Catecholaminergic neurons in the ventrolateral medulla and nucleus of the solitary tract in the...

17
THE JOURNAL OF COMPARATIVE NEUROLOGY 273:224-240 (1988) Catecholaminergic Neurons in the Ventrolateral Medulla and Nucleus of the Solitary Tract in the Human VICTORIA ARANGO, DAVID A. RUGGIERO, JANIE L. CALLAWAY, MUHAMMAD ANWAR, J. JOHN M A ” , AND DONALD J. REIS Department of Psychiatry, Laboratory of Psychopharmacology (V.A., J. J.M.), and Department of Neurology, Division of Neurobiology (V.A., D.A.R., J.L.C., M.A., D.J.R.), Cornell University Medical College, New York, New York 10021 ABSTRACT Catecholaminergic neurons in the ventrolateral medulla (VLM) and nucleus of the solitary tract (NTS) are important because of their presumed roles in autonomic regulation, including the tonic and reflex control of arterial pressure, neuroendocrine functions, and the chemosensitivity asso- ciated with the ventral medullary surface. However, little is known about the connections of these neurons in the human brain. As a first step in analyzing the functional biochemical anatomy of catecholamine neurons in the human, we used antisera against tyrosine hydroxylase (TH) and phenyl- ethanolamine N-methyltransferase (PNMT) to localize medullary catechol- amine-containing neurons and processes in the VLM and the NTS. Cells staining for TH were located throughout the VLM. Most cells staining for TH and PNMT, which are therefore adrenergic, occurred in an area of the VLM probably corresponding to the rostroventrolateral reticular nucleus. Axons of TH-immunoreactive neurons in the VLM projected (1) dorsally, in a series of parallel transtegmental trajectories, toward the dor- somedial reticular formation, the NTS, and vagal motor nucleus, (2) longi- tudinally, through the central tegmental field, as fascicles running parallel to the neuraxis, (3) ventrolaterally toward the ventral surface (VS) of the rostra1 VLM where they appeared to terminate, and (4) medially into the raphe, where they arborized. Similar systems of fibers were labeled for PNMT; the longitudinal bundles of PNMT-labeled axons were limited to the principal tegmental bundle and concentrated dorsally. Fibers containing PNMT were also identified in the medullary raphe, on the medullary ventral surface, and contacting intraparenchymal blood vessels. In the NTS, neu- rons exhibited immunoreadivity to both TH and PNMT Four principal subgroups of TH-immunoreactive neurons were seen: a ventral, an interme- diate, a medial, and a dorsal group. Perikarya containing PNMT were restricted to the dorsolateral aspect of the NTS. Processes containing TH and PNMT immunoreactivity were identified in the medial and dorsolateral NTS; others appeared to project between the NTS and the VLM and within the solitary tract. The presence of catecholaminergic fibers of the VLM interconnecting with the NTS, raphe, intraparenchymal microvessels, VS, and possibly the spinal cord suggests that the autonomic and chemoreceptor functions attrib- uted to these neurons also may apply to the human. Key words: medulla oblongata, autonomic nuclei, tyrosine hydroxyiase, phenylethanolamine N-methyltransferase Accepted February 15,1988. 0 1988 ALAN R. LISS, INC.

Transcript of Catecholaminergic neurons in the ventrolateral medulla and nucleus of the solitary tract in the...

THE JOURNAL OF COMPARATIVE NEUROLOGY 273:224-240 (1988)

Catecholaminergic Neurons in the Ventrolateral Medulla and Nucleus of the

Solitary Tract in the Human

VICTORIA ARANGO, DAVID A. RUGGIERO, JANIE L. CALLAWAY, MUHAMMAD ANWAR, J. JOHN M A ” , AND DONALD J. REIS

Department of Psychiatry, Laboratory of Psychopharmacology (V.A., J. J.M.), and Department of Neurology, Division of Neurobiology (V.A., D.A.R., J.L.C., M.A., D.J.R.),

Cornell University Medical College, New York, New York 10021

ABSTRACT Catecholaminergic neurons in the ventrolateral medulla (VLM) and

nucleus of the solitary tract (NTS) are important because of their presumed roles in autonomic regulation, including the tonic and reflex control of arterial pressure, neuroendocrine functions, and the chemosensitivity asso- ciated with the ventral medullary surface. However, little is known about the connections of these neurons in the human brain. As a first step in analyzing the functional biochemical anatomy of catecholamine neurons in the human, we used antisera against tyrosine hydroxylase (TH) and phenyl- ethanolamine N-methyltransferase (PNMT) to localize medullary catechol- amine-containing neurons and processes in the VLM and the NTS.

Cells staining for TH were located throughout the VLM. Most cells staining for TH and PNMT, which are therefore adrenergic, occurred in an area of the VLM probably corresponding to the rostroventrolateral reticular nucleus. Axons of TH-immunoreactive neurons in the VLM projected (1) dorsally, in a series of parallel transtegmental trajectories, toward the dor- somedial reticular formation, the NTS, and vagal motor nucleus, (2) longi- tudinally, through the central tegmental field, as fascicles running parallel to the neuraxis, (3) ventrolaterally toward the ventral surface (VS) of the rostra1 VLM where they appeared to terminate, and (4) medially into the raphe, where they arborized. Similar systems of fibers were labeled for PNMT; the longitudinal bundles of PNMT-labeled axons were limited to the principal tegmental bundle and concentrated dorsally. Fibers containing PNMT were also identified in the medullary raphe, on the medullary ventral surface, and contacting intraparenchymal blood vessels. In the NTS, neu- rons exhibited immunoreadivity to both TH and PNMT Four principal subgroups of TH-immunoreactive neurons were seen: a ventral, an interme- diate, a medial, and a dorsal group. Perikarya containing PNMT were restricted to the dorsolateral aspect of the NTS. Processes containing TH and PNMT immunoreactivity were identified in the medial and dorsolateral NTS; others appeared to project between the NTS and the VLM and within the solitary tract.

The presence of catecholaminergic fibers of the VLM interconnecting with the NTS, raphe, intraparenchymal microvessels, VS, and possibly the spinal cord suggests that the autonomic and chemoreceptor functions attrib- uted to these neurons also may apply to the human.

Key words: medulla oblongata, autonomic nuclei, tyrosine hydroxyiase, phenylethanolamine N-methyltransferase

Accepted February 15,1988.

0 1988 ALAN R. LISS, INC.

CATECHOLAMINE NEURONS IN HUMAN MEDULLA 225

Abbreviatwns

bv blood vessel CNV DAO dorsal accessory olive DCN dorsal cochlear nucleus D ECN external cuneate nucleus I ICP inferior cerebellar peduncle IVN inferior vestibular nucleus LRN(s) lateral reticular nucleus (subtrigeminal division) M medial division of the nucleus of the solitary tract MA0 medial accessory olive MLF medial longitudinal fasciculus MVN medial vestibular nucleus NC nucleus cuneatus NG nucleus gracilis NTS(r) nucleus of the solitary tract (rostral division) P pyramid PION principal inferior olivary nucleus PP nucleus prepositus PT principal tegmental adrenergic bundle RO nucleus raphe obscurus RVL rostroventrolateral reticular nucleus Rd dorsoreticular nucleus SGs subependymal glial layer TR solitary tract Tt transtegmental adrenergic fiber systems V VS ventral surface V spinal trigeminal nucleus X XI1 hypoglossal nucleus

commissural nucleus of the vagus

dorsal division of the nucleus of the solitary tract

intermediate division of the nucleus of the solitary tract

ventral division of the nucleus of the solitary tract

dorsal motor nucleus of the vagus

"he importance of the central catecholaminergic neuron lies in its presumed role in nociception (Jensen and Yaksh, '84; Jones and Gebhart, '86), neuroendocrine control mub- instein and Sawyer, '701, the sleep-wake cycle (see review by Marley and Stephenson, ,721, various disease states in- cluding Parkinson's disease (Hornykiewicz, '66), affective disorders (Schildkraut, '651, and as described recently, in autonomic functions including the regulation of respiration and arterial blood pressure (Chalmers et al., '78; Reis, '85). However, most of the information presently available on the distribution of catecholaminergic neurons has been ob- tained from studies in the cat (Coote and McLeod, '74; Poitras and Parent, '78; Blessing et al., '80; Jones and Friedman, '83; Ciriello et al., '86; Kitahama et al., '86; Reiner and Vincent, '86; Ruggiero et al., '86), rat (Hokfelt et al., '74, '84; Armstrong et al., '82; Everitt et al., '83; Kalia et al., '85a,b; Ruggiero et al., '85; Tucker et al., '87), and nonhuman primate (Di Carlo et al., '73; Hubbard and Di Carlo, '73; Feften et al., '74; Gamer and Sladek, '75; Schofield and Everitt, '81; Tanaka et al., '82; Satoh and Fibiger, '85). Much less is known about these systems in the human brain.

Catecholaminergic cell bodies in the human medulla ob- longata have been described on the basis of catecholamine histofluorescence, the regional activity and content of tyro- sine hydroxylase (TH: the rate-limiting enzyme of catechol- amine biosynthesis), phenylethanolamine N-methyl- transferase (PNMT the enzyme synthesizing adrenaline), and neuromelanin (Lew et al., '77; Mefford et al., '77; Kopp et al., '79; Saper and Petito, '82). These methods, however, lack the resolving power of immunocytochemistry, particu- larly for localization of fiber tracts.

As a first step in analyzing the biochemical anatomy of central autonomic neurons in the human, we have used

antisera against TH and PNMT to localize catecholamine- containing neurons and fibers in the medulla oblongata, immunocytochemically. We have focused on the adrenergic connections of the C1 area (Hokfelt et al., '74) of the rostral ventrolateral medulla (VLM) and the nucleus of the solitary tract (NTS). Data from this report extend previous findings in the human based on immunocytochemistry (Pearson et al., '83; Kitahama et al., '85) and reveal several new cate- cholamine systems through which adrenergic neurons may influence cardiopulmonary and neuroendocrine functions, intracerebral blood flow, and the chemosensitivity associ- ated with the ventral medullary surface (VS).

Some of this work was previously published in abstract form (Callaway et al., '87).

MATERIALS AND METHODS Human brainstems (< 9 hours postmortem) were col-

lected in crushed ice at autopsy in New York Hospital. Samples were collected in the course of routine autopsies being carried out for medical indications. Permission for the autopsies and use of the tissue samples was obtained by the responsible physicians in accordance with the re- quirements of the Human Rights in Research Committees. Neuropathology was excluded by macroscopic and micro- scopic examination of the brain. Samples from the medulla oblongata of six subjects (Table 1) were blocked into 4-6- nun slabs, fixed in four changes of 4% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.4) for 6 hours at 4"C, and cryoprotected by infiltration in 30% sucrose in 0.1 M phosphate buffer overnight.

Sections were cut transversely (35 pm thick) while frozen with a sliding microtome and collected in spot wells con- taining 0.1 M phosphate buffer. Endogenous peroxidase was removed by incubation in 3% Hz02 in 0.1 M Tris-saline (pH 7.5) for 30 minutes. Sections were labeled, immunocy- tochemically, by using a modification (Pickel, '81) of the peroxide-antiperoxidase (PAP) method of Sternberger ('79). Briefly, nonspecific binding was blocked by 1:30 nor- mal goat serum in 0.1 M Tris-saline for 30 minutes. Sec- tions were incubated overnight in primary antiserum with 0.2% Triton-X-100 in 1% goat serum in Tris-saline. Antisera to TH (1500) and PNMT (1:2,000) were provided by Drs. Tong Joh and Dong Park, respectively. Both antisera were prepared in rabbits and checked for specificity by methods described previously (Joh and Goldstein, '73; Park et al., '82; Joh and Ross, '83). After rinsing in 1% goat serum in Trk-saline, sections were incubated in a 1:50 dilution of goat antirabbit immunoglobulin (IgG, Miles) for 45 min- utes, rinsed, and transferred to a 1:lOO dilution of PAP (prepared in our laboratory) for 1 hour. The chromogen used for visualization of the bound peroxidase was 3' 3'-diamino- benzidine (DAB, 22 mgllOO ml Tris-saline and 10 pl 30% HzOZ for 6 minutes). The reaction was stopped by immer- sion in a bath of Tris-saline. Controls were obtained by omitting the primary antibody from the sequential incuba- tions described above. Sections were rinsed in water, mounted from phosphate buffer containing 0.5% gelatin, air dried, defatted, dehydrated, cleared, and coverslipped. The sections, examined under brightfield or darMield mi- croscopy, were drawn with the aid of an overhead projector or a camera lucida and photographed. Adjacent sections, stained for Nissl substance, were used to determine nuclear boundaries.

226 V. AHANGO ET AL. TABLE 1. Patient Statistics

Postmortem delay Age Gender Race Cause of death (hours)

85 M Caucasian Cardiac arrest 3.5 63 M Caucasian Lymphoma 9 60 M Caucasian Renal cell carcinoma 9 31 M Hispanic Carcinoma of stomach 8 64 M Black Precipitation 6.5 66 M Caucasian Carcinoma of lung 9

RESULTS Axons of TH-immunoreactive neurons in the VLM pro- N~~~~~ containing TH and PNMT were identified in the jected dorsally in a series of parallel trajectories (the trans-

human medulla. I-unoreactive neuTons were recognized tegmental tracts, described in the rat by Granata et al., '85) under brightfield illumination by a characteristic dark toward the dorsomedial reticular formation and the NTS- brown to black reaction produd which distinguished from x. In sections though the rostra1 medulla some Of these background and densely filled perikarya (cytoplasm) and fibers, originating in the rostral medulla, contributed to a their processes. Brightfield illumination was best suited for longitudinally organized bundle of fascicles (the principal viewing immunoreactive cell bodies (Figs. 2,3, 8, s), adrenergic tegmental bundle, PT) concentrated in the dor- whereas processes were seen more clearly when viewed so~ed ia l tegmentm. In the rostra1 m M , the tract was with darkfield optics (Figs. 4,5,6a). located ventral to the medial vestibular nucleus and ventro-

Specificity of staining was judged by the above criterion medial to the rostral aspect of the NTS. At caudal levels, and by the preferential labeling of structures and fiber the entire tract shifted medially and laid adjacent to the trads shown previously to be catecholaminergic. Control hypoglossal nucleus and ventromedial to the medial aspect sections did not contain immunoreactivity. Cells containing of t@ NTS (Fig. 1 b d Although there was a major concen- TH and PNMT immunoreactivity were observed primarily tration of fibers in the P", a more widespread but contig- in the VLM and in the NTS (Fig. 1). uous system of longitudinally organized fiber tracts

extended farther ventrally toward the VLM. These tracts (arranged as groups of fascicles cut in cross section) were

Cells were immunolabeled for both TH and P m T in the intersected bY fibers belonging to the transtegmental arc VLM. Perikarya containing TH immunoreactivity ex- systems, as described above. The more laterally situated tended from the rostral aspect of the VLM (at the level of fibers ofthe transte-ental arcs originating in part in the the rostral subvestibular portion of the NTS) into the tau- VLM and cowsing within, or oblique to, the plane of sec- dal aspect of the VLM (at the level of the commissural tion) traversed the entire medullW' te€Wentm and

of the vagus, CNV, Fig. 1a-c). me majority of sprayed dorsally into the NTS along both sides of the soli- labeled cells occurred in close proximity to the ventral tW' tract (Fig- 1b). 5h-m of the fibers labeled for TH also surface in an area in the rostra1 VLM equivalent to the may have coursed in an antiparallel direction (from the rostroventrolateral reticular nucleus (RVL) as defined NTs to the VLM) Or Stemmed from fibers tracts descending physiologically, immunocytochemically, and on the basis of from other catecholaminergic nuclei. The darkfield photo- connectivity data in rats and cats (Ross et a]., '83, 'Ma$, montage in Figure 4 dm-~onstrates the complex networks '88; Gatti et al., '85; Granata et al., '85; Ruggiero et al., '85, formed by the TH-immunoreactive '86; Benarroc. et Other fibers labeled for TH projected ventrolaterally to- m a y of m-immunoreactive neurons extended dorsally ward the ventral medullary surface of the rostral VLM and from the VLM toward the NTS, dorsal motor nucleus ofthe ended in a field Of PUnctate varicosities lining the surface vagus (X), and intercalated nucleus. The photomicrograph (Fig. 5a). small bundle of axons also descended close to in Figure 2 illustrates TH-labeled cells in the rostra1 VLM the ventral surface and, at caudal levels, sprayed medially and their topographic relationship to the ventral medullary along the surface and ventral to the lateral reticular "cere- surface. N~~~~ in the VLM and dorsal reticular forma- bellar relay'' nucleus (Fig. Id. It is presumed that these tion were fmzOrm, triangular, and multipolar in shape and fibers were destined for the ventral raphe and/or the spinal ranged between 18 and 35 pm in their long axes, A few cord. A few fibers were also followed dorsolaterally from neurons were larger (> 35p.m; Fig. 9 ~ ) . The long axes ofthe the rostral VLM~ passing laterally to the spinal t r igemid

bundles located dorsal to it) were followed medially into the N~~~~~ exhibiting PNMT immunoreact~v~ty were iden- raphe where they arborized. In the nucleus raphe obscurus,

morpho~ogi~a~~y to those containing TH @ig. 3). They vertically organized fiber networks (axons) admixed with restricted in localization and occurred punctate varicosities (putative terminals) predominated at

Ventrolateral medulla

'86). Throughout the medulla, an

TH-immunoreactive pefikarya were generally oriented in tract. a ventrolateral to dorsomedial direction; on the average,

were to trace the% fibers any further- fiojections from the rostral VLM (and the longitudinal

these neurons had two to four dendrites.

were, however, in the rostral VLM, and to a lesser extent, at intermediate levels of the VLM (Fig. ld-0. No PNMT-immunoreactive cells were found in the caudal aspect of the VLM (Fig. 10

Fig. 1. Camera lucida drawings of TH (a-c) and PNMT (d-0 immuno- in the reactive cell bodies (dots), processes (lines), and putative terminal fields

(stippling) in the medulla oblongata of the human, illustrated at three rostrocaudal levels. The drawings at the top of the page are most anterior.

and

bers immunostained for TH.

few were Observed farther tegmental in contrast to the large

CATECHOLAMINE NEURONS IN HUMAN MEDULLA 227

a

Figure 1

228 V. ARANGO ET AL.

Fig. 2. Low-power photomicrograph of TH-immunoreactive perikarya in the rostral VLM (RVL), taken at the level of Figure lb . Note the topographic relationship of these catecholamine-synthesizing neurons to the ventral

rostral medullary levels although some were identified at caudal levels as well (Figs. la-c, 5b). Numerous close con- tacts were observed between TH-immunoreactive processes and blood vessels throughout the VLM (Fig. 6a).

Similar systems of fibers were also labeled with the PNMT antiserum. The longitudinal fiber systems, in part originat- ing in the rostral VLM, and containing PNMT immunore- activity were limited to the principal tegmental bundle and concentrated dorsally. PNMT-containing processes ap- peared to be in close apposition to intraparenchymal blood vessels (Fig. 6b). PNMT-labeled fibers (resembling axons) were also identified projecting from neurons in the rostral VLM onto the ventral medullary surface and into the med- ullary raphe. Markedly fewer axons were found projecting into or away from the NTS.

Nucleus of the solitary tract In the dorsomedial medulla, neurons showed immunore-

activity to both TH and PNMT. Perikarya containing TH were concentrated in the intermediate one-third of the NTS and extended caudally into the C N V and rostrally into the subvestibular aspect of the nucleus Figs. lb,c, 7). Four principal subgroups of TH-immunoreactive neurons were seen: (1) A ventral group-a linear array of fusiform and ovoid (some pear-shaped) cell bodies, lying adjacent to, or overlapping, the dorsal motor nucleus of the vagus and the intercalated nucleus, containing two to three dendrites and

medullary surface, which is indicated by arrowheads. Bar = 100 pm. Pro- cesses reaching the ventral surface are not visible under brightfield illumi- nation (see Fig. 5a). Arrow indicates a cell which is enlarged in Figure 9c.

measuring approximately 25-40 pm in their long axes. The orientation of many of these cells (and their processes) was primarily in a ventrolateral to dorsomedial direction. Pho- tomicrographs of this group of cells are provided in Figures 8a and 9a. (2) A dorsal group-arrays of smaller, ovoid and fusiform cell bodies lining the dorsal and lateral extremi- ties of the NTS and CNV, containing one to two dendrites, and on the average measuring approximately 20-25 pm in their long axes. Some larger neurons (40-50 pm) were also seen, although it was the smaller neurons that character- ized this subgroup (see arrows in Fig. 8b). Depending on their location the long axes of these perikarya were ori- ented medially, dorsally, or dorsomedially. (3) An interme- diate group-the TH-immunoreactive cells were situated in the intermediate subnucleus of the NTS, medial to the solitary tract, and between the above cell groups. Processes of these neurons were organized in a dorsal (dorsomedial) to ventral (ventrolateral) direction (Figs. 8c, 9b). The size and shape of these immunoreactive neurons resembled those of the ventral group. (4) A medial group-a smaller group of cells, scattered and comparatively lightly labeled, was also admixed with a massive field of punctate varicosi- ties in an area of the medial NTS directly underlying the area postrema and rostral to it. A few cells were also seen in the dorsal reticular formation underlying the NTS. In contrast to our data in the rat (and as observed in the cat, Ruggiero et al., '861, few or no neurons were labeled for TH

CATECHOLAMINE NEURONS IN HUMAN MEDULLA 229

Fig. 3. a: Low-power photomicrograph of a 30-pm section of human ros- tral medulla treated with antibodies to PNMT (1:2,000). The rostrocaudal

level is that depicted in Figure Id. b: Enlargement of the area boxed in a, illustrating the adrenaline-synthesizing neurons of the C1 area of the RVL. Bar = 100 pm.

230 V. ARANGO ET AL.

Fig. 4. A low-power photomicrograph of a 30-pm section (inset) through the human medulla treated with antibodies to TH (1:500). On the left, a montage of darkfield photomicrographs (corresponding to the bracketed area in inset) illustrates the complexity of fiber systems containing TH

immunoreactivity. The upper portion of the montage demonstrates fields of punctate varicosities in the NTS derived partly from fibers (transtegmental tracts depicted in Fig. l a d originating in the RVL (lower right) (Fig. la- c). These photographs correspond to the level in Figure lb.

in the rostra1 aspect of the dorsomedial medullary tegmen- TH-labeled axons were traced from cells in the VLM into tum, medial longitudinal fasciculus, or periventricular gray. the ventral aspect of the NTS and, as described previously,

Neurons showing PNMT immunoreactivity were limited entered the nucleus from both sides of the solitary tract to the dorsalmost aspect of the NTS. These were similar (Fig. la-c). Some fibers also appeared to project between morphologically and topographically to those containing the dorsal adrenergic bundle and the ventromedial division TH immunoreactivity (Fig. 1 e-f). of the NTS. Labeled punctate varicosities were concen-

Fig. 5. Darkfield photomicrographs demonstrating TH-immunoreactive varicosities lining the ventral medullary surface (arrows) of the RVL (a) as well as labeled axons (characterized by en passant swellings) and putative

terminals (small punctate varicosities) in the nucleus raphe obscurus (b). Both of these fields were traced from immunoreactive perikarya in the RVL. Bars = 100 pm.

Fig. 6. The upper panel (a) is a darkfield photomicrograph of TH-immu- noreactive processes in the medulla. Note the close proximity of numerous labeled processes to a blood vessel located in an area of the medullary reticular formation lying dorsomedially to the RVL. The lower panel (b) is

a brightfield photomicrograph of TH-immunoreactive processes encircling a blood vessel in the NTS. Bars = 100 pm. Similar associations were found between PNMT-immunoreactive processes and blood vessels, as seen in the rat and cat (Ruggiero et al., '85, '86).

CATECHOLAMINE NEURONS IN HUMAN MEDULLA 233 CATECHOLAMINE NEURONS IN HUMAN MEDULLA 233

/

\ '

.

\

Fig. 7. Camera lucida drawings of an intermediate level of the human medulla oblongata. Inset demonstrates neurons labeled immunocytochemi- cally for TH in the NTS, the RVL, and the central tegmental field. The enlargement demonstrates terminal-like fields containing TH in the dorsal (D) and medial (M) subnuclei of the NTS. Note the substructural organiza- tion of TH-immunoreactive neurons into dorsal (D), intermediate (I), and

medial (MI subnuclei. Labeled neurons in the ventral subnucleus (V), al- though similar morphologically, formed a separate cluster, distinct from the intermediate subgroup, and in close proximity to the dorsal motor nucleus of the vagus (XI. Whether some of these catecholaminergic cell bodies, in V, are constituents of X remains to be established. The medial border of subgroup V is indicated by an arrow.

Fig. 8. Brightfield photomicrographs of TH-immunoreactive perikarya in the subgroups of the NTS of the right side. a: The ventral group consists of large neurons whose long axes (or dendritic trees) are oriented primarily in a ventrolateral (or dorsomedial) direction. b: The dorsal group consists of

arrays of smaller neurons (arrows) lining, in this case, the dorsolateral extremity of the NTS. On some sections, as shown here, large perikarya are also observed. c: The intermediate group, located medial to the solitary tract, contains mostly large neurons. Bar = 50 fim.

CATECHOLAMINE NEURONS I N HUMAN MEDULLA

trated at caudal and adjacent intermediate levels of the nucleus and were less dense in its rostral one-third. An extremely dense field was present in the medial aspect of NTS subadjacent to the glial subependymal layer (SGs, Fig. 7) and appeared to extend rostrally into the medial aspect of the subvestibular NTS and dorsal motor nucleus of the vagus. Axons extending from the dorsolateral area of the NTS ran ventromedially in the NTS adjacent to the SGs; whether these axons originated from neurons in the NTS or another source could not be determined. A smaller con- tiguous field of fine processes, admixed with immunoreac- tive cell bodies, was seen in the dorsolateral NTS. Den- drites, as well as axons and terminal-like varicosities of the intrinsic neurons, also appeared to ramify within the NTS although their precise contribution could not be differen- tiated from the processes derived from extrinsic sources. Dendrites sectioned transversely or obliquely were, of course, often difficult or impossible to distinguish from cut axons with the light microscope. Both TH (Fig. 6b)- and PNMT-immunoreactive processes were identified in the NTS in close proximity to blood vessels. Other immunoreac- tive processes, sectioned obliquely or transversely and or- ganized as fascicles, were identified in the solitary tract (Fig. 10).

Comparatively fewer fibers were labeled in the NTS on tissues treated with the PNMT antiserum. PNMT-immu- noreactive fibers traversed the body of the intermediate one-third of the nucleus and ended in the region of the dorsal motor nucleus of the vagus. Some of the fibers ap- peared to originate from neurons labeled in the dorsal group. This impression was based on studies of their fiber trajectories and on the fact that relatively few fibers labeled for PNMT entered the ventral NTS via the transtegmental fiber system as described above.

DISCUSSION This study presents new observations on the distribution

of catecholamine-synthesizing neurons and processes in the human VLM and the NTS. By using darkfield optics in conjunction with the sensitive PAP technique, we were able to observe in greater detail than observed previously the distributions of TH- and PNMT-immunoreactive dendrites, axons, and fine punctate varicosities resembling terminal boutons. Another factor which contributed to higher reso- lution was that our data were obtained from brains which had a postmortem delay of 3.5-9 hours. The short postmor- tem delay, while essential for achieving intense immuno- reactive staining of cell bodies, is even more crucial in attaining good fiber staining. In the present study, immu- noreactive fibers were best visualized in those cases with short postmortem intervals, which appears to be the most important factor influencing staining, more so than modi- fying the parameters of the technique.

The purpose of this study was to explore further the distribution of the catecholaminergic enzymes TH and PNMT in the human medulla and focus on the connectivity of the rostral VLM. As observed from our material, labeled neurons in the rostral VLM identify a substructure of the quadrangular nucleus paragigantocellularis lateralis de- fined by the cytoarchitectonic studies of Olszewski and Bax- ter ('54), probably corresponding to the RVL of Ross et al. ('84a, '86). Results of this immunocytochemical study ex- tend those which investigated the anatomy of catechol- amine neurons in the brainstem of humans and other primates (Di Carlo et al., '73; Hubbard and Di Carlo, '73; Nobin and Bjorklund, '73; Felten et al., '74; Garver and

235

Sladek, '75; Pearson et al., '83; Kitahama et al., '85; Satoh and Fibiger, '85; Spencer et al., '85). The present study defined new connections of two areas critical in the regula- tion of arterial blood pressure and other autonomic func- tions, the VLM and the NTS, and established several homologies between humans and other species (i.e., the rat, cat, and non-human primate):

(1) The rostral VLM and a dorsolateral subgroup of the NTS of the human harbor perikarya containing both TH and PNMT and, therefore, likely synthesize adrenaline, as reported in the cat (Ciriello et al., '86; Kitahama et al., '86; Ruggiero et al., '86) and rat (Armstrong et al., '82; Hokfelt et al., '74, '84; Kalia et al., '85a,b; Ruggiero et al., '85). These findings confirm and extend another study in the human, in which PNMT-immunoreactive cells were re- ported to lie in the rostral VLM and in a dorsal region of the NTS (Kitahama et al., '85). Based on our studies of the rostral VLM, the geometries of neurons labeled for TH and PNMT in the human agree with the observations of Nissl- stained material by Olszewski and Baxter ('54): "medium- sized, multipolar and slightly elongated . . . with their long axes directed dorsomedially." In the rat, catecholaminergic neurons in the C1 area of the RVL contain both TH and PNMT and primarily synthesize adrenaline. At caudal lev- els (the A1 area, where most neurons contain TH but not PNMT immunoreactivity), catecholaminergic cells are pri- marily noradrenergic. Whether or not there is a 1:l corre- spondence between neurons staining for TH and PNMT in the rostral VLM, as observed in the rat, remains to be quantified in the human.

( 2 ) The oblique transtegmental arcs of adrenergic fibers interconnecting the rostral VLM and the NTS and also contributing to the longitudinal projections, in part, origi- nating in the rostral VLM, exist in humans in a manner identical to the distribution reported in the rat and cat (Hokfelt et al., '74, '84; Jones and Friedman, '83; Ross et al., '84b; Granata et al., '85; Ruggiero et al., '85, '86; Ben- arroch et al., '86; Kitahama et al., '86). As seen in the rat, some fibers of the transtegmental tracts which contain TH and PNMT bend perpendicularly to the transverse plane and contribute to a concentration of transversely sectioned fascicles projecting longitudinally in the dorsal medullary tegmentum. Other fibers were found to bypass these bun- dles and continue directly into the NTS (see below). Fibers of the longitudinal bundle were followed as far caudally as the spinomedullary junction and presumably continued into the spinal cord. Pearson et al. ('831, in fact, demonstrated TH-containing fibers in the human running longitudinally throughout the neuraxis including the medulla and project- ing directly to an area of the thoracic spinal cord bordering the intermediolateral cell column (ILCC) as described in the rat (Ross et al., '84a). Unfortunately, neither the work of Pearson et al. nor that of us proved that the fibers ended in the ILCC. Corroborating the findings of Kitahama et al. ('851, axons containing PNMT were concentrated in the dorsal tegmentum and formed a subset of the more exten- sive distribution of longitudinal projection fibers labeled for TH. Fibers containing TH which descend through the me- dulla may, of course, as seen in other species, also be de- rived from other rostrally located catecholaminergic nuclei such as the A5 area, locus coeruleus and the hypothalamus (Pickel et al., '74; Saper et al., '76; Hokfelt et al., '79; Loewy et al., '79). Whether PNMT (and TH-)-labeled axons contrib- ute to the descending projections to autonomic spinal neu-

Fig. 9. Photomicrographs of TH-immunoreactive perikarya in the ventral (a) and intermediate (b) nuclear subgroups of the NTS and the RVL (c). Bar = 25 pm.

CATECHOLAMINE NEURONS IN HUMAN MEDULLA 237

Fig. 10. Brightfield photomicrograph demonstrating TH-immunoreac- tive fascicles in the solitary tract, admixed with unlabeled noncatecholami.

rons, as seen in the rat, is currently being explored (in progress).

(3) Processes were traced directly into the raphe from immunoreactive perikarya in the rostral VLM. The trajec- tory of these fibers was homologous to that illustrated in the rat (Granata et al., '85; Ruggiero et al., '85). In humans, as in rats, these fibers appeared to stem, in part, from cells in the rostral VLM (C1 area) and not from the caudal VLM (A1 area). Based on the cytoarchitectonic divisions estab- lished by Olszewski and Baxter ('54), adrenergic fibers ap- peared to course through the nucleus raphe pallidus but also invaded the nucleus raphe obscurus. These findings extend those of Kitahama et al. ('85), who also identified a few PNMT-immunoreactive fibers limited to the ventral aspect of the medullary raphe. (4) Close contacts were observed between TH and PNMT

axons and brainstem intraparenchymal blood vessels. Al- though not described previously in the human, similar types of appositions with medullary vessels have been noted both in the rat muggier0 et al., '85) and cat (Ruggiero et al., '86). Electron microscopic studies in the rat (Milner et al., '87) have shown that PNMT-immunoreactive dendrites and ter- minals are found adjacent t o the basal lamina of blood vessel endothelial cells, and in some cases, are separated by a thin astrocytic process. In human material, processes containing PNMT and TH were also discovered (this study) lining, or in close proximity to, the ventral subpial surface of the rostral ventrolateral medulla. Similar fields, derived from axons and dendrites, were observed in the rat (Ruggi- ero et al., '85) and cat muggier0 et al., '86) and, as seen in the human, were traced directly from immunoreactive per- ikarya in the rostral VLM. This finding is significant and

nergic processes. Whether the labeled axons originate from peripheral afferents or intrinsic neurons remains to be established. Bar = 100 pm.

suggests that the chemoreceptor functions attributed to the ventral surface in rats and cats (for example, see Feldberg and Guertzenstein, '72, '76; Yamada et al., '82; Benarroch et al., '86) may also apply to the human. The nature of the contacts and origins of other afferents to the ventral surface are unknown.

(5) A more complex substructural organization of TH- labeled cells was also discovered in the human NTS; these cell bodies occurred in a more heterogeneous distribution than has been reported in human or nonhuman primates (Hubbard and Di Carlo, '73; Gamer and Sladek, '75; Demir- jian et al., '76; Jacobowitz and MacLean, '78; Schofield and Everitt, %l;, Feiten and Sladek, '83; Satoh and Fibiger, '85). Data from nonhuman primates, by contrast, suggest that catecholaminergic perikarya in the dorsomedial me- dulla (corresponding to the C2 region) are sparse and lie primarily external (medially and ventrally) to the NTS; some reports have also indicated monoaminergic cells as embedded within the solitary tract, situated within the motor vagal complex-intercalated nucleus or lying in the cranial aspect of the area postrema or along the floor of the fourth ventricle. In a recent study in the baboon, for exam- ple, relatively few neurons exhibiting catecholamine fluo- rescence were located in the NTS, area postrema, or surrounding the dorsomotor nucleus of the vagus (see Satoh and Fibiger, '85, Fig. 5 PL 35-39). These discrepancies may reflect species differences or differences in the sensitivity of the techniques employed.

In contrast to our findings, those of Pearson et al. ('83) suggest that the majority of TH-immunoreactive cells in humans are (as described in nonhuman primates) located

238

outside of the NTS whereas cells in the NTS were described as scarce and lacking organization (see their Fig. li-k). Extending their work, our material suggests that large numbers of cells contain TH in the NTS. They are intensely immunoreactive and are, in fact, organized into several well-defined cell groups as described in other species (Kalia et al., '85a,b), including four major clusters or arrays of cells located in the ventral (ventromedial), intermediate, medial, and dorsal (dorsolateral) divisions of the complex. The distribution of TH-immunoreactive somata in the NTS, moreover, corresponds to the distribution of cells containing neuromelanin in the human (Saper and Petito, '82). The discrete immunoreactive cell clusters in the dorsal and intermediate subnuclei of the NTS correspond, respectively, in location, size, and geometry to the dorsal and ventral sensory nuclei of the vagus as described by Olszewski and Baxter ('54). Immunoreactive cells in the ventral NTS ap- pear to surround and lie within the dorsal motor nucleus of the vagus (plates XI and XI11 of Olszewski and Baxter, '54). As observed by Saper and Petito ('82), these neurons showed greater similarities to other neuromelanin-containing (or, as observed in the present study, catecholaminergic) cell groups in the NTS than to neurons of the dorsal motor nucleus of the vagus or the intercalated nucleus. Finally, we have also confirmed the finding of Kitahama et al. ('85) that PNMT-immunoreactive perikarya have a more re- stricted localization in the dorsolateral aspect of the NTS.

In addition to immunoreactive perikarya, hitherto-unrec- ognized fields of punctate varicosities were labeled for TH

V. ARANGO ET AL. associated with the ventral medullary surface (Benarroch et al., '86). TH neurons in the human may also give rise to an extensive system of ascending fibers derived from both noradrenergic and adrenergic cells in the NTS and VLM and play a role in other autonomic and behaviorally related functions. Projections to paraventricular and supraoptic hy- pothalamic, preoptic, and "nonspecific thalamic" nuclei, if they exist in humans, may integrate autonomic with neu- roendocrine functions (for example: vasopressin secretion) and those related to the sleep-wake cycle (Marley and Ste- phenson, '72; Hokfelt et al., '74; McAllen et al., '82; Mc- Kellar and Loewy, '82; Day et al., '84; Sawchenko et al., '85; Tanaka et al., '85; Tucker et al., '87). Most of the recent studies of catecholamines in the nervous system have been performed in the monkey, cat, and rat. The results of the present investigation have revealed several important new pathways that may contribute to the anatomical circuitry of autonomic and related functions in the human, including putative interconnections between the NTS and the rostral VLM, short projections from the rostral VLM into the med- ullary raphe and onto chemosensitive areas of the ventral medullary surface, close contacts between catecholaminer- gic processes and intraparenchymal microvessels, and fi- nally, local circuits within the NTS. These homologies, if confirmed ultrastructurally, suggest that the autonomic, and chemoreceptor functions attributed to catecholamines in the VLM and the NTS may also apply to the human.

ACKNOWLEDGMENTS in two areas of the NTS: medially, in an area homologous to the medial subnucleus of Kalia et al. ('85a,b), and dorso-

caudal pole of the medial vestibular nucleus and the rostral aspect of the nucleus gracilis. The above terminal-like fields were found to extend caudally into contiguous loci in the

ergic (PNMT-positive) elements in the NTS and the nucleus

tivity determinations in the human brain using micro- punches, which took the highest activity in the c1 area to be a loo%, showed that the activity in the NTS was 5.58 pmol/hour/mg protein or 67.1%. There was no activity in the various raphe nuclei, except for the raphe pallidus, which contained 1.82 pmol/hour/mg protein, or 20.8% of PNMT activity found in the C1 area (Kopp et al., '79).

Another new &servation was the TH immunolabeling of obliquely and transversely sectioned processes within the solitary tract. This suggests that local intranuclear projec- tions may course via the solitary tract, that axons of cate- cholaminergic neurons in brain may project peripherally, exiting via the tract, or, alternatively, that primary visceral afferents may catecholamines as their neurotrans- mitters in the human.

Epinephrine and norepinephrine are thought to function as central neurotransmitters modulating the of arterial blood pressure and other autonomic and pain-re- lated functions. several studies from our laboratory and others have suggested that adrenergic neurons of the c1 area of the rostral VLM are of importance due to their interconnections with cardiopulmonary segments of the

This work was supported by PHs grants MH40210, HL18974 and "303346. by an Irma T, Hirschl Research

gatorship and Grant in Aid to D,A,R. from the American Heart Association,

we would like to thank Dr. T.A. Milner for critically

assistance, M ~ . W. ~~i~ for the photography, and D ~ ~ , H.

lecting the samples. The brains were obtained from the New York City Medical and from Dr, carol petite in the D~~~~~~~~ of pathology of N~~ york Hospital.

LITERATURE CITED

laterally, in the dorsolateral subnucleus, adjacent to the Scientist 'Award to J. J.M.; and by an Established Investi-

commissural Of the va@ls. The presence Of reading this manuscript, Ms. A. Pico~lo for her technical

Tierney and A.B. Racuma for their invaluable help in col- raphe pallidus is also supported biochemically. PNMT

Armstrong, D.M., C.A. Ross, V.M. Pickel, T.H. Joh, and D.J. Reis (1982) Distribution of dopamine, noradrenaline and adrenaline containing cell bodies in the rat medulla oblongata demonstrated by the immunocyto- chemical localization of catecholamine biosynthetic enzymes. J. Comp.

Benarrach, E.E., A.R. Granata, D.A. Ruggiero, D.H. Park, and D.J. Reis (1986) Neurons of C1 area mediate cardiovascular responses initiated from ventral medullary surface. Am. J. Physiol. 250:R932-R945.

Blessing, W.W., P. Frost, and J.P. Furness (1980) Catecholamine cell groups of the cat medulla oblongata. Brain Res. 19269-75.

Callaway, J.L., V. Arango, D.A. Ruggiero, M. Anwar, J.J. Mann, and D.J. Reis (1987) Catecholaminergic neurons in the human rostral ventrolat-

Chalmers, J.P., S.W. White, S.W. Geffen, and R. Rush (1978) The role of central catecholamine in the control of blood pressure through the baroreceptor reflex and the nasopharyngeal reflex in the rabbit. In W. De Jong et al. (eds): Hypertension and Brain Mechanisms. Progress in Brain Research, Vol. 47. Amsterdam: Elsevier, pp. 85-94.

Ciriello, J., M.M. Caverson, and D.H. Park (1986) Immunohistochemical identification of noradrenaline- and adrenaline-synthesizing neurons in

Coote, J.H., and V.H. McLeod (1974) The influence of bulbospinal monoam- inergic pathways on sympathetic nerve activity. J. Physiol. iL0nd.J

2z2:173-187.

Neurosci. Abstr. 13:277.

NTS and the intermediolateral and -medial cell columns of the spinal cord (Ross et '83, '84a, '86; Tucker et '87); the cat "entro]ateral medulla, J, Camp. Neuro]. 253;218-230. and due to their presumed Of

pressure, the baroreceptor reflex (Granata et al., '83, '85; Ross et al., '84b), and the chemosensitivity

in the tonic

2 4 1 ~ 3 - 4 7 5 .

CATECHOLAMINE NEURONS IN HUMAN MEDULLA 239

Day, T.A., A.V. Ferguson, and L.P. Renaud (1984) Facilitatory influence of noradrenergic afferents on the excitability of rat paraventricular nu- cleus neurosecretory cells. J. Physiol. (Lond.) 355:237-249.

Demirjian, C., R. Grossman, R. Meyer, and R. Katzman (1976) The catechol- amine pontine cellular groups locus coeruleus, A4, subcoeruleus in the primate Cebus upella Brain Res. 115:395411.

Di Carlo, V., J.E. Hubbard, and P. Pate (1973) Fluorescent histochemistry of monoaminecontaining cell bodies in the brain stem of the squirrel monkey (Suirniri sciureus) N. An atlas. J. Comp. Neurol. 152:347-372.

Everitt, B.J., T. Hakfelt, L. Terenius, K. Tatemoto, V. Mott, and M. Gold- stein (1983) Differential coexistence of neuropeptide Y (NPY)-like im- munoreactivity with catecholamines in the central nervous system of the rat. Neuroscience 11:443-462.

Feldberg, W., and P.G. Guertzenstein (1972) A vasopressor effect of pento- barbitone sodium. J. Physiol. (Lond.) 224:83-103.

Feldberg, W., and P.G. Guertzenstein (1976) Vasopressor effects obtained by drugs acting on the ventral surface of the brain stem. J. Physiol. (Lond.) 229~395-408.

Felten, D.L., A.M. Laties, and M.B. Carpenter (1974) Monoaminecontain- ing cell bodies in the squirrel monkey brain. Am. J. Anat. 139:153-166.

Felten D.L., and J.R. Sladek (1983) Monoamine distribution in primate brain. V. Monoaminergic nuclei: Anatomy, pathways and local organi- zation. Brain Res. Bull. 10:171-284.

Garver, D.L., and J.R. Sladek, Jr. (1975) Monoamine distribution in primate brain. I. Catecholaminecontaining perikarya in the brain stem of Mu- caca speciosa. J. Comp. Neurol. 159~289-304.

Gatti, P.J., A.M. Taveira Da Silva, P. Hamosh, and R.A. Gillis (1985) Car- diorespiratory effects produced by application of L-glutamic acid and kainic acid to the ventral surface of the cat hindbrain. Brain Res. 330~21-29.

Granata, A.R., D.A. Ruggiero, D.H. Park, T.H. Joh, and D.J. Reis (1983) Lesions of epinephrine neurons in rostral ventrolateral medulla abolish the vasodepressor components of baroreflex and cardiopulmonary reflex. Hypertension 5(SuppL v):V80-V84.

Granata, A.R., D.A. Ruggiero, D.H. Park, T.H. Joh, and D.J. Reis (1985) Brain stem area with C1 epinephrine neurons mediates baroreflex va- sodepressor responses. Am. J. Physiol. 248:H547-H567.

Hikfelt, T., K. Fuxe, M. Goldstein, and 0. Johansson (1974) Immunohisto- chemical evidence for the existence of adrenaline neurons in the rat brain. Brain Res. 66:235-251.

Hokfelt, T., 0. Phillipson, and M. Goldstein (1979) Evidence for a dopami- nergic pathway in the rat descending from the A l l cell group to the spinal cord. Acta Physiol. Scand. 107:393-395.

Hokfelt, T., 0. Johansson, and M. Goldstein (1984) Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bjorklund and T.Hokfelt (eds): Handbook of Chemical Neuroanatomy. Vol. 2: Classical Transmitters in the CNS, Part I. Amsterdam: Elsevier, pp. 157-276.

Hornykiewicz, 0. (1966) Metabolism of brain dopamine in human parkin- sonism: Neurochemical and clinical aspects. In E. Costa et al. (eds): Biochemistry and Pharmacology of the Basal Ganglia. New York: Raven Press, pp. 171-185.

Hubbard, J.E., and V. Di Carlo (1973) Fluorescent histochemistry of mono- aminecontaining cell bodies in the brain stem of the squirrel monkey (Suirniri sciureus). I. The locus coeruleus. J. Comp. Neurol. 147:553- 565.

Jacobowitz, D.M., and P.D. MacLean (1978) A brainstem atlas of catechol- aminergic neurons and serotonergic perikarya in a pygmy primate (Cebuella pygrnaeu). J. Comp. Neurol. 177:397-416.

Jensen, T.S., and Y.L. Yaksh (1984) Spinal monoamine systems partly me- diate the antinociceptive effects produced by glutamate at brainstem sites. Brain Res. 321:287-297.

Joh, T.H., and M. Goldstein (1973) Isolation and characterization of multiple forms of phenylethanolamine~N-methyltransferase. Mol. Pharmacol. 9t117-129.

Joh, T.H., and M.E. Ross (1983) Preparation of catecholamine synthesizing enzymes as immunogen for immunocytochemistry. In A.C. Cuello (ed): Immunocytochemistry. mRO Handbook Series: Methods in the Neuro- sciences, Vol. 3. Oxford: John Wiley and Sons, Ltd., pp. 121-138.

Jones, B.E., and L. Friedman (1983) Atlas of catecholamine perikarya, varicosities and pathways in the brain stem of the cat. J. Comp. Neurol. 215~382-396.

Jones, S.L., and G.F. Gebhart (1986) Characterization of coeruleospinal inhibition of the nociceptive tail flick reflex in the rat: Mediation by spinal a2-adrenoceptors. Brain Res. 364t315-320.

Kalia, M., K. Fuxe, and M. Goldstein (1985a) Rat medulla oblongata. 11. Dopaminergic, noradrenergic (A1 and A2) and adrenergic neurons, nerve fibers and presumptive terminal processes. J. Comp. Neurol. 233~308- 332.

Kalia, M., K. Fuxe, and M. Goldstein (1985b) Rat medulla oblongata. 111. Adrenergic neurons, nerve fibers and preterminal processes. J. Comp. Neurol. 233:333-349.

Kitahama, K., L. Denoroy, A. Berod, and M. Jouvet (1986) Distribution of PNMT-immunoreactive neurons in the cat medulla oblongata. Brain Res. Bull. 17:197-208.

Kitahama, K., J. Pearson, L. Denoroy, N. Kopp, J. Ulrich, T. Maeda, and M. Jouvet (1985) Adrenergic neurons in human brain demonstrated by immunohistochemistry with antibodies to phenylethanolamine-N-meth- yltransferase (PNMT): Discovery of a new group in the nucleus tractus solitarius. Neurosci. Lett. 53~303-308.

Kopp, N., L. Denoroy, B. Renaud, J.F. Pujol, A. Tabib, and M. Tommasi (1979) Distribution of adrenaline-synthesizing enzyme activity in the human brain. J. Neurol. Sci. 41~397409.

Lew, J.Y., Y. Matsumoto, J. Pearson, M. Goldstein, T. Hokfelt, and K. Fuxe (1977) Localization and characterization of phenylethanolamine N- methyltransferase in the brain of various mammalian species. Brain Res. 119:199-210.

h w y A.D. S. McKellar, and C.B. Saper (1979) Direct projection from the A5 catecholamine cell group to the intermediolateral cell column. Brain Res. 174:309-314.

Marley, E., and J.D. Stephenson (1972) Central actions of catecholamines. In H. Blaschko and E. Muscholl (eds): Catecholamines. Berlin: Springer, pp. 463-537.

McAllen, R.M., J.J. Neil, and A.D. Loewy (1982) Effects of kainic acid applied to the ventral surface of the medulla oblongata on vasomotor tone, the baroreceptor reflex and hypothalamic autonomic responses. Brain Res. 238:65-76.

McKellar, S., and A.D. Loewy (1982) Efferent projections of the A1 catechol- amine cell group in the rat An autoradiographic study. Brain Res. 241 :11-29.

Mefford, I., A. Oke, R.N. Adams, and G. Jonsson (1977) Epinephrine local- ization in human brain stem. Neurosci. Lett. 5:141 -145.

Milner, T.A., V.M. Pickel, D.H. Park, T.H. Joh, and D.J. Reis (1987) Phenyl- ethanolamine N-methyltransferasecontaining neurons in the rostral ventrolateral medulla of the rat: I. Normal ultrastructure. Brain Res. 411:28-45.

Nobin, A., and A. Bjorklund (1973) Topography of the monoamine neuron systems in the human brain as revealed in fetuses. Acta Physiol. Scand. 88 (Supp. 388) 2:1-40.

Olszewski, J., and D. Baxter (1954) Cytoarchitecture of the Human Brain Stem. Basel: S. Karger A.G.

Park, D.H., E.E. Baetge, B.B. Kaplan, V.R. Albert, D.J. Reis, and T.H. Joh (1982) Different forms of adrenal phenylethanolamine-N-methyltrans- ferase: Species-specific posttranslational modification. J. Neurochem. 38t410-414.

Pearson, J., M. Goldstein, K. Markey, and L. Brandeis (1983) Human brain- stem catecholamine neuronal anatomy as indicated by immunocyto- chemistry with antibodies to tyrosine hydroxylase. Neuroscience 8:3- 32.

Pickel, V.M., M. Segal, and F.E. Bloom (1974) A radioautographic study of the efferent pathway of the nucleus locus coeruleus. J. Comp. Neurol. 155t15-42.

Pickel, V.M. (1981) Immunochemical methods. In L. Heimer and M.J. Ro- bards (eds): Neuroanatomical Tract-Tracing Methods. New York Plenum Press, pp. 483-509.

Poitras, D., and A. Parent (1978) Atlas of the distribution of monoamine containing cell bodies in the brain stem of the cat. J. Comp. Neurol. 179~699-718.

Reiner, P.B., and S.R. Vincent (1988) The distribution of tyrosine hydroxyl- ase, dopamine-6-hydroxylase, and phenylethanolamine-N-methyltrans- ferase immunoreadive neurons in the feline medulla oblongata. J. Comp. Neurol. 248518-531.

Reis, D.J. (1985) Brain-stem mechanisms governing resting and reflex tone of precapillary vessels. J. Cardiovasc. Pharmacol. 7(Suppl. 3):S160-S166.

Ross, C.A., D.A. Ruggiero, T.H. Joh, D.H. Park, and D.J. Reis(1983) Adren- aline synthesizing neurons in the rostral ventrolateral medulla: A pos- sible role in tonic vasomotor control. Brain Res. 273:358-381.

Ross, C.A., D.A. Ruggiero, T.H. Joh, D.H. Park, and D.J. Reis (1984a) Rostra1 ventrolateral medulla: Selective projections to the thoracic auto- nomic cell column from the region containing C1 adrenaline neurons. J.

240 V. ARANGO ET AL.

Comp. Neurol. 228r168-185. Ross, C.A., D.A. Ruggiero, D.H. Park, T.H. Joh, A.F. Sved, J. Fernandez-

Pardal, J.M. Saavedra, and D.J. Reis (1984b) Tonic vasomotor control by the rostral ventrolateral medulla: Effect of electrical or chemical stim- ulation of the area containing C1 adrenaline neurons on arterial p res sure, heart rate, and plasma catecholamines and vasopressin. J. Neurosci. 4~474-494.

Ross, C.A., D.A. Ruggiero, and D.J. Reis (1986) Projections from the nucleus tractus solitarii to the rostral ventrolateral medulla. J. Comp. Neurol. 242:511-534.

Rubinstein, L., and C.H. Sawyer (1970) Role of catecholamines in stimulat- ing the release of pituitary ovulating hormone(s) in rats. Endocrinology 86:988-995.

Ruggiero, D.A., C.A. Ross, M. Anwar, D.H. Park, T.H. Joh, and D.J. Reis (1985) Distribution of neurons containing phenylethanolamine N-meth- yltransferase in medulla and hypothalamus of rat. J. Comp. Neurol. 239:127-154.

Ruggiero, D.A., P.J. Gatti, R.A. Gillis, W.P. Norman, M. Anwar, and D.J. Reis (1986) Adrenaline-synthesizing neurons in the medulla of the cat. J. Comp. Neurol. 252532442.

Saper, C.B., A.D. Loewy, L.W. Swanson, and M. Cowan (1976) Direct hy- pothalamo-autonomic connections. Brain Res. II7:305-312.

Saper, C.B., and C. Petito (1982) Correspondence of melanin-pigmented neurons in human brain with A1-A14 catecholamine cell groups. Brain 105:87-101.

Satoh, K., and H.C. Fibiger (1985) Distribution of central cholinergic neu- rons in the baboon (Papiopapio). 11. A topographic atlas correlated with catecholamine neurons. J. Comp. Neurol. 236:215-233.

Sawchenko, P.E., L.W. Swanson, R. Grzanna, P.R.C. Howe, S.R. Bloom, and J.M. Polak (1985) Co-localization of neuropeptide Y immunoreactivity in brainstem catecholaminergic neurons that project to the paraventric- ular nucleus of the hypothalamus. J. Comp. Neurol. 241r138-153.

Schildkraut, J.J. (1965) The catecholamine hypothesis of affective disorders: A review of supporting evidence. Am. J. Psychiatry 122~509-522.

Schofield, S.P.M., and B.J. Everitt (1981) The organization of catecholamine- containing neurons in the brain of the rhesus monkey (Macaca muluttu). J. Anat. I32:391-418.

Spencer, S., C.B. Saper, T. Joh, D.J. Reis, M. Goldstein, and J.D. Raese (1985) Distribution of catecholamine-containing neurons in the normal human hypothalamus. Brain Res. 328r73-80.

Sternberger, L.A. (1979) Immunocytochemistry. 2nd ed. New York John Wiley and Sons.

Tanaka, C., M. Ishikawa, and S. Shimada (1982) Histochemical mapping of catecholaminergic neurons and their ascending fiber pathways in the rhesus monkey brain. Brain Res. Bull. 9:255-270.

Tanaka, J., H. Kaba, H. Saito, and K. Seto (1985) Inputs from the A1 noradrenergic region to the hypothalamic paraventricular neurons in the rat. Brain Res. 335:368-371.

Tucker, D.C., C.B. Saper, D.A. Ruggiero, and D.J. Reis (1987) Organization of central adrenergic pathways: I. Relationships of ventrolateral med- ullary projections to the hypothalamus and spinal cord. J. Comp. Neu- rol. 259r591-603.

Yamada, K.A., W.P. Norman, P. Hamosh, and R.A. Gillis (1982) Medullary ventral surface GABA receptors affect respiratory and cardiovascular function. Brain Res. 248:71-78.