Phenylethanolamine N-methyltransferase-containing neurons in the rostral ventrolateral medulla of the rat. I. Normal ultrastructure

Depletion of C1 neurons attenuates the salt-induced hypertension in unanesthetized rats

High salt intake is able to evoke neuroendocrine and autonomic responses that include vasopressin release and sympathoexcitation resulting in increasing in the arterial blood pressure (BP). The C1 neurons are a specific population of catecholaminergic neurons located in the RVLM region and they control BP under homeostatic imbalance. Thus, here we hypothesized that the ablation of C1 neurons mitigate the high blood pressure induced by high-salt intake. To test this hypothesis, we injected anti-DβH-SAP saporin at the RVLM and monitored the BP in unanesthetized animals exposed to high salt intake of 2% NaCl solution for 7 days. The injection of anti-DβH-SAP into the RVLM depleted 80% of tyrosine hydroxylase-positive neurons (TH⁺ neurons) in the C1, 38% in the A5, and no significant reduction in the A1 region, when compared to control group (saline as vehicle). High salt intake elicited a significant increase in BP in the control group, while in the anti-DβH-SAP group the depletion of TH⁺ neurons prevents the salt-induced hypertension. Moreover, the low frequency component of systolic BP and pulse interval were increased by high-salt intake in control animals but not in anti-DβH-SAP group, which indirectly suggests that the increase in the BP is mediated by increase in sympathetic activity. In conclusion, our data show that hypertension induced by high-salt intake is dependent on C1 neurons.

Vesicular glutamate transporter 2 is required for the respiratory and parasympathetic activation produced by optogenetic stimulation of catecholaminergic neurons in the rostral ventrolateral medulla of mice in vivo

Catecholaminergic neurons of the rostral ventrolateral medulla (RVLM-CA neurons; C1 neurons) contribute to the sympathetic, parasympathetic and neuroendocrine responses elicited by physical stressors such as hypotension, hypoxia, hypoglycemia, and infection. Most RVLM-CA neurons express vesicular glutamate transporter (VGLUT)2, and may use glutamate as a ionotropic transmitter, but the importance of this mode of transmission in vivo is uncertain. To address this question, we genetically deleted VGLUT2 from dopamine-β-hydroxylase-expressing neurons in mice [DβH(Cre/0) ;VGLUT2(flox/flox) mice (cKO mice)]. We compared the in vivo effects of selectively stimulating RVLM-CA neurons in cKO vs. control mice (DβH(Cre/0) ), using channelrhodopsin-2 (ChR2-mCherry) optogenetics. ChR2-mCherry was expressed by similar numbers of rostral ventrolateral medulla (RVLM) neurons in each strain (~400 neurons), with identical selectivity for catecholaminergic neurons (90-99% colocalisation with tyrosine hydroxylase). RVLM-CA neurons had similar morphology and axonal projections in DβH(Cre/0) and cKO mice. Under urethane anesthesia, photostimulation produced a similar pattern of activation of presumptive ChR2-positive RVLM-CA neurons in DβH(Cre/0) and cKO mice. Photostimulation in conscious mice produced frequency-dependent respiratory activation in DβH(Cre/0) mice but no effect in cKO mice. Similarly, photostimulation under urethane anesthesia strongly activated efferent vagal nerve activity in DβH(Cre/0) mice only. Vagal responses were unaffected by α1 -adrenoreceptor blockade. In conclusion, two responses evoked by RVLM-CA neuron stimulation in vivo require the expression of VGLUT2 by these neurons, suggesting that the acute autonomic responses driven by RVLM-CA neurons are mediated by glutamate.

C1 neurons: the body's EMTs

The C1 neurons reside in the rostral and intermediate portions of the ventrolateral medulla (RVLM, IVLM). They use glutamate as a fast transmitter and synthesize catecholamines plus various neuropeptides. These neurons regulate the hypothalamic pituitary axis via direct projections to the paraventricular nucleus and regulate the autonomic nervous system via projections to sympathetic and parasympathetic preganglionic neurons. The presympathetic C1 cells, located in the RVLM, are probably organized in a roughly viscerotopic manner and most of them regulate the circulation. C1 cells are variously activated by hypoglycemia, infection or inflammation, hypoxia, nociception, and hypotension and contribute to most glucoprivic responses. C1 cells also stimulate breathing and activate brain stem noradrenergic neurons including the locus coeruleus. Based on the various effects attributed to the C1 cells, their axonal projections and what is currently known of their synaptic inputs, subsets of C1 cells appear to be differentially recruited by pain, hypoxia, infection/inflammation, hemorrhage, and hypoglycemia to produce a repertoire of stereotyped autonomic, metabolic, and neuroendocrine responses that help the organism survive physical injury and its associated cohort of acute infection, hypoxia, hypotension, and blood loss. C1 cells may also contribute to glucose and cardiovascular homeostasis in the absence of such physical stresses, and C1 cell hyperactivity may contribute to the increase in sympathetic nerve activity associated with diseases such as hypertension.

Rostroventrolateral medulla neurons with commissural projections provide input to sympathetic premotor neurons: anatomical and functional evidence

The activity of neurons in the rostral ventrolateral medulla (RVLM) is critical for the generation of vasomotor sympathetic tone. Multiple pre-sympathetic pathways converge on spinally projecting RVLM neurons, but the origin and circumstances in which such inputs are active are poorly understood. We have previously shown that input from the contralateral brainstem contributes to the baseline activity of this population: in the current study we investigate the distribution, phenotype and functional properties of RVLM neurons with commissural projections in the rat. We firstly used retrograde transport of fluorescent microspheres to identify neurons that project to the contralateral RVLM. Labelled neurons were prominent in a longitudinal column that extended over 1 mm caudal from the facial nucleus and contained hybridisation products indicating enkephalin (27%), GABA (15%) and adrenaline (3%) synthesis and included 6% of bulbospinal neurons identified by transport of cholera toxin B. Anterograde transport of fluorescent dextran-conjugate from the contralateral RVLM revealed extensive inputs throughout the RVLM that frequently terminated in close apposition with catecholaminergic and bulbospinal neurons. In urethane-anaesthetised rats we verified that 28/37 neurons antidromically activated by electrical stimulation of the contralateral pressor region were spontaneously active, of which 13 had activity locked to central respiratory drive and 15 displayed ongoing tonic discharge. In six tonically active neurons sympathoexcitatory roles were indicated by spike-triggered averages of splanchnic sympathetic nerve activity. We conclude that neurons in the RVLM project to the contralateral brainstem, form synapses with sympathetic premotor neurons, and have functional properties consistent with sympthoexcitatory function.

Ultrastructural localization of extranuclear progestin receptors relative to C1 neurons in the rostral ventrolateral medulla

To better understand the role of progestins in the C1 area of the rostral ventrolateral medulla (RVLM), immunocytochemical localization of progestin receptors (PRs) was combined with tyrosine hydroxylase (TH) in single sections of RVLM from proestrus rat brains prepared for light and electron microscopy. By light microscopy, PR-immunoreactivity (-ir) was detected in a few nuclei that were interspersed between TH-labeled perikarya and dendrites. Electron microscopy revealed that PR-ir was in several extranuclear locations. The majority of PR-labeling was in non-TH immunoreactive axons (51+/-9%) near the plasma membrane. Additional dual labeling studies revealed that PR-immunoreactive axons could give rise to terminals containing the GABAergic marker GAD65. PR-ir also was found in non-neuronal processes (29+/-9%), some resembling astrocytes. Occasionally, PR-ir was in non-TH-labeled terminals (10+/-3%) affiliated with clusters of small synaptic vesicles, or in patches contained in the cytoplasm of dendrites (10+/-1%). These findings suggest that progestins can primarily modulate neurons in the C1 area of the RVLM by presynaptic mechanisms involving GABAergic transmission. Moreover, they suggest that PR activation may contribute to progestin's effects on arterial blood pressure during pregnancy as well as to sex differences in central cardiovascular regulation.

Central chemoreceptors and sympathetic vasomotor outflow

The present study explores how elevations in brain P(CO(2)) increase the sympathetic nerve discharge (SND). SND, phrenic nerve discharge (PND) and putative sympathoexcitatory vasomotor neurons of the rostral ventrolateral medulla (RVLM) were recorded in anaesthetized sino-aortic denervated and vagotomized rats. Hypercapnia (end-expiratory CO(2) from 5% to 10%) increased SND (97 +/- 6%) and the activity of RVLM neurons (67 +/- 4%). Injection of kynurenic acid (Kyn, ionotropic glutamate receptor antagonist) into RVLM or the retrotrapezoid nucleus (RTN) eliminated or reduced PND, respectively, but did not change the effect of CO(2) on SND. Bilateral injection of Kyn or muscimol into the rostral ventral respiratory group (rVRG-pre-Bötzinger region, also called CVLM) eliminated PND while increasing the stimulatory effect of CO(2) on SND. Muscimol injection into commissural part of the solitary tract nucleus (commNTS) had no effect on PND or SND activation by CO(2). As expected, injection of Kyn into RVLM or muscimol into commNTS virtually blocked the effect of carotid body stimulation on SND in rats with intact carotid sinus nerves. In conclusion, CO(2) increases SND by activating RVLM sympathoexcitatory neurons. The relevant central chemoreceptors are probably located within or close to RVLM and not in the NTS or in the rVRG-pre-Bötzinger/CVLM region. RVLM sympathoexcitatory neurons may be intrinsically pH-sensitive and/or receive excitatory synaptic inputs from RTN chemoreceptors. Activation of the central respiratory network reduces the overall sympathetic response to CO(2), presumably by activating barosensitive CVLM neurons and inhibiting RTN chemoreceptors.

Modulatory inputs on sympathetic neurons in the rostral ventrolateral medulla in the rat

1. The first part of this study looks at spontaneously active neurons located in the rostral ventrolateral medulla (RVLM) with projections to the thoracic spinal cord. Sixteen neurons were intracellularly recorded in vivo. Four out of 16 neurons were antidromically activated from the thoracic spinal cord (axonal conduction velocities varied from 1.8 m/s to 9.5 m/s). 2. The simultaneous averages of the neuronal membrane potential and arterial blood pressure triggered by the pulsatile arterial wave or the EKG-R wave demonstrated changes in membrane potential (hyperpolarization or depolarization) locked to the cardiac cycle in four neurons in this group. These neurons (three of them bulbospinal) were further tested for barosensitivity by characterizing the responses to electrical stimulation of the aortic depressor nerve. Four neurons responded with inhibitory hyperpolarizing responses characterized as inhibitory postsynaptic potentials (IPSP) to aortic nerve stimulation (onset latency: 32.3 +/- 5.0 ms; mean +/- SEM). 3. In two neurons in the RVLM, one of them characterized as barosensitive, electrical stimulation of the opposite RVLM (0.5 Hz, 1.0 ms pulse duration, 25-100 microA) elicited excitatory postsynaptic potentials (EPSPs) with latencies of 9.07 and 10.5 ms. At resting membrane potential, the onset latency of the evoked EPSPs did not change with increasing stimulus intensities. Some of the recorded neurons were intracellularly labelled with biocytin for visualization. They were found in the RVLM. 4. These experiments in vivo would support the idea of a functional commissural pathway between the RVLM of both sides. 5. Anatomical data have shown that some of those commissural bundle fibers originate in the C1 adrenergic neuronal group in the RVLM. In the second part of this study, we used an intracellular recording technique in vitro to investigate the effects of the indirect adrenergic agonist tyramine on neurons in the RVLM with electrophysiological properties similar to premotor sympathetic neurons in vivo. 6. Tyramine (0.5-1 mM) produced a pronounced inhibitory effect with hyperpolarization and increase in membrane input resistance on two neurons characterized as regularly firing (R), and on one neuron characterized as irregularly firing (1). This effect was preceded by a transient depolarization with increases in firing rate. 7. These results would indicate that neurons in the RVLM recorded in vitro and with properties similar to premotor sympathetic neurons can be modulated by catecholamines released from terminals probably making synaptic contacts.

Development of catecholaminergic neurons in the human medulla oblongata

Distribution and maturation of catecholaminergic (CA) neurons have been studied by tyrosine hydroxylase immunohistochemistry in the medulla oblongata of human fetuses aged 14.5-25 weeks of gestation. Already at 14.5 weeks, CA neurons were observed in two longitudinally oriented cell clusters, one located ventrolaterally in the area of the lateral reticular and ambiguous nuclei, the other one dorsomedially forming 4 groups related to the dorsal vagal nucleus, the commissural nucleus of the vagus, the nucleus of the tractus solitarius and the area postrema. CA neurons in the area postrema were often found close to blood vessels. Scattered intermediate CA neurons were seen between these two larger clusters. CA neurons still appeared immature exhibiting bipolar morphology with only one or two short stout processes, which hardly branched. At 21 weeks, CA neurons occupied essentially the same location, but had a more mature morphology. Though still bipolar in shape, they had thinner and much longer processes which frequently branched. Both in the ventrolateral and the dorsomedial cell clusters, these processes were frequently lying close to blood vessels. At 25 weeks, CA cells had matured into multipolar neurons with long thin processes forming fine fiber networks in the ventrolateral medulla as well as around and within the dorsal vagal and solitarius nuclei. Only at this stage, a distinct CA fiber tract was seen located in the region of the tractus solitarius. Our results indicate that CA neurons in the human medulla, which are presumably involved in the control of ventilation and blood pressure, though generated rather early during development, mature relatively late.

C1 adrenergic neurons are contacted by presynaptic profiles containing DELTA-opioid receptor immunoreactivity

Ligands of the delta-opioid receptor tonically influence sympathetic outflow. Some of the actions of delta-opioid receptor agonists may be mediated through C1 adrenergic neurons in the rostral ventrolateral medulla. The goal of this study was to determine whether C1 adrenergic neurons or their afferents contain delta-opioid receptors. Single sections through the rostral ventrolateral medulla were labeled for delta-opioid receptor using the immunoperoxidase method and the epinephrine synthesizing enzyme phenylethanolamine N-methyltransferase (PNMT) using the immunogold method, and examined at the light and electron microscopic level. Few ( approximately 5% of 903) profiles dually labeled for PNMT and delta-opioid receptor were detected; most of these were dendrites with diameters < 1.5 microm. delta-Opioid receptor immunoreactivity was affiliated with multivesicular bodies in dually labeled perikarya, whereas delta-opioid receptor immunoperoxidase labeling appeared as isolated clusters within both singly and dually labeled dendrites. The majority ( approximately 83% of 338) of delta-opioid receptor-immunoreactive profiles were axons and axon terminals. delta-Opioid receptor-immunoreactive terminals averaged 0.75 microm in diameter, contained numerous large dense-core vesicles and usually formed appositions or asymmetric (excitatory-type) synapses with their targets. The majority (>50% of 250) of delta-opioid receptor-immunoreactive axons and axon terminals contacted PNMT-immunoreactive profiles. Most of the contacts formed by delta-opioid receptor-immunoreactive profiles ( approximately 75% of 132) were on single-labeled PNMT-immunoreactive dendrites with diameters <1.5 microm. The prominent localization of delta-opioid receptors to dense-core vesicle-rich presynaptic profiles suggests that delta-opioid receptor activation by endogenous or exogenous agonists may modulate neuropeptide release. Furthermore, the presence of delta-opioid receptors on axon terminals that form excitatory-type synapses with PNMT-immunoreactive dendrites suggests that delta-opioid receptor ligands may modulate afferent activity to C1 adrenergic neurons. The observation that some PNMT-immunoreactive neurons contain delta-opioid receptor immunoreactivity associated with multivesicular bodies and other intracellular organelles suggests that some C1 adrenergic neurons may present, endocytose and/or recycle delta-opioid receptors.

Cytoarchitecture of central chemoreceptors in the mammalian ventral medulla

We reviewed the previous reports on the fine anatomy of the mammalian ventral medulla with special attention to the cytoarchitecture of the superficial chemosensitive regions to summarize what is known, what is not yet known, and what should be studied in the future. We also reviewed studies on anatomical relationship between neurons and vessels, and morphological studies on dendrites of respiratory or chemosensitive neurons. When we compared the morphological reports on the ventral and dorsal putative chemosensitive regions, similarities were found as follows. Chemosensitive cells were often found not only near the ventral surface but near the dorsal surface of the brainstem. Dendritic projection towards the surface was a common characteristic in the ventral and dorsal chemosensitive neurons. Morphological abnormality in the brainstem of sudden infant death syndrome victims was also summarized. On the basis of the previous reports we discussed the perspective on the future study on central chemoreception. Among various unanswered questions in central chemosensitivity studies, physiological significance of surface cells and surface extending dendrites is the most important topic, and must be thoroughly investigated.

Selective distribution of mu-opioid receptors in C1 adrenergic neurons and their afferents

Agonists of the mu-opioid receptor (MOR) have profound effects on blood pressure, heart rate, and respiration that may be mediated by C1 adrenergic neurons in the rostral ventrolateral medulla (RVL). C1 neurons are sympathoexcitatory and are involved in both tonic and reflex regulation of sympathetic outflow. This study was designed to determine whether C1 neurons, or their afferents, contain MOR. C1 neurons were identified by using an antibody against the epinephrine synthesizing enzyme phenylethanolamine-N-methyl transferase (PNMT), whereas MOR was localized by using an antipeptide antibody that recognizes the cloned MOR, MOR1. Combined immunoperoxidase and immunogold methods were used to examine the cellular distribution of MOR1 relative to PNMT-containing neurons in the RVL. MOR1 was found in 22% of PNMT-containing dendrites (n = 392), whereas MOR1-containing axons or axon terminals contacted 14% of PNMT-containing dendrites. This distribution was heterogenous with regard to dendritic size: PNMT-labeled dendrites containing MOR1 were usually large (60% were >1.2 microm), whereas PNMT-containing dendrites that received MOR1-labeled afferents were usually small (79% were <1.2 microm). Individual dendrites rarely contained MOR1 at both pre- and postsynaptic sites. Together these results suggest that MOR agonists may directly influence the activity of C1 neurons, as well as the activity of select afferents to these cells. Plasmalemmal membrane labeling for MOR1 was more frequent in smaller PNMT-containing dendrites, suggesting that postsynaptic receptors are more readily available for ligand binding in small dendrites, although the receptor was more frequently detected in larger PNMT dendrites. The selective distribution of MORs to specific pre- and postsynaptic sites suggests the receptor may be selectively trafficked to positions where it may regulate afferent activity that is heterogeneously distributed along the dendritic tree of C1 neurons.

Mu-opioid receptors are present in functionally identified sympathoexcitatory neurons in the rat rostral ventrolateral medulla

Agonists of the mu-opioid receptor (MOR) produce profound hypotension and sympathoinhibition when microinjected into the rostral ventrolateral medulla (RVL). These effects are likely to be mediated by the inhibition of adrenergic and other presympathetic vasomotor neurons located in the RVL. The present ultrastructural studies were designed to determine whether these vasomotor neurons, or their afferents, contain MORs. RVL bulbospinal barosensitive neurons were recorded in anesthetized rats and filled individually with biotinamide by using a juxtacellular labeling method. Biotinamide was visualized by using a peroxidase method and MOR was identified by using immunogold localization of an antipeptide antibody that recognizes the cloned MOR, MOR1. The subcellular relationship of MOR1 to RVL neurons with fast- or slow-conducting spinal axons was examined by electron microscopy. Fast- and slow-conducting cells were not morphologically distinguishable. Immunogold-labeling for MOR1 was found in all RVL bulbospinal barosensitive neurons examined (9 of 9). MOR1 was present in 52% of the dendrites from both types of cells and in approximately half of these dendrites the MOR1 was at nonsynaptic plasmalemmal sites. A smaller portion of biotinamide-labeled dendrites (16%) from both types of cells were contacted by MOR1-containing axons or axon terminals. Together, these results suggest that MOR agonists can directly influence the activity of all types of RVL sympathoexcitatory neurons and that MOR agonists may also influence the activity of afferent inputs to these cells. The heterogenous distribution of MORs within individual RVL neurons indicates that the receptor is selectively targeted to specific pre- and postsynaptic sites.

Catecholamine systems in the brain of vertebrates: new perspectives through a comparative approach

A comparative analysis of catecholaminergic systems in the brain and spinal cord of vertebrates forces to reconsider several aspects of the organization of catecholamine systems. Evidence has been provided for the existence of extensive, putatively catecholaminergic cell groups in the spinal cord, the pretectum, the habenular region, and cortical and subcortical telencephalic areas. Moreover, putatively dopamine- and noradrenaline-accumulating cells have been demonstrated in the hypothalamic periventricular organ of almost every non-mammalian vertebrate studied. In contrast with the classical idea that the evolution of catecholamine systems is marked by an increase in complexity going from anamniotes to amniotes, it is now evident that the brains of anamniotes contain catecholaminergic cell groups, of which the counterparts in amniotes have lost the capacity to produce catecholamines. Moreover, a segmental approach in studying the organization of catecholaminergic systems is advocated. Such an approach has recently led to the conclusion that the chemoarchitecture and connections of the basal ganglia of anamniote and amniote tetrapods are largely comparable. This review has also brought together data about the distribution of receptors and catecholaminergic fibers as well as data about developmental aspects. From these data it has become clear that there is a good match between catecholaminergic fibers and receptors, but, at many places, volume transmission seems to play an important role. Finally, although the available data are still limited, striking differences are observed in the spatiotemporal sequence of appearance of catecholaminergic cell groups, in particular those in the retina and olfactory bulb.

Localization of mu-opioid receptors to locus coeruleus-projecting neurons in the rostral medulla: morphological substrates and synaptic organization

The increase in discharge activity of locus coeruleus (LC) neurons following precipitated opiate withdrawal has been reported to be caused, in part, by excitatory amino acid release most likely originating from the nucleus paragigantocellularis lateralis (PGCl) in the rostral ventral medulla. Activation of glutamate-containing neurons in the PGCl may depend on changes in the occupancy of opioid receptive sites located on LC-projecting neurons which subsequently effect excitatory amino acid release in the LC during opiate withdrawal. To determine whether the mu-opioid receptor (MOR) is localized to plasmalemmal sites of LC-projecting neurons in the PGCl, we combined retrograde transport of the protein-gold tracer, wheat germ agglutinin-conjugated to inactive horseradish peroxidase (WGA-AU-apoHRP), from the LC with immunocytochemical detection of MOR in the same section of tissue throughout the rostral medulla. Light microscopic analysis indicated that neurons containing either the retrograde tracer or immunoperoxidase labeling for the MOR were numerous throughout the ventral medulla and that individual PGCl neurons contained both WGA-Au-apoHRP as well as MOR. By electron microscopy, WGA-Au-apoHRP was commonly identified in lysosomes within somata and large proximal dendrites. The somata contained either spherical or invaginated nuclei and were often surrounded by numerous myelinated axons. Gold deposits could also be identified in the cytoplasm of smaller dendritic processes in the PGCl, although these were not necessarily associated with lysosomes. The smaller dendritic processes were often the target of afferent input by axon terminals containing heterogeneous types of synaptic vesicles. Of 150 cellular profiles exhibiting WGA-Au-apoHRP retrograde labeling, 31% contained immunoperoxidase labeling for MOR. These results indicate that the MOR is distributed along plasmalemmal sites of morphologically diverse neurons in the PGCl which project to the LC.

Lesions of the C1 catecholaminergic neurons of the ventrolateral medulla in rats using anti-DbetaH-saporin

Phenylethanolamine-N-methyltransferase (PNMT)-containing neurons in the rostral ventrolateral medulla (RVLM) are believed to play a role in cardiovascular regulation. To determine whether injection of anti-dopamine beta-hydroxylase (DbetaH)-saporin directly into the RVLM in rats could selectively destroy these cells and thereby provide an approach for evaluating their role in cardiovascular regulation, we studied rats 2 wk after unilateral injection of 21 ng anti-DbetaH-saporin into the RVLM. There was an approximately 90% reduction in the number of PNMT-positive neurons in the RVLM, although the number of non-C1, spinally projecting barosensitive neurons of this area was not altered. The A5 cell group was the only other population of DbetaH-containing cells that was significantly depleted. The depressor response evoked by injection of tyramine into the RVLM was abolished by prior injection of toxin. The pressor response evoked by injection of glutamate into the RVLM was attenuated ipsilateral to the toxin injection but was potentiated contralateral to the toxin injection. Thus anti-DbetaH-saporin can be used to make selective lesions of PNMT-containing cells, allowing for the evaluation of their role in cardiovascular regulation.

Alpha2A-adrenergic receptors are primarily presynaptic heteroreceptors in the C1 area of the rat rostral ventrolateral medulla

The 2A subtype of the alpha-adrenergic receptor (alpha2A-AR) is necessary for the hypotensive effects of clonidine and other sympathoinhibitory adrenergic agonists. This hypotensive response appears to be due to the inhibition of sympathoexcitatory reticulospinal neurons found in the rostral ventrolateral medulla (RVL), including neurons of the C1 adrenergic cell group. The cellular mechanisms underlying this inhibition have not been established. Thus, this study examined the ultrastructural relationships between profiles containing alpha2AAR-immunoreactivity (alpha2AAR-I) and those containing the catecholamine synthesizing enzyme tyrosine hydroxylase (TH) to determine potential cellular substrates for alpha2A-AR inhibition of C1 neuron activity. Consistent with previous light microscopic studies, alpha2AAR-I was found in perikarya and large dendrites and the majority of these profiles also contained TH-labeling (approximately 70% of 140). However, alpha2AAR-I in these cells was primarily found within endosomes and Golgi complexes and in clusters associated with the endoplasmic reticula, probable sites for synthesis and/or trafficking of receptors. In contrast, most of the alpha2AAR-I profiles (n=646) in the RVL were axons and axon terminals (approximately 68%) which lacked TH immunoreactivity. alpha2AAR-labeled axons were small and unmyelinated and labeled terminals usually formed symmetric synapses on the shafts of catecholaminergic or unlabeled dendrites. Most of these alpha2AAR-labeled axons were found in close proximity to TH-labeled profiles and approximately one-fifth (17% of 408) of the alpha2AAR-labeled axons and axon terminals directly contacted TH-labeled profiles, mostly dendrites. These studies suggest that alpha2AARs in the C1 area of the RVL function primarily as heteroreceptors on presynaptic axons and terminals of non-catecholaminergic cells, some of which provide inhibitory synaptic input to C1 neurons. These receptors may be activated by catecholamines released either from the dendrites of C1 neurons or from the terminals of other catecholaminergic neurons via volume transmission.

Evidence of disynaptic projections from the rostral ventrolateral medulla to the thoracic spinal cord

Sympathetic outflow is regulated by a direct pathway of the rostral ventrolateral reticular formation (rvlm) to the thoracic spinal cord. For the first time, a dual retrograde/anterograde transport technique was used to demonstrate by light microscopy, potential disynaptic pathways from the rvlm to the thoracic spinal cord in the rat. An anterograde tracer, biotinylated dextran amine (BDA) was injected into the rvlm and a retrograde tracer, FluoroGold (FG) deposited into the upper thoracic spinal cord in the same animal. Rostral ventrolateral medullary efferents labeled with BDA were apposed to thoracic reticulospinal neurons labeled with FG in the ventrolateral tegmentum, ipsilateral and contralateral to the injection site in the rvlm. Suggestive evidence was obtained of synaptic interactions with neuronal somata and proximal dendrites. The results support the idea that the rvlm projects to the thoracic cord via disynaptic, intrareticular pathways paralleling the well established monosynaptic projection.

Molecular cloning and characterization of an L-epinephrine transporter from sympathetic ganglia of the bullfrog, Rana catesbiana

Chemical signaling by dopamine (DA) and L-norepinephrine (L-NE) at synapses is terminated by uptake via specialized presynaptic transport proteins encoded by the DA transporter (DAT) and L-NE transporter (NET) genes, respectively. In some vertebrate neurons, particularly the sympathetic neurons of amphibians, L-NE is converted to L-epinephrine (L-Epi, adrenaline) and released as the primary neurotransmitter. Although evidence exists for a molecularly distinct L-Epi transporter (ET) in the vertebrate brain and peripheral nervous system, a transporter specialized for extracellular L-Epi clearance has yet to be identified. To pursue this issue, we cloned transporter cDNAs from bullfrog (Rana catesbiana) paravertebral sympathetic ganglia and characterized functional properties via heterologous expression in non-neuronal cells. A cDNA of 2514 bp (fET) was identified for which the cognate 3.1 kb mRNA is highly enriched in frog sympathetic ganglia. Sequence analysis of the fET cDNA reveals an open reading frame coding for a protein of 630 amino acids. Inferred fET protein sequence bears 75, 66, and 48% amino acid identity with human NET, DAT, and the 5-hydroxytryptamine transporter (SERT), respectively. Transfection of fET confers Na+- and Cl--dependent catecholamine uptake in HeLa cells. Uptake of [3H]-L-NE by fET is inhibited by catecholamines in a stereospecific manner. L-Epi and DA inhibit fET-mediated [3H]-L-NE uptake more potently than they inhibit [3H]-L-NE uptake by human NET (hNET), whereas L-NE exhibits equivalent potency between the two carriers. Moreover, fET exhibits a greater maximal velocity (Vmax) for the terminal products of catecholamine biosynthesis (L-Epi > L-NE >> DA), unlike hNET, in which a Vmax rank order of L-NE > DA > L-Epi is observed. fET-mediated transport of catecholamines is sensitive to cocaine and tricyclic antidepressants, with antagonist potencies significantly correlated with hNET inhibitor sensitivity. Amino acid conservation and divergence of fET with mammalian catecholamine transporters help define residues likely to be involved in catecholamine recognition and translocation as well as blockade by selective reuptake inhibitors.

Adrenergic responses in silent and putative inhibitory pacemaker-like neurons of the rat rostral ventrolateral medulla in vitro

Noradrenaline and adrenergic agonists were tested on pacemaker-like and silent neurons of the rat rostral ventrolateral medulla using intracellular recording in coronal brainstem slices as well as in punches containing only the rostral ventrolateral medullary region. Noradrenaline (1-100 microM) depolarized or increased the frequency of discharge of all cells tested in a dose-dependent manner. The noradrenaline-induced depolarization was associated with an apparent increase in cell input resistance at low concentrations and a decrease or no significant change at higher concentrations. Moreover, it was voltage dependent and its amplitude decreased with membrane potential hyperpolarization. Noradrenaline caused a dose-related increase in the frequency and amplitude of spontaneous inhibitory postsynaptic potentials. The alpha 1-adrenoceptor antagonist prazosin (0.5 microM) abolished the noradrenaline depolarizing response as well as-the noradrenaline-evoked increase in synaptic activity and unmasked an underlying noradrenaline dose-dependent hyperpolarizing response associated with a decrease in cell input resistance and sensitive to the alpha 2-adrenoceptor/antagonist yohimbine (0.5 microM). The alpha 1-adrenoceptor agonist phenylephrine (10 microM) mimicked the noradrenaline depolarizing response associated with an increase in membrane resistance as well as the noradrenaline-induced increase in synaptic activity. The alpha 2-adrenoceptor agonists UK-14,304 (1-3 microM) and clonidine (10-30 microM) produced only a small hyperpolarizing response, whereas the beta-adrenoceptor agonist isoproterenol (10-30 microM) had no effect. Baseline spontaneous postsynaptic potentials were abolished by strychnine (1 microM), bicuculline (30 microM) or both. However, only the strychnine-sensitive postsynaptic potentials had their frequency increased by noradrenaline or phenylephrine and they usually occurred with a regular pattern. Tetrodotoxin (1 microM) eliminated 80-95% of baseline spontaneous postsynaptic potentials and prevented the increase in synaptic activity evoked by noradrenaline and phenylephrine. Similar results were obtained in rostral ventrolateral medulla neurons impaled in both coronal slices and punches of the rostral ventrolateral medulla. It is concluded that noradrenaline could play an important inhibitory role in the rostral ventrolateral medulla via at least two mechanisms: an alpha 2-adrenoceptor-mediated hyperpolarization and an enhancement of inhibitory synaptic transmission through activation of alpha 1-adrenoceptors located on the somatic membrane of glycinergic interneurons. Some of these interneurons exhibit a regular discharge similar to the pacemaker-like neurons and might, at least in part, constitute a central inhibitory link in the baroreceptor-vasomotor reflex pathway.

Angiotensin II decreases a resting K+ conductance in rat bulbospinal neurons of the C1 area

In the rostral ventrolateral medulla (RVLM), angiotensin II (Ang II) receptors are concentrated in the region that contains neurons innervating sympathetic preganglionic neurons. We sought to determine whether these bulbospinal cells are sensitive to Ang II. Retrogradely labeled bulbospinal RVLM neurons (N = 125) were recorded in thin slices from neonatal rats. Most (33 of 46) histologically recovered bulbospinal neurons were C1 cells (immunoreactive for tyrosine hydroxylase [TH-ir] or phenylethanolamine N-methyltransferase [PNMT-ir]). Bulbospinal RVLM neurons were spontaneously active (2.7 +/- 0.2 spikes per second, n = 69) with 'resting' potential of -54 +/- 0.4 mV (n = 77) and input resistance of 879 +/- 53 M omega (n = 47). Ang II (0.3 to 1 mumol/L) increased the spontaneous firing rate of most bulbospinal neurons (+250%, 28 of 39). In current-clamp mode, Ang II (1 mumol/L) produced depolarization (+6.8 +/- 0.6 mV, n = 59 neurons) and increased input resistance (+21 +/- 2%, n = 36 neurons). In voltage-clamp mode, Ang II elicited an inward current (9.7 +/- 0.9 pA; holding potential, -40 to -55 mV; n = 25 neurons) that reversed polarity at the K+ equilibrium potential (n = 8 neurons) and was barium sensitive (n = 4 neurons). Ang II-evoked conductance change was voltage independent (-40 to -140 mV, n = 8 neurons). The effects of Ang II were blocked by losartan (9 of 9 neurons) but persisted in low Ca2+/high Mg2+ (7 of 7 neurons). Ang II-sensitive cells were inhibited by alpha 2-adrenergic receptor agonists (12 of 15 neurons). Ang II excited 91% (30 of 33) of TH-ir or PNMT-ir cells but 23% (3 of 13) of non-TH-ir neurons. In conclusion, RVLM bulbospinal cells express Ang II type-1 receptors whose activation leads to a reduction in resting K+ conductance.

Monosynaptic input from Leu5-enkephalin-immunoreactive terminals to vagal motor neurons in the nucleus ambiguus: comparison with the dorsal motor nucleus of the vagus

Vagal motor neurons in the rat dorsal motor nucleus of the vagus (DMN) are known to receive direct synaptic input from enkephalin-containing terminals. We examined 1) whether the vagal motor neurons within the nucleus ambiguus (NA) also received monosynaptic input from enkephalin-immunoreactive terminals and 2), if so, whether their ultrastructural relations differed from those in the DMN. In both regions, terminals containing Leu5-enkephalin-like immunoreactivity (LE-LI) were examined in relation to motor neurons identified by retrograde transport of wheat germ-agglutinated horseradish peroxidase (WGA-HRP) applied to the cut end of the cervical vagus nerve in single sections of the medulla oblongata of adult rats. By light microscopy, the most significant overlap between varicose processes with LE-LI and WGA-HRP-containing neurons was seen in the rostral compact portion of the NA and the DMN at the level of the obex. Thus, only these regions were examined by electron microscopy. The most distinguishing ultrastructural feature of WGA-HRP-labeled neurons in the NA compared to the DMN was their higher incidence of nonsynaptic appositions with other neurons. In both the NA and the DMN, terminals with LE-LI formed primarily symmetric synapses on smaller (presumably distal) dendrites; many of these dendrites, as well as most target perikarya, contained WGA-HRP. Additionally, in the compact portion of the NA compared to the DMN 1) multiple LE-labeled terminals more frequently contacted single perikarya or dendrites and 2) single terminals with LE-LI more commonly showed two contacts or active zones and contained more abundant LE-immunoreactive large (80-100 nm) dense-core vesicles (dcvs). In contrast to small (40-50 nm), clear vesicles, which were usually aggregated near active zones, the immunoreactive dcvs were usually located near glial processes distal to these zones. These results indicate that enkephalin immunoreactivity is intensely localized to dcvs within terminals that may have direct inhibitory (symmetric synapses) actions on vagal motor neurons in both the compact portion of the NA and the DMN. Moreover, because numbers of dcvs and active zones have been equated with synaptic strength, our findings suggest enhanced potencies of enkephalin-immunoreactive terminals in the compact portion of the NA. Our findings support a prominent role for enkephalin in the coordinated activity of esophageal motor neurons located in the compact portion of the NA.

Neurons in the barosensory area of the caudal ventrolateral medulla project monosynaptically on to sympathoexcitatory bulbospinal neurons in the rostral ventrolateral medulla

Neurons in the caudal ventrolateral medulla may function as interneurons in the baroreceptor reflex are by inhibiting sympathoexcitatory bulbospinal neurons in rostral ventrolateral medulla. While some caudal ventrolateral medullary neurons are excited orthodromically by baroreceptors and antidromically from the rostral ventrolateral medulla, there is no anatomical evidence to prove that these barosensory neurons of the caudal ventrolateral medulla monosynaptically innervate the bulbospinal neurons in the rostral ventrolateral medulla. To establish the presence of such a direct projection, barosensory neurons were identified in the rostral caudal ventrolateral medulla of anesthetized rats by criteria that they spontaneously discharged with a cardiac rhythm and were excited by baroreceptor stimulation. The anterograde tracer biocytin was iontophoresed onto these neurons and, in the same animal, the retrograde tracer wheatgerm-agglutinated apo-horseradish peroxidase conjugated to gold particles was injected by micropressure into the ipsilateral spinal (thoracic level 3) intermediolateral cell column to label bulbospinal neurons. After 18-24 h, rats were killed and sections through the rostral ventrolateral medulla were processed for both markers. By light microscopy, numerous biocytin-labeled varicose processes overlapped rostral ventrolateral medullary neurons containing wheatgerm-agglutinated apo-horseradish peroxidase conjugated to gold particles. By electron microscopy, biocytin was found in axons and terminals. The terminals (n = 76) were large (0.6-1.2 microns in diameter), contained numerous small, clear vesicles and formed primarily symmetric synapses on perikarya and large (1.5-4.5 microns) dendrites within the rostral ventrolateral medulla. Some of these target neurons contained wheatgerm-agglutinated apo-horseradish peroxidase conjugated to gold particles associated with lysosomes and multivesicular bodies in the cytoplasm. The results indicate that: (i) neurons in the barosensory sympathoinhibitory region of the caudal ventrolateral medulla directly synapse on bulbospinal neurons in the rostral ventrolateral medulla; and (ii) the synaptic profile (symmetric synapse) and location (perikarya and large dendrites) is consistent with the conclusion that baroreceptor neurons of the caudal ventrolateral medulla potently and monosynaptically inhibit sympathoexcitatory neurons of the rostral ventrolateral medulla. The findings support the hypothesis that the barosensory region of the rostral caudal ventrolateral medulla is an intermediate relay in the baroreceptor reflex are.

A1 catecholamine cell group: fine structure and synaptic input from the nucleus of the solitary tract

Preembedding immunoperoxidase staining methods were used to characterize tyrosine hydroxylase-immunoreactive (TH-ir) elements in the caudal ventrolateral medulla, and to determine the extent to which neurons of the A1 cell group are directly innervated by projections of the nucleus of the solitary tract (NTS). TH-ir neurons in the A1 region were medium-sized and multipolar. They possessed rounded nuclei with infrequent invaginations, well-developed Golgi apparati, high cytoplasmic densities of mitochondria, and a low to moderate tendency for rough endoplasmic reticulum (RER) to align in parallel stacks. A1 cell bodies were commonly juxtaposed to TH-positive and TH-negative neurons, myelinated profiles, glia and/or vascular elements, but close membrane appositions were only seen with glial elements. Synaptic input to A1 neurons was predominantly asymmetric, provided virtually exclusively by non-TH-ir terminals, and directed principally to dendritic shafts; A1 somata are relatively sparsely innervated. In a second experiment, silver-intensified immunogold localization of TH-ir was combined with immunoperoxidase labeling for anterogradely transported Phaseolus vulgaris-leucoagglutinin (PHA-L), following tracer injections in the caudal aspect of the medial division of the NTS. These experiments revealed a small proportion of PHA-L-labeled axon terminals that made asymmetric contacts with dendritic shafts of TH-ir neurons. These results suggest that the fine structure and synaptic input of A1 neurons are somewhat distinct from that of rostrally situated C1 catecholamine cells. In addition, while they document a direct NTS-A1 projection that may participate in the interoceptive control of vasopressin secretion, the bulk of ventrolaterally directed projections from the caudomedial NTS contact noncatecholaminergic elements in the A1 region, some of which may correspond to so-called depressor neurons implicated in the baroreflex control of sympathetic outflow and vasopressin secretion.

Distribution of the alpha-ketoglutarate dehydrogenase complex in rat brain

The alpha-ketoglutarate dehydrogenase complex (KGDHC) is a key enzyme in mitochondrial oxidation that appears critical to neurodegenerative diseases. Its activity in the brain declines in thiamine-deficient animals, Alzheimer's disease, and Wernicke-Korsakoff syndrome. Since selective cell populations are affected in these disorders, understanding the cellular distribution of KGDHC is important in order to define its role in the pathophysiology of these diseases. We used antisera against both bovine KGDHC and its E1k component to determine the immunocytochemical distribution of the enzyme and compare it with that of another mitochondrial enzyme, pyruvate dehydrogenase complex (PDHC) and a cholinergic neuronal marker, choline acetyltransferase (ChAT) in rat brain. Although low levels of immunoreactivity occurred in neurons, glia, and neuropil throughout the brain, some regions displayed relatively high perikaryal KGDHC enrichment. In the cerebral cortex, high immunoreactivity occurred mostly in layers III, V, and VI. The hippocampal pyramidal layer in CA1 and CA2 exhibited more intense staining than CA3. In the mammillary body, intensely labeled cells occurred in the supramammillary and lateral nuclei, while moderately stained cells predominated in the medial nucleus. The basal forebrain, basal ganglia, reticular and midline thalamic nuclei, red nucleus, pons, cranial nerve nuclei, inferior and superior colliculi, and cerebellar nuclei also contained highly immunoreactive neurons. The distribution of KGDHC overlapped with that of PDHC and colocalized to a limited extent with ChAT. These data are the first to demonstrate KGDHC immunoreactivity in discrete areas of rat brain and are vital to our understanding of selective vulnerability to metabolic insults and disease.

Relationship of calbindin D-28k with afferent neurons to the rostral ventrolateral medulla in the rat

The phenylethanolamine-N-methyltransferase (PNMT)-containing neurons in the rostral ventrolateral medulla (RVLM) (the C1 adrenergic group) have been implicated in the generation of the tonic sympathetic nerve activity. Using a double-labeling immunohistofluorescence technique, we found that 34.6 +/- 11.4% (mean +/- S.D.) of PNMT immunoreactive neurons in the RVLM were immunoreactive for Calbindin D-28k (CaBP), a Vitamin D-dependent calcium binding protein. Since CaBP is probably involved in regulating intracellular calcium concentrations in cells that are metabolically or electrically very active, our results suggest that at least some C1 adrenergic neurons (those containing calbindin) may have calcium mediated high metabolic or electrophysiologic activity that is associated with generating tonic nerve function. The RVLM has wide connections with many different nuclei in the brain which are known to contain clusters of neurons that express immunoreactivity to CaBP. In order to determine whether CaBP could be used as a molecular marker for projection neurons to the RVLM or to identify a subpopulation of projection neurons containing CaBP, we sought to determine the relationships between CaBP and the neurons that project to RVLM. Following injections of the retrograde tracer FluoroGold (FG) into the rat RVLM, sections containing retrogradely labeled neurons in (1) the nucleus tractus solitarii (NTS), (2) the contralateral RVLM, (3) the area postrema, (4) the mesencephalic central gray (mCG), (5) the lateral hypothalamus (LH), (6) the substantia innominata (SI), and (7) the paraventricular hypothalamic nucleus (PV) were tested for CaBP immunoreactivity. Although many retrogradely labeled neurons were found amidst many CaBP immunoreactive neurons in each of these nuclei, only a subpopulation of the retrogradely labeled neurons expressed CaBP immunoreactivity. The NTS demonstrated the higher proportion of double-labeled cells (mean 31.5 +/- 4.3%), whereas the lower proportion corresponded to the contralateral RVLM (mean 9.6 +/- 3.2%). On the other hand, both the retrogradely labeled neurons and the CaBP immunoreactive neurons in each of these nuclei were often found in regions containing a great number of adrenergic axons (i.e. immunoreactive for PNMT). Our results suggest that: (1) Two types of adrenergic RVLM neurons could be found, those containing CaBP and those lacking this calcium binding protein. (2) CaBP is not a common marker for the afferent neurons to the RVLM, but rather is found in selective subsets of them. (3) Both the non-CaBP projection neurons and the CaBP immunoreactive neurons in these nuclei may be innervated by adrenergic fibers.

Lesions of the rostral ventrolateral medulla reduce the cerebrovascular response to hypoxia

Sympathoexcitatory neurons of the rostral ventrolateral medulla are tonically active and required for maintenance of resting levels of arterial pressure. They are also selectively excited by hypoxia and responsible for the associated sympathoexcitation. Since electrical or chemical stimulation of RVL will increase regional cerebral blood flow (rCBF) independently of changes in regional cerebral glucose utilization (rCGU) we investigated whether the RVL was also required to maintain resting levels of rCBF and also participated in the cerebrovascular vasodilation elicited by hypoxia. Rats were anesthetized (chloralose; 40 mg/kg, s.c.), paralyzed (tubocurarine) and ventilated (100% O2). rCBF was measured in 10 dissected brain regions using [14C]iodoantipyrine; rCGU was measured by 2-deoxy-D-[14C]glucose. In controls (n = 6) rCBF ranged from 56 +/- 5 in corpus callosum to 101 +/- 6 ml/min x 100 g in inferior colliculus. Hypoxic-hypoxia (PaO2 = 36 +/- 1 mmHg, n = 6) increased rCBF in all structures maximally, at 204% of control, in occipital cortex. Hypercapnia (PaCO2 = 63.5 +/- 0.9, n = 5) also increased rCBF (P < 0.01) maximally to 299% of control in superior colliculus. Spinal cord transection with maintenance of arterial pressure did not affect resting rCBF and increased the vasodilation to hypoxia (PaO2 = 39 +/- 1 mmHg, n = 5) from 2- to 3-fold in all structures (P < 0.01). Bilateral lesions within the RVL had no effect on resting rCBF or rCGU. However, they significantly reduced, in all areas by 50-69% (P < 0.01, n = 5), the cerebrovascular dilation elicited by hypoxia but not hypercapnia.(ABSTRACT TRUNCATED AT 250 WORDS)

Ultrastructural localization and afferent sources of corticotropin-releasing factor in the rat rostral ventrolateral medulla: implications for central cardiovascular regulation

We investigated the ultrastructural localization, afferent sources, and arterial pressure effects of corticotropin-releasing factor (CRF) in the nucleus reticularis rostroventrolateralis (RVL), a region of the ventrolateral medulla containing C1 adrenergic neurons and sympatho-excitatory reticulospinal afferents to sympathetic preganglionic neurons. A polyclonal antibody to CRF was localized in acrolein-fixed sections through the rat RVL by the peroxidase-antiperoxidase (PAP) method. Light microscopy showed that 1-7 perikarya/30 micron section and numerous varicose processes contained CRF-like immunoreactivity (CRF-LI). By electron microscopy, CRF-LI was most intensely localized to large (80-100 nm) dense-core vesicles within numerous terminals and a few perikarya and large dendrites. Approximately half of the terminals containing CRF-LI were in direct contact with unlabeled perikarya or dendrites; the remainder were in apposition to either unlabeled terminals or astrocytes. Most synaptic specializations were asymmetric synapses on small, unlabeled dendrites. To examine potential extrinsic sources of CRF-containing terminals in the C1 area of the RVL, PAP immunocytochemical localization of CRF was combined with retrograde transport of wheat germ agglutinin-conjugated horseradish peroxidase (WGA-HRP). In all cases examined, a number of dually labeled neurons were found in the paraventricular nucleus (PVN) of the hypothalamus and a few dually labeled neurons were observed in the nuclei of the solitary tract; these labeled neurons were ipsilateral to the unilateral injection of WGA-HRP into the C1 area. Fewer dually labeled perikarya were detected in the lateral hypothalamic area and the lateral parabrachial nuclei, ipsilateral to the WGA-HRP injection. Additional physiological studies showed that bilateral microinjections of CRF into the C1 area of the RVL of urethane-anesthetized rats elicited a dose-related increase in arterial pressure. The results suggest that within the C1 area of the RVL, CRF released from terminals, arising predominantly from the PVN of the hypothalamus and probably from local neurons as well, may excite sympathoexcitatory reticulospinal neurons.

Localization of NADPH diaphorase in neurons of the rostral ventral medulla: possible role of nitric oxide in central autonomic regulation and oxygen chemoreception

We studied whether neurons containing nitric oxide synthase (NOS) are localized to the rostral ventrolateral medulla (RVM) and, if so, whether they are distinct from the adrenergic neurons of the C1 group. NOS-containing neurons and/or C1 neurons were visualized using NADPH diaphorase histochemistry and phenylethanolamine N-methyltransferase (PNMT) immunohistochemistry, respectively. A column of NADPH diaphorase-positive neurons, extending 2 mm in the rostrocaudal plane, was observed lateral to the inferior olive and medial to the C1 neurons. Double labelling studies showed that NADPH diaphorase-positive neurons were not immunoreactive for PNMT, indicating that the two enzymes were localized in the different cells. Furthermore, only a small fraction of NADPH diaphorase neurons were retrogradely labelled after injections of fluorogold into the thoracic cord. We conclude that the RVM contains a well-defined group of neurons endowed with NOS that are distinct from the adrenergic neurons of the C1 group and have only limited monosynaptic projections to the spinal cord. Since the RVM is involved in the control of arterial pressure and in oxygen-conserving reflexes, the findings raise the possibility that nitric oxide participates in central autonomic regulation and oxygen chemoreception.

Presumptive adrenergic neurons containing phenylethanolamine N-methyltransferase immunoreactivity in the medulla oblongata of neonatal swine

Given the importance of the swine (Sus scrofa) as an animal model for human development, physiology and disease, neurons containing the epinephrine-synthesizing enzyme, phenylethanolamine N-methyltransferase (PNMT), were mapped in the medulla oblongata of neonatal swine as a first step in identifying their roles in central autonomic control. Neurons were labeled immunocytochemically by using an antiserum to PNMT raised in rabbits against trypsin-treated enzyme purified from the bovine adrenal gland. The general regional organization of neurons expressing PNMT (-like) immunoreactivity (ir) in the neonatal swine was similar to data obtained in other species and, in some aspects, more closely resembled the pattern observed in the primate brain. Immunolabeled cells appeared to be more abundant and caudally more extensive than observed in other adult animals. PNMT-immunoreactive (ir) neuronal somata, however, were largely confined to the reticular formation in the ventrolateral quadrant and the nucleus tractus solitarii (NTS) and more restricted in distribution than those expressing tyrosine hydroxylase (TH) and dopamine beta-hydroxylase (D beta H)-ir on serial transverse sections. A close correspondence was observed between the distributions of TH- and PNMT-ir neurons and processes throughout the C1 and C2 areas. However, in the C1 and C3 regions TH-ir neurons outnumbered those containing D beta H and PNMT-ir. In contrast, cell groups enriched in PNMT-ir neurons and processes were characterized by relatively weak D beta H-ir. In the ventrolateral medulla (VLM), PNMT-ir cell bodies were concentrated rostrally and extended from the caudal pole of the facial nucleus to a level posterior to the calamus scriptorius. The rostral VLM was characterized by an admixture of bipolar and multipolar primarily medium-diameter immunostained neurons. A prominent cell column (condensation) organized ventromedially to the nucleus ambiguus pars compactus (NAc). A loosely organized cluster bordered the lateral aspect of the special visceral efferent column; another smaller aggregate was located in the ventromedial reticular formation adjacent to the inferior olive. At middle medullary levels, PNMT-ir neurons formed two distinct subgroups (dorsal and ventral) interrupted by a band of precerebellar relay neurons that extended between the medial and lateral limbs of the lateral reticular nucleus of Walberg. At obex, the dorsal cell group formed a diagonal array and assumed a position dorsal and dorsolateral to the medial limb of LRN. This group was distinguished by bipolar neurons with axes of orientation directed perpendicularly to the majority of neurons in the rostal VLM or those lying near the caudal ventromedullary surface.(ABSTRACT TRUNCATED AT 400 WORDS)

Stimulation of Cl area neurons globally increases regional cerebral blood flow but not metabolism

We examined the effects of electrical and chemical stimulation of the Cl area of the rostral ventrolateral medulla (RVL) on regional cerebral blood flow (rCBF) and regional cerebral glucose utilization (rCGU) in anesthetized (chloralose), paralyzed (curare) and ventilated rats. rCBF and rCGU were measured using 14C-iodoantipyrine (IAP) and 14C-deoxyglucose (2-DG), respectively, as indicators, with bilateral regional dissection of 11 brain regions. Electrical stimulation of the RVL elicited increases in arterial pressure (AP), heart rate (HR) and plasma concentration of epinephrine (EPI) and norepinephrine (NE). In addition, stimulation of the RVL, but not the adjacent medial longitudinal fasciculus, with AP maintained, increased rCBF (p less than 0.05, n = 6), but not rCGU, bilaterally and symmetrically (134-169% of control) throughout the brain. Bilateral adrenalectomy abolished the increase in plasma EPI elicited by stimulation of the RVL but did not affect resting rCBF (n = 5) or the elevation in rCBF elicited by RVL stimulation (n = 5). Increases in rCBF elicited by RVL stimulation were also unaffected by acute transection of the superior cervical ganglion (p greater than 0.05). Kainic acid (KA) microinjected into the RVL unilaterally (n = 6) at a dose producing sustained elevation in AP (5 nmol in 100 nl), elicited changes in rCBF similar to those elicited by electrical stimulation. We conclude that neurons within the RVL, possibly those of the adrenergic Cl group, can initiate a global cerebrovasodilation, but not an increase in rCGU, largely through neural pathways intrinsic to the brain. The responses may represent activation of networks in RVL mediating circulatory adjustments to hypoxia.

Ultrastructural localization of tyrosine hydroxylase immunoreactivity in the rat diagonal band of Broca

The present study sought to establish the cellular basis for the catecholaminergic (i.e., noradrenaline and dopamine) modulation of neurons in the horizontal limb of the diagonal band of Broca (HDB) in the rat brain. The light and electron microscopic localization of antigenic sites for a polyclonal antibody directed against the catecholamine synthesizing enzyme, tyrosine hydroxylase (TH), were examined in the HDB using a double-bridged, peroxidase-antiperoxidase method. By light microscopy, numerous punctate, varicose processes with intense TH-immunoreactivity (TH-I) were detected in the HDB. Additionally, a few small, bipolar, or multipolar TH-immunoreactive neurons were observed. Ultrastructural analysis of single sections revealed that the TH-labeled processes were axons and axon terminals. Axons (n = 134) with TH-I were primarily unmyelinated. Terminals with TH-I (n = 169) were 0.3-1.4 microns in diameter and contained many small, clear vesicles and 0-5 larger dense-core vesicles. The types of associations (i.e., asymmetric synapses, symmetric synapses, and appositions which lacked a membrane specialization in the plane of section analyzed) formed by the TH-labeled terminals were quantitatively evaluated. The TH-labeled terminals: (1) formed associations with unlabeled perikarya and dendrites (134 out of 169), (2) were closely apposed without glial intervention to unlabeled and TH-labeled terminals (11 out of 169), or (3) had no neuronal associations in the plane of section analyzed (24 out of 169). The relatively rare (n = 4) associations with unlabeled perikarya were mostly characterized by symmetric synaptic specializations. The majority of the TH-labeled terminals were associated with the shafts of small dendrites (66% of 134). Moreover, most of the associations on dendrites and dendritic spines were further characterized by asymmetric synaptic specializations; however, many were also appositions without any apparent glial intervention in the plane of section analyzed. Additionally, the TH-labeled terminals were often associated with only one dendrite, which, in the same plane of section, was sparsely innervated by other terminals. Astrocytic processes usually surrounded the portions of the terminals and dendrites not involved in the region of association. The TH-immunoreactive perikarya were small (7-12 microns), ovoid, and had an indented nucleus with some heterochromatin. Their scant cytoplasm contained mitochondria, Golgi complexes, and endoplasmic reticulum. A few immunoreactive dendrites, presumably derived from the local neurons, were also detected. Both TH-immunoreactive perikarya and dendrites were associated primarily with unlabeled terminals, although a few terminals with TH-I also contacted them.(ABSTRACT TRUNCATED AT 400 WORDS)

Morphology of rostral medullary neurons with intrinsic pacemaker activity in the rat

Neurons with regular ongoing activity attributable to intrinsic pacemaker properties were recorded in coronal tissue slices within the nucleus reticularis rostroventrolateralis of the rat medulla oblongata (RVL). The cells were injected with horseradish peroxidase or Lucifer yellow and their dendritic and proximal axonal characteristics were investigated (n = 15). These small-to-medium-sized neurons had a simple dendritic arborization (3-6 primary dendrites branching up to 3 times) apparently confined within the limits of nucleus RVL and with limited extension in the rostrocaudal direction. Their axons originated either from the cell body or from a primary dendrite and coursed in a dorsomedial direction without giving rise to local arborizations. It is concluded that RVL pacemaker neurons, presumed to represent a non-adrenergic class of sympathoexcitatory premotor neurons, exhibit characteristics reminiscent of the archetypal 'reticular core' neurons.

Strain difference in phenylethanolamine N-methyltransferase activity and immunoreactivity of medulla oblongata of Sprague-Dawley and Long-Evans hooded rats

Previously, we reported that retinal phenylethanolamine N-methyltransferase (PNMT) activity of Sprague-Dawley (SD) rats was significantly higher than that of Long-Evans (LE) hooded rats. However, there were no noticeable differences observed in PNMT-immunostaining patterns between the retinae of LE and SD rats. In order to examine this discrepancy, we extended this study to areas of the medulla oblongata harboring PNMT-containing cell bodies. In the present report, we demonstrate that the enzyme activity as well as the immunoreactivity of PNMT in medulla oblongata of SD rats were significantly higher than that observed in the LE strain.

Evidence for decreased transport of PNMT protein in advanced Alzheimer's disease

Phenylethanolamine N-methyltransferase (PNMT) is the rate-limiting enzyme in the synthesis of epinephrine and a specific marker for adrenergic neurons. PNMT protein is decreased in axon terminals in brains from patients with Alzheimer's disease due to retrograde degeneration of epinephrine neurons. To determine the subcellular mechanism underlying retrograde degeneration, the distribution of PNMT between axon terminal and cell body was calculated in early and advanced Alzheimer cases compared with age-matched controls. In early Alzheimer's disease there is a decrease in PNMT in axon terminals and in total PNMT in epinephrine cell bodies and terminals compared with control values. There is no difference in the ratio of PNMT in cell body/axon terminal compared with controls. In contrast, in advanced Alzheimer's disease, PNMT activity increases by 124% in epinephrine neuronal cell bodies compared with controls. Immunochemical titration shows that this increased enzyme activity is due to an increase in PNMT protein. The cell body/axon terminal ratio of PNMT is increased 2.5-fold in advanced Alzheimer's disease compared with controls. These findings are consistent with the hypothesis that in early Alzheimer's disease there is a decreased synthesis or increased degradation of PNMT. However, in advanced Alzheimer's disease we propose that the accumulation of this enzyme in the perikarya results from a diminished transport of PNMT to axon terminals. We further postulate that epinephrine, the product of PNMT, and its further metabolites are endogenous neurotoxins. Therefore, the accumulation of PNMT in epinephrine cell bodies may contribute to the degeneration of these neurons in Alzheimer's disease.

Excitation of rostral medullary pacemaker neurons with putative sympathoexcitatory function by cyclic AMP and beta-adrenoceptor agonists 'in vitro'

This study explores the mechanism of action of catecholamines on rostral medullary pacemaker neurons with putative sympathoexcitatory function, in tissue slices. The firing rate of the pacemaker neurons of nucleus reticularis rostroventrolateralis (RVL pacemakers) was reversibly increased by agents which elevate intracellular levels of cAMP (forskolin and 8-br-cAMP). Forskolin dideoxy, an analog without action on adenylate cyclase, was ineffective and adenosine, a potential degradation product of 8-br cAMP produced inhibition exclusively and only in high doses (0.1-1 mM). The firing rate of these cells was uniformly increased by epinephrine and isoproterenol (10 microM) but unaffected by both phenylephrine (100 microM) and clonidine (up to 1 microM). These effects were abolished by pretreatment with the beta-adrenoceptor antagonist propranolol (10 microM) but they were unaffected by the alpha-antagonist phentolamine (100 microM). The indirectly-acting sympathomimetic amine tyramine (0.1-1 mM) activated all the cells tested. The effect of tyramine was antagonized by the beta-blocker pindolol and was absent 7 days after microinjection of the neurotoxin 6-hydroxydopamine into the lateral aspect of the RVL. Intracellular recordings indicated that both isoproterenol and tyramine enhanced the rate of depolarization of the pacemaker neurons during the interspike interval and produced a decrease in input resistance. After tetrodotoxin (TTX) pretreatment, isoproterenol produced a depolarization also associated with a reduction in input resistance. Three conclusions are proposed. First, RVL pacemakers have functional beta-adrenergic receptors whose activation increases their discharge rate via the intracellular production of cAMP. The effect of cAMP is due at least in part to the activation of an inward current which may be carried by a cation. Secondly, RVL neurons are in close proximity to a releasable pool of catecholamines which is susceptible to destruction by the cytotoxic agent 6-hydroxydopamine (6-OHDA). Finally it is tentatively suggested that the reduction in sympathetic tone produced by centrally acting beta-blockers could be due, at least in part, to an action of these agents on RVL pacemaker cells.

Evidence for projections from the rostral medullary raphe onto medullary catecholamine neurons in the rat

Following the iontophoretic deposition of Phaseolus vulgaris leucoagglutinin (PHA-L) into the rostral medullary raphe, which included portions of the caudal nucleus raphe magnus, rostral nucleus raphe pallidus, rostral nucleus raphe obscurus and rostral nucleus reticularis paragigantocellularis, two-color immunoperoxidase staining was employed to demonstrate contiguity between PHA-L-immunoreactive (PHA-LI) varicose fibers and boutons and medullary catecholamine (CA) cells. Raphe projections were contiguous with phenylethanolamine N-methyltransferase-immunoreactive (PNMTI) neurons in the C1, C2 and C3 cell groups and with tyrosine hydroxylase-immunoreactive (THI) neurons in the A1 and A2 cell groups. Contiguity between PHA-LI processes and medullary CA cells was observed most frequently in the C1 cell group. Preliminary findings of this study have been presented previously.

Ultrastructural localization of somatostatin-like immunoreactivity in the rat dentate gyrus

Neurons containing somatostatin (SOM) are enriched in the dentate gyrus. We sought to establish the ultrastructural localization of this peptide in the dentate gyrus of the rat brain with a double-bridged peroxidase-antiperoxidase (PAP) method localizing antisera directed against somatostatin (SOM)-28 and SOM-28. Initial light microscopic observations confirmed that the majority of perikarya and thick varicose processes with intense SOM-like immunoreactivity (SOM-LI) were observed in the hilus. Fine varicose processes with SOM-LI were found throughout all layers of the dentate gyrus but were most intense in the outer third of the molecular layer (ML), where an occasional perikaryon with SOM-LI was seen. By electron microscopy, SOM-LI was found in neuronal perikarya, dendrites, axons, and axon terminals. Two types of SOM-containing perikarya were observed. The first type was small (6-10 microns), round or avoid, and had a labeled cytoplasma with abundant Golgi complexes and a dense accumulation of PAP-reaction product. The second type of perikarya was larger (11-16 microns) and had a more abundant cytoplasm than the first type, but the Golgi complexes did not appear labeled. Most (96% of 374) of the synapses on the SOM-labeled perikarya and dendrites were from terminals without SOM-LI which formed nearly equal proportions of asymmetric and symmetric junctions. The remainder of the presynaptic terminals contained SOM-LI and made primarily symmetric synapses. Synaptic junctions from both unlabeled and labeled terminals were primarily on the shafts of the small (0.5-1.5 microns) SOM-immunoreactive dendrites. The terminals with SOM-LI (0.25-1.3 microns) contained many small, clear vesicles and from zero to four large dense-core vesicles. Terminals with SOM-LI were associated 1) with one unlabeled perikaryon or dendrite (49% of 215 in the hilus; 76% of 326 in the ML); 2) with two unlabeled perikarya or dendrites simultaneously (5% hilus; 4% ML); and 3) with one SOM-containing perikaryon or dendrite (6% hilus; 3% ML). In all three types of associations, synaptic contacts on perikarya were few while the majority were with small (distal) dendrites. Moreover, most of the terminals with SOM-LI formed symmetric junctions or lacked membrane specializations but were without any apparent glial intervention in the plane of section analyzed. The remaining SOM-labeled terminals (40% hilus; 17% ML) were without any apparent synaptic relations. However, a few of these terminals were in direct apposition to other terminals, some of which were also SOM-immunoreactive.(ABSTRACT TRUNCATED AT 400 WORDS)

The C1 area of rostral ventrolateral medulla: a central site integrating autonomic responses to hemorrhage

Activation of the sympathetic neurons and release of adrenomedullary catecholamines are the principal early reflex responses to hemorrhage. These are initiated by arterial baro- and chemoreceptors, from other cardiopulmonary receptors, and by intracerebral receptors responding to ischemia. A principal gateway for integrating the autonomic responses are a small collection of neurons in a region of the rostral ventrolateral medulla (RVL), containing a cluster of neurons of the C1 adrenergic cell group, the C1 area. Neurons in the C1 area of RVL project exclusively to autonomic nuclei of the spinal cord, are tonically active, and fire with a rhythm linked to the cardiac cycle. They are essential for maintaining resting discharge of sympathetic nerves and, consequently, arterial pressure (AP) and heart rate. They also are critical for reflex changes in AP in the baro- and chemoreceptor, somato-sympathetic (pain), and cerebral ischemic reflexes. Neurons of the C1 area are under tonic excitatory and inhibitory control by pathways from other autonomic centers. They are controlled by a range of neurotransmitters including, gamma aminobutyric acid (GABA), acetylcholine, catecholamines, enkephalin, and several neuropeptides. They also serve as a site of action for the hypotensive actions of several clinically important neurotransmitters. The C1-area of RVL may play a critical role in the autonomic responses to hemorrhage and may be an important target for drugs seeking to treat hemorrhagic shock.

Ultrastructural localization of choline acetyltransferase in the rat rostral ventrolateral medulla: evidence for major synaptic relations with non-catecholaminergic neurons

Pharmacological and biochemical studies suggest that interactions between cholinergic and catecholaminergic and catecholaminergic neurons, particularly those of the C1 adrenergic cell group, in the rostral ventrolateral medulla (RVL) may be important in cardiovascular control. Ultrastructural localization of choline acetyltransferase (ChAT), the biosynthetic enzyme for acetylcholine, and its relation to neurons exhibiting immunoreactivity for catecholamine- (tyrosine hydroxylase; TH) or adrenaline (phenylethanolamine-N-methyltransferase; PNMT) -synthesizing enzymes were examined in the RVL using dual immunoautoradiographic and peroxidase anti-peroxidase (PAP) labeling methods. By light microscopy, the ChAT-immunoreactive neurons were located both dorsally (i.e. the nucleus ambiguus) and ventromedially to those labeled with TH or PNMT (TH/PNMT). A few ChAT-labeled processes were dispersed among TH/PNMT-containing neurons with the majority of overlap immediately ventral to the nucleus ambiguus. By electron microscopy, ChAT-immunoreactivity (ChAT-I) was detected in neuronal perikarya, dendrites, axons and axon terminals and in the vascular endothelial cells of certain blood vessels. The ChAT-labeled perikarya in the ventromedial RVL were medium-sized (15-20 microns), elongated, contained abundant cytoplasm and had had slightly indented nuclei. Synaptic junctions on ChAT-immunoreactive perikarya and dendrites were primarily symmetric with 64% (45 out of 70) of the presynaptic terminals unlabeled. The remaining terminals were immunoreactive for ChAT (30%) or TH/PNMT (6%). Terminals with ChAT-I were large (0.8-2.0 microns) and contained numerous small clear vesicles and 1-2 dense core vesicles. Seventy-seven percent (112 out of 145) of the ChAT-labeled terminals formed symmetric synapses with unlabeled perikarya and dendrites, whereas only 8% were with TH/PNMT-labeled perikarya and dendrites, and 15% were with ChAT-immunoreactive perikarya and dendrites. We conclude (1) that cholinergic neurons in the RVL principally terminate on and receive input from non-catecholaminergic neurons, and (2) that the reported sympathetic activation following application of cholinergic agents to the RVL may be mediated by cholinergic inhibition of local inhibitory interneurons. The observed synapses between ChAT and TH/PNMT-containing neurons suggests that cholinergic and adrenergic neurons additionally may exert a minor reciprocal control on each other and thus may modulate their response to the more abundant input from afferents containing other transmitters.

Cellular substrates for interactions between neurons containing phenylethanolamine N-methyltransferase and GABA in the nuclei of the solitary tracts

Adrenaline and gamma-aminobutyric acid (GABA) have been implicated in autonomic functions involving the intermediate and caudal portions of the medial nuclei of the solitary tracts (m-NTS). We sought to determine whether there was a cellular basis for direct intracellular or synaptic interactions between these transmitters in neurons in the m-NTS of rat brain by using dual-labeling immunocytochemical methods. Light microscopy revealed immunoautoradiographic labeling for the adrenaline-synthesizing enzyme phenylethanolamine N-methyltransferase (PNMT) in perikarya and processes in close proximity to cells demonstrating peroxidase reaction product for GABA. Electron microscopy of the intermediate m-NTS at the level of the area postrema further established the localization of immunoautoradiographic and peroxidase labels for PNMT and GABA in common as well as separate perikarya and dendrites. All axon terminals were labeled separately for PNMT and GABA. The PNMT-labeled terminals formed both symmetric and asymmetric synapses, whereas the GABA-labeled terminals formed exclusively symmetric synapses. Twenty-four percent (n = 42) of the PNMT- and 39% (n = 128) of the GABA-labeled terminals formed synaptic junctions on unlabeled soma and dendrites. Occasionally both types of terminals converged on a common unlabeled dendrite and on GABA-labeled dendrites. Only 3% of the PNMT- and 12% of the GABA-containing terminals formed synapses on PNMT-labeled soma and dendrites, whereas 7% of each type formed synapses with GABA-labeled profiles. The remaining labeled terminals lacked synaptic relations within the sections examined. The autoradiographic results were confirmed and extended by means of immunogold labeling for PNMT in combination with peroxidase-antiperoxidase localization of the GABA-synthesizing enzyme glutamic acid decarboxylase (GAD). GAD-labeled terminals formed symmetric synapses with dendrites that were either unlabeled or contained low levels of PNMT (gold particles) or PNMT and GAD. We conclude that in caudal, more cardiovascular portions of the NTS, adrenaline and GABA may coexist, but they are more commonly detected in separate populations of neurons having receptive sites for both transmitters and innervating certain common target neurons.

Cytoplasmic loop of beta-adrenergic receptors: synaptic and intracellular localization and relation to catecholaminergic neurons in the nuclei of the solitary tracts

Pharmacological studies suggest that beta-adrenergic receptors (beta AR) in the medial nuclei of the solitary tracts (m-NTS) facilitate presynaptic release of catecholamines and also function at postsynaptic sites. We have localized the antigenic sites for a monoclonal antibody against a peptide corresponding to amino acids 226-239 of beta AR in the m-NTS of rat brain. By light microscopy, immunoperoxidase labeling for this antibody was detected in somata and proximal processes of many small cells that were distributed throughout the rostrocaudal extent of the m-NTS. Electron microscopy confirmed the cytoplasmic localization of beta AR in perikarya and proximal dendrites of neurons. Immunoreactivity occurred as discrete patches associated with cytoplasmic surfaces of plasma membrane and with irregularly-shaped saccules with clear lumen in the immediate vicinity. Select regions of nuclear envelopes, mitochondrial membranes, and rough endoplasmic reticulum were also immunoreactive along their cytoplasmic surfaces. In contrast, the Golgi apparatus was labeled, but infrequently. Immunoreactivity was also detected at numerous post- and occasional presynaptic membrane specializations of select axodendritic junctions. Dual labeling for the beta AR-antibody by the immunoperoxidase method and for a rabbit antiserum against the catecholamine-synthesizing enzyme, tyrosine hydroxylase (TH), by the immunoautoradiographic method within the same sections, further established the precise cellular relations between beta AR and catecholaminergic neurons. Immunoreactivity for beta AR was detected in numerous perikarya and proximal dendrites that did not show detectable levels of TH. However, a few cells were dually labeled for both antigens, as seen by both light and electron microscopy. The TH-labeled terminals formed synapses at junctions both with and without beta AR-like immunoreactivity. These results from the single and dual labeling studies: (1) confirm biochemical predictions that amino acids 226-239 of beta AR protein reside intracellularly; (2) provide the first ultrastructural evidence for beta AR localization within both pre- and postsynaptic membrane specializations of a subset of catecholaminergic synapses; and (3) suggest select intracellular sites that may be involved with synthesis and/or internalization and degradation of the receptor protein.

Ultrastructural localization of tyrosine hydroxylase-like immunoreactivity in the rat hippocampal formation

The light and electron microscopic localization of antigenic sites for a polyclonal antiserum directed against the catecholamine synthesizing enzyme, tyrosine hydroxylase (TH), was examined in the hippocampal formation of the rat brain with a double-bridged peroxidase-antiperoxidase method. By light microscopy, the majority of varicose processes with intense TH-like immunoreactivity (LI) were contained in the hilus of the dentate gyrus (DG) and strata radiatum and lacunosum-moleculare of the CA3 region of the hippocampus. Only a few immunoreactive fibers were observed in the molecular and granule cell layers of the DG, in strata oriens and pyramidale of CA3, and in all layers of CA1. Electron microscopy confirmed that these labeled processes were primarily axons and axon terminals. Terminals with TH-LI were 0.4-1.1 micron in diameter and contained many small clear vesicles and from 0 to 3 larger dense-core vesicles. The number and types of associations formed by terminals with TH-LI were remarkably similar in the DG and hippocampus proper despite known differences in intrinsic cells and function. In both regions, the majority of terminals with TH-LI formed junctions on small (distal dendrites (52% of 112 in the DG; 67% of 116 in CA3) and dendritic spines (30% in the DG; 18% in CA3) that were both asymmetric and symmetric. In the DG, axosomatic junctions (2% of 112) were symmetric and occurred exclusively on the perikarya of granule cells, whereas junctions on large (proximal) dendrites were more numerous (16%), exhibited symmetric as well as asymmetric membrane specializations, and were of both granule (molecular layer) and nongranule (hilus) cell origin. In CA3, synaptic contacts on perikarya (5% of 116) and large (proximal) dendrites (10%) of both pyramidal cell and nonpyramidal cell origin were few and all symmetric. The distribution and types of synaptic associations formed by terminals with TH-LI in the CA1 region paralleled that seen in the CA3 region. In both the dentate and hippocampus proper, 10% of the terminals with TH-LI were observed closely apposed to unlabeled terminals that formed asymmetric synapses with dendrites and dendritic spines. In rare instances, TH-immunoreactive terminals were found in close association with the basement membrane of blood vessels, astrocytic processes, or with other unlabeled terminals not forming recognizable junctions. In addition TH-LI was occasionally detected within the cytoplasm of a minority of astrocytes.(ABSTRACT TRUNCATED AT 400 WORDS)