac neurones, granule-containing cells and nerves associated with ganglia and blood vessels in rat and guinea-pig heart

Intrinsic neuropeptide Y-containing neurones in rat and guinea-pig hearts were studied at the ultrastructural level by the pre-embedding peroxidase-antiperoxidase immunocytochemical technique. Intracardiac neuronal cell bodies were often weakly or moderately immunostained, and the labelling was usually pronounced in the Golgi complex, multivesicular bodies, some cisterns of granular endoplasmic reticulum and large granular vesicles. Neuropeptide Y-immunoreactive nerve fibres were also observed in association with intracardiac neurones. A subpopulation of neuropeptide Y-immunoreactive granule-containing cells in the rat heart are described for the first time and were very heavily labelled; other granule-containing cells were non-immunoreactive, but were contacted by neuropeptide Y-containing nerves. Preterminal regions of nerve fibres that were located in nerve bundles were only weakly neuropeptide Y-immunoreactive, in contrast to the heavy labelling observed in varicosities that contained many synaptic vesicles. Many neuropeptide Y-immunoreactive nerve fibres were associated with the coronary vasculature and were particularly prominent in the walls of small arteries and arterioles where labelled nerve varicosities were present close to the smooth muscle cells. Immunoreactive nerves were also seen in the myocardium, usually near to capillaries. In axonal varicosities, the central core of large granular vesicles was immunolabelled, and electron-dense immunoreactive material outlined the membranes of small and large clear vesicles. The significance of neuropeptide Y-immunoreactive intracardiac neurones and granule-containing cells and the origin of associated labelled nerve fibres in the heart are discussed.

Peripheral course of genioglossal motor axons within the hypoglossal nerve of the rat.

The spatial compartmentalization of motor axons in cranial nerves has not been previously demonstrated. In the present study, motor axons in the medial lingual ramus of the hypoglossal nerve of rats were labeled using horseradish peroxidase and diaminobenzidine staining methods. Neuronal axons were segregated into fascicles within the main nerve trunk (average total length: 10.4 mm). Light microscopic examination of stained nerve sections revealed reaction product within individual axons and showed that the grouping of labeled nerve fibers in the medial ramus was maintained for at least the peripheral half of the nerve before the stain faded. The authors propose that the fascicular anatomy of the rat hypoglossal cranial nerve resembles that of peripheral spinal nerves in humans. This functional and structural compartmentalization may have a clinical impact on the repair and treatment of cranial nerve pathologies.

The cochlear place-frequency map of the adult and developing mongolian gerbil

Gerbils are frequently used in auditory research; however, only limited information on the exact frequency representation in the cochlea is available. Therefore the place-frequency map of the gerbil cochlea was determined by iontophoretic application of horseradish peroxidase in physiologically characterized single auditory nerve fibers. Locations of the labeled nerve fibers at the inner hair cells were determined from a ‘3-dimensional’ reconstruction of the cochleae. The map was established for frequencies between 0.3 and 31.9 kHz. As in other mammals, a baso-apical frequency gradient from high to low frequencies was found. The slope of the place-frequency map amounted to 1.5 mm/octave for higher frequencies. Below 4 kHz the slope decreased to a value of 1 mm/octave at 0.5 kHz. A shift in the frequency map occurs during cochlear maturation. The place-frequency map in subadult gerbils (18 days after birth) is shifted by a distance corresponding to an octave, at least for frequencies above 6 kHz.

Localization of nitrergic neuronal and non-neuronal cells in the ultimobranchial glands of the chicken.

The ultimobranchial glands of 20 chickens, aged 2-3 months, were investigated for their nicotinamide adenine dinucleotide hydrogen phosphate-diaphorase (NADPH-d) reactivity and the distribution of nitric oxide synthase (NOS), using NADPH-d histochemistry and NOS immunocytochemistry respectively. Formazan, the blue reaction product of NADPH-d, was localised in the neuronal cell bodies and nerve fibres. Most of the cell bodies were found in the parenchyma. Some of them occurred in the wall of the ultimobranchial cysts, and a few in the immediate vicinity of the blood vessels. Labelled nerve fibres mostly travelled with blood vessels, while few of them appeared in the cystic lining. In addition to neuronal profiles, some C cells, cystic lining, and vascular endothelium were also labelled. NOS staining was found in neuron-like cells and fibres that were confirmed as neurons in adjacent sections stained with antibodies against neuron-specific enolase. It was also detected in cystic lining and in some C cells, but not in vascular endothelium. The distribution patterns of NADPH-d and NOS suggest that NO may play a role in the regulation of the secretory activity of and the blood flow through the ultimobranchial glands.

Role of nitric oxide in human esophageal circular smooth muscle in vitro

The role of nitric oxide in human esophageal smooth muscle was examined. Immunostaining for constitutive nitric oxide synthase labeled nerve fibers and bundles within longitudinal and circular smooth muscle layers of resected tissue samples. Strips of circular muscle mounted in organ baths exhibited spontaneous contractions and active tone. When exposed to 5-second trains of electric field stimulation at 20 Hz, most strips exhibited intrastimulus "on" and poststimulus "off" contractions. Exposure to a 0.1 mumol/L (or greater) concentration of atropine converted "on" contractions to "on" relaxations and reduced "off" contractions by 63%. Exposure to NG-nitro-L-arginine resulted in concentration-dependent enhancement of "on" contractions and abolition of "off" contractions. Excess L-arginine enhanced the reversal of these effects. Sodium nitroprusside inhibited both spontaneous and evoked contractions. These results suggest that nitric oxide synthesis is a mediator of neural inhibition of human esophageal circular smooth muscle and is necessary for the occurrence of "off" contractions.

Ultrastructural investigation of nitric oxide synthase-immunoreactive nerves associated with coronary blood vessels of rat and guinea-pig.

Ultrastructural investigation of nitric oxide synthase-immunoreactive nerves closely associated with blood vessels in rat and guinea-pig hearts revealed many labelled nerve fibres in the walls of the main branches of the coronary arteries, and in arterioles, capillaries and post-capillary venules. The number of nitric oxide synthase-containing nerve fibres associated with different vessels, even those of the same calibre, varied. Terminal regions of nitric oxide synthase-immunoreactive fibres were observed in the endocardium and myocardium. Nitric oxide synthase-labelled fibres displayed electron-dense immunoproduct in both varicose and intervaricose regions. Immunoreactive axonal varicosities contained both small and large synaptic vesicles. The characteristics of the nitric oxide synthase-immunoreactive nerve fibres observed in the heart and the possibility that these fibres represent the processes of intracardiac neurones and/or sensory neurones of extrinsic origin are discussed.

Are breast tumours innervated? Immunohistological investigations using antibodies against the neuronal marker protein gene product 9.5 (PGP 9.5) in benign and malignant breast lesions

The aim of the present study was to assess the innervation pattern of benign and malignant breast lesions using the neuronal marker protein gene product (PGP) 9.5. An unlabelled antibody technique (using streptavidin biotin complex formation) was used on paraffin wax sections of tissues fixed in neutral buffered formalin. In 2 4 cases of chronic mastopathy, PGP 9.5 immunoreactivity was seen in relation to blood vessels and the ductal system. No immunoreactivity for PGP 9.5 was seen in the affected tissues of 9 10 cases of fibroadenomata. In 9 16 breast cancers, PGP 9.5-labelled perivascular nerve fibres were detected in connective tissue stroma supporting carcinoma tissue, though not in the immediate vicinity of such tumour tissue. Labelled nerve fibres were detected in large bundles at the periphery of tumours, possibly unrelated to the latter. Our results indicate that the newly formed blood vessels within a tumour are not innervated, though major blood vessels which supply the tumour are innervated.

Distinct subsets of neuropeptide y-negative principal neurons receive basket-like innervation from enkephalinergic and gabaergic axons in the superior cervical gaglion of adult rats

The distributions of axons immunoreactive for [Leu]- or [Met]enkephalin and GABA were studied in the superior cervical ganglion of adult rats. The antigens were visualized separately and in combination with neuropeptide Y by the immunoperoxidase technique, using reaction end-products of different colors. Similarities and differences were found in the light-microscopic innervation patterns of enkephalin- and GABA-immunoreactive nerve fibers. Both fiber systems were heterogeneously distributed within the superior cervical ganglion, forming denser networks in its rostral part than elsewhere in the ganglion. The appearance of labeled nerve fibers differed in the two systems. Enkephalin-immunoreactive axons exhibited dotted profiles due to a strong immunoreaction in the axonal varicosities as compared with that in the intervaricose segments, whereas GABA-positive fibers were evenly labeled in both parts of the axons. The most marked difference between the innervation patterns from enkephalin- and GABA-immunoreactive axons was the presence of bundles of varicose axons in conjunction with the basket-like aggregation of enkephalin-immunoreactive nerve terminals. The possibility that enkephalins and GABA are co-localized in certain axons was excluded in double-labeling studies, silver intensification being used for the first antigen and the nickel-enhanced diaminobenzidine reaction for the second antigen. Different subsets of principal neurons were richly innervated in a basket-like manner by axons immunoreactive for enkephalins and GABA. Additionally, combined staining with antisera against either enkephalin and neuropeptide Y or GABA and neuropeptide Y revealed that both subsets of principal neurons richly innervated either by enkephalin-immunoreactive or by GABA-immunoreactive axons were devoid of neuropeptide Y immunoreactivity. Thus, the enkephalinergic and GABAergic axons have different subpopulations of neuropeptide Y-negative principal neurons as targets in the superior cervical ganglion. These results provide further evidence that sympathetic ganglion cells can be classified on the basis of their receiving input from different sources.

Development of the sympathetic innervation to the cerebral arterial system in neonatal rats as revealed by anterograde labeling with wheatgerm agglutinin-horseradish peroxidase.

In order to clarify the developmental pattern in the sympathetic nerve fibers innervating the cerebral arterial system during the postnatal period in rats, we labeled the postganglionic nerve fibers originating in the superior cervical ganglion (SCG) and directly observed their extension and plexus formation by means of anterograde labeling with wheat germ agglutinin-horseradish peroxidase (WGA-HRP). The WGA-HRP solution was injected into the right SCG 1-7 days after birth. The rats were killed 48 h after trace injection, and the cerebral arteries were reacted with tetramethylbenzidine, then observed as a whole mount preparation. The labeled nerve fibers appeared as a few relatively straight bundles with branching fibers running longitudinally to the long axis of the artery in the ipsilateral right side of the circle of Willis and proximal portion of their main branching arteries at 3 days after birth. The nerve fibers started to form a circular pattern of nerve plexus only on the wall of the circle of Willis as early as 1 week after birth. At the beginning of postnatal week 2, labeled nerve fibers extended the collateral projections into the collateral side of the circle of Willis, and these expanding projections could not be observed at postnatal week 3. We observed a route of the sympathetic nerve fibers advancing into the cerebral arterial system which has not been described in previous studies; bundle of labeled nerve fibers entered into the wall of the middle portion of the basilar artery in half of the animals, in any postnatal period.(ABSTRACT TRUNCATED AT 250 WORDS)

Spinal and facial innervation of the skin in the gadid fish Ciliata mustela (Teleostei).

The pattern of innervation of the skin of the rockling Ciliata mustela was investigated to sort out spinal from facial nerve innervation of cutaneous chemosensory and mechanosensory systems. This fish has a variety of appendages with different functional sensory specializations, i.e., the chin barbel, pelvic fin, anterior dorsal fin, and dorsal trunk skin. The carbocyanine dye, diI, was applied to nerve stumps in dissected aldehyde-fixed tissue. In the case of the chin barbel, the dye was applied to both the trigeminal and facial nerve components. In the other cases, the dye was applied either selectively to the spinal nerves, to the facial nerves, or to both components. In the chin barbel, diI labeled nerve fibers associated with taste buds (TBs) and solitary chemosensory cells (SCCs) as well as relatively blunt free nerve endings, which closely approach the epidermal surface. In the pelvic fin, anterior dorsal fin, and dorsal trunk skin, taste buds, solitary chemosensory cells, and their innervation were labeled only after diI was applied to the facial nerve stumps. Application of diI to spinal nerves labeled delicate, free nerve endings and nerve fibers associated with small cells deep in the epidermis with features characteristic of Merkel cells. Transmission electron microscopy supports these results; after denervation of the facial component of the anterior dorsal fin, synaptic contacts with Merkel cells remained intact, whereas the synapses with the SCCs vanished.

Parasympathetic Cerebrovascular Innervation

To elucidate parasympathetic cerebrovascular innervation of the rat sphenopalatine ganglion (SPG), we injected wheat germ agglutinin-conjugated horseradish peroxidase into the SPG of one side and traced anterogradely labeled nerve fibers. Three days after the injection, the animals were killed and tissues including the SPG, major cerebral arteries, and ethmoidal arteries were reacted by the tetramethylbenzidine method. A number of cells and nerve fibers labeled with wheat germ agglutinin-conjugated horseradish peroxidase were observed in the SPG. The nasal mucous membrane, the periorbital soft tissue, and the lacrimal gland of the injected side contained numerous labeled nerve fibers. In cerebral vessels, anterogradely labeled nerve fibers were observed around the internal ethmoidal, anterior cerebral, middle cerebral, internal carotid, and posterior cerebral arteries of both sides. A few labeled nerve fibers were seen on the wall of the basilar arteries of the distal portion, and the vertebral arteries contained no labeled nerve fiber. In animals of which the ethmoidal nerve and the external ethmoidal artery were cut together with the surrounding periorbital soft tissues just before entering the ethmoidal foramen, no labeled nerve fiber was identified on the wall of the major cerebral arteries. Although labeled gangliocytes were found in the trigeminal and superior cervical ganglia after the tracer injection to the ipsilateral SPG, the chronic maxillary neurotomy and superior cervical ganglionectomy did not alter the distribution of the labeled nerve fibers on the wall of the cerebral arteries.(ABSTRACT TRUNCATED AT 250 WORDS)

Transcellular labeling by DiI demonstrates the glossopharyngeal innervation of taste buds in the lingual epithelium of the axolotl.

Innervation of the axolotl lingual epithelium by the glossopharyngeal nerve was examined to reveal its sensory target cells. The carbocyanine dye diI was applied to the nerve stump in the tongue fixed with paraformaldehyde. After a diffusion period of several months, the tongues were examined with a conventional epifluorescence microscope and a confocal laser scanning microscope (LSM) in wholemounts or preparations sectioned with a vibratome. Beneath the epithelium the labeled nerve fibers spread horizontally to form a meshwork of fibers, from which fascicles of fibers extended upward perpendicularly to the epithelium to innervate taste buds. Numerous taste buds were labeled by possible transcellular diffusion of diI. At the base of the taste bud, the nerve fibers branched and formed a basal plexus of fine fibers, on which numerous varicosities were seen. One or at most several taste cells were labeled in a taste bud. In the basal part of taste buds, the cell without an apical process, the basal cell, was also labeled. In the epithelium, between the taste buds, a few solitary cells were labeled. In some cases, a single fascicle of fibers innervating these cells was clearly shown by the LSM. In addition, fine fibers apparently formed free nerve endings in the epithelial cell layer. The results showed that the IX nerve innervated not only taste cells, but also presumed mechanosensory basal cells in the taste bud and the solitary cells of unknown function in the non-taste lingual epithelium. Afferent nerve responses to mechanical stimulation of the tongue may be explained by these non-taste cellular elements in the epithelium.

Neuropeptides in the human celiac/superior mesenteric ganglionic complex: An immunohistochemical study

The occurrence of vasoactive intestinal polypeptide (VIP), peptide histidine-isoleucine (PHI), calcitonin gene-related peptide (CGRP), substance P (SP), somatostatin (SOM), galanin (GAL) and enkephalins (ENK) is studied in the human celiac/superior mesenteric ganglionic complex of pre- and full-term new-borns, and adult subjects by means of immunohistochemistry. The antisera used labelled nerve fibres and terminal-like networks for each examined peptide, as well as VIP- and SOM-positive postganglionic neurons. Differences in the relative amount and density of the structures immunoreactive to the various peptides were observed. Moreover, variations in the amount and type of labelled elements were appreciable for each peptide when specimens from subjects at perinatal and adult ages were compared. Double-labelling immunofluorescence for SP and each other peptide showed that co-localization with SP is very frequent for CGRP, moderate to scarce for GAL and SOM, and rare to absent for PHI, VIP and ENK. VIP-, ENK- and CGRP-immunolabelled perikarya bearing the morphological features of the small intensely fluorescent (SIF) cells occurred in the organ. The presence of a paraganglion in one of the specimens examined allowed the detection of VIP- and ENK-positive cell bodies and VIP-, ENK-, SP- and GAL-like immunoreactive varicose nerve fibres in it. The results obtained provide substantial morphological data in support of the involvement of the examined peptides in the chemical interneuronal signalling in the human celiac/superior mesenteric ganglia.

Primary sensory afferents in the thymus of the guinea pig demonstrated with anterogradely transported horseradish peroxidase conjugates

After injections of wheat germ agglutinin- or B-subunit of cholera toxin-horseradish peroxidase (HRP) conjugates into the C 2–C 4 dorsal root ganglia of the guinea pig and anterograde axonal transport of the conjugates, labeled nerve fibers were found in the thymus mainly on the ipsilateral side. The distribution of the HRP-labeled fibers was similar to that of nerves immunoreactive to substance P and calcitonin gene-related peptide, which were partly colocalized. The results show that the thymus is innervated by primary sensory afferents which have their cell bodies in dorsal root ganglia.

The organization of the crossed geniculogeniculate pathway of the rat: A Phaseolus vulgaris-leucoagglutinin study

The intergeniculate leaflet of the thalamus is known to give rise to neuronal projections to the suprachiasmatic nuclei and the rostral part of the pineal gland. Via these projections the intergeniculate leaflet is considered to play a role in regulation of circadian rhythms. Iontophoretic injections of the anterograde tracer Phaseolus vulgaris-leucoagglutinin were placed in various subnuclei of the lateral geniculate nucleus in order to study the topographical organization of the crossed geniculogeniculate pathway in the rat. Injections involving neurons in the intergeniculate leaflet or the medial subpart of the ventral nucleus (which presumably is part of the intergeniculate leaflet of the thalamus too) gave rise to labeled nerve fibers in the opposite lateral geniculate nucleus. The axons contained in this pathway were followed either medially via the posterior commissure, or via the optic tracts and optic chiasm, to the contralateral hemisphere. In the contralateral lateral geniculate nucleus, the intergeniculate leaflet was most densely innervated, but a substantial innervation of the ventral lateral geniculate nucleus was observed as well. Only a few labeled fibers were observed in the dorsal subnucleus. However, the dense innervation of the contralateral intergeniculate leaflet not only covered the small zone between the dorsal and ventral nuclei, but also a dorsomedial part of the ventral nucleus that merged caudally with the lateral part of the zona incerta. In the remaining part of the ventral nucleus, single Phaseolus vulgaris-leucoag-glutinin-labeled fibers surrounded specific cells. The demonstration of a divergent projection between the intergeniculate leaflet and specific subparts of the contralateral geniculate nuclei indicates that the two lateral geniculate nuclei are regulating each other. The function of this pathway is suggested to be related to the regulation of circadian rhythmicity, but experimental evidence for this hypothesis is still lacking.

Functional organization of vagal reflex systems in the brain stem of the goldfish, Carassius auratus.

The coordination of secretory and motor responses to food within the alimentary canal requires well organized brain stem reflex systems. In the goldfish, Carassius auratus, three vagal reflex systems control three phases of ingestion and digestion. The orobranchial system sorts food from substrate, the pharyngeal chewing organ prepares items deemed to be food for digestion and absorption, and the abdominal system regulates the digestion of food. Each system is represented in the central nervous system by separate sensory and motor nuclei. The aim of the present study was to determine whether the interrelationships among the vagal sensory and motor nuclei reflect the peripheral organization. The sensory nucleus of each vagal system was injected with the neuronal tracer horseradish peroxidase (HRP), in separate cases. HRP injections into the vagal lobe sensory layers (orobranchial system) labeled fibers projecting topographically to the vagal lobe motor layer, but not at all to the pharyngeal or abdominal motor nuclei. Similarly, injections of HRP into the pharyngeal and abdominal sensory nuclei selectively labeled nerve fibers projecting to the pharyngeal and abdominal motor nuclei, respectively. All injections resulted in labeled fibers and/or cells in the lateral reticular formation, and in fibers ascending in the secondary gustatory-visceral tract. Gustatory information from the pharynx is apparently processed in the same brain stem system as pharyngeal general visceral information, suggesting that functional or regional characteristics of visceral sensory information may be more important for brain stem processing than the traditional "special" (gustatory) versus "general" visceral dichotomy. These results indicate that anatomically and functionally separate reflex systems exist within the goldfish vagal visceral nuclei.

Ultrastructural identification of trigeminal nerve terminals in the pterygopalatine ganglion of rats: an anterograde tracing and immunohistochemical study

Trigeminal nerve terminals in the rat pterygopalatine ganglion (PPG) were ultrastructurally identified using anterograde tracing with Phaseolus vulgaris-leucoagglutinin (PHA-L). Electron microscopic immunohistochemistry was used to demonstrate the presence of substance P (SP) and calcitonin gene-related peptide (CGRP) in nerve terminals of the PPG. Adjacent to the rostral part of the PPG an additional minor area was described. Perikarya in this minor rostral part were more spherical and had irregular outlines. Ultrastructurally, the glial enwrapment of the nerve terminals seemed to be more loosely arranged in comparison to that in the major rostral part of the PPG. With PHA-L, numerous labelled nerve fibres and terminals were found in all parts of the PPG. The ultrastructure of these terminals was uniform, many of them showing synaptic contacts. Numerous terminals in the PPG were SP-positive, whereas only a few were CGRP-positive. Fibres stained positive for both neuropeptides. The PPG is shown to be synaptically innervated by sensory fibres arising in the trigeminal ganglion, with the strong suggestion of SP and CGRP acting as neurotransmitters. A modulatory interaction between the autonomic and sensory system, resembling an axon reflex mechanism in the peripheral nervous system is endorsed.