The relationship of efferent projections from the area postrema to vagal motor and brain stem catecholamine-containing cell groups: An axonal transport and immunohistochemical study in the rat

Cepyri~t Q 1994I3W Printed inGreat Britain.Ali rightsreserved 030~5~/94$6.~+0.~

THE RELATIONSHIP OF EFFERENT PR~~E~I~NS FROM THE AREA PUSTREMA TO VAGAL MOTOR AND BRAIN STEM CATECHOLAM~NE-CO~AINING CELL GROUPS: AN AXONAL TRANSPORT AND IMMUNOHISTOCHEMICAL STUDY IN THE RAT E. T. CUNNIKGHAM JR,*?

R. R. MEELI@ and P. E. SAWCHENKO@

*Laboratory of Neuronal Structure and Function, The Salk Institute for Biological Studies, and The Foundation for Medical Research La Jolla, CA 92037, U.S.A. $Departmcnt of Animal Biology, University of Pennsylvania Veterinary School, Philadelphia, PA 19104,U.S.A. A~~a~t-T~~ ma postrema has been irnp~~~d as a major station for the processing of visceral sensory information, involved primarily in eliciting rapid homeostatic reponses to fluid and nutrient imbalances. Yet the precise relationship of efferent projections from the area postrema to medullary motor and relay nuclei involved in such functions remains unclear. In this study, axonal transport and immuaohistochem~ ical techniques were used to investigate the relationship of efferent projections from the area postrema to vagal motor neurons and medullary catecholamine-containing cell groups in the rat. The results may be summarized as follows: (1) The area postrema’gives rise to dense inputs to the commissural and medial parts of the nucleus of the solitary tract. Many of these projections are intimately associated with ~tcchoi~ine-~ntaini~~ neurons in the A2 and C2 cell groups, including a particufarly prominent input to a caudaily placed cluster of adrenergic neurons (the C2d cell group) in the dorsai aspect of the medial part of the nucieus of the solitary tract. (2) The area postrema provides a dense input to the external lateral part of the parabrachial nucleus. (3) The area postr~a does not project si~~~ntly to vagal motor neurons in eitherthe dorsal motor nucleus or the nucleus ambiguus, ~though the ~ssibility for inputs to distal dendrites of dorsal vagal motor neurons cannot be excluded. (4) En route to the parab~c~al nucleus, axons of areapostrema neurons travere the regions of the AI, Cf and AS cell groups, although these fibers make few arborizations, suggesting little functional contact. Together, these results suggest that sensory jnfo~ation received by the area postrema is dispatched to a restricted set of neurons in the commissural, medial, and dorsal parts of the nucleus of the solitary tract, most probably including catecholamine-containing cells in the A2, C2, and C2d cell groups, and to the external lateral portion of the parabrachial nucleus. The targets of area postrema projections are, in turn, in a position to effect adaptive changes in the activities of hypothalamic neurosecretory neurons, vagal motor neurons, and limbic forebrain regions in response to perturbations in fluid and nutrient homeostasis.

The area postrema (AP) has been implicated in a number of behaviors important to maintaining fluid and nutrient homcostssis,1a,s3,~incl~~ng emesis,‘414 conditioned taste aversion Iearning,4~30*@ the regulation of food and water intake,24s59s62 blood pressure “9” 29,43~73 heart rate,3,2’,2Band the secretion of hydothalamic hormones, most notably arginine vasopressin33s37,76 and oxytocin. *s*33However, projections from the AP to autonomic pregan~~onjc neurons appear sparse, and direct inputs to the ne~o~reto~ hypothalamus have not been deTPresent address: Department of ~phthalmolo~ K301, School of pricing, UCSF, IO Kirkham Street. San Francisco~ CA 94143, U.&i. $To whomcorrqmnd~n~ should be addressed at: The Salk Institute, P.O. Box 85800, San Diego, CA 92186, U.S.A. A&evi&ons: AP, ama postrema; CT-NRP, cholera toxinhorseradish peroxidase; NTS, nucleus of the solitary tract; PB, parabrachial nucleus; PHA-L, Phusecdus vtlfgaris-leucoagglutinin; TH, tyrosine hydroxylase.

scribed.69 It would appear, therefore, that the behav-

ioral effects of AP activation are largely indirect, presumably mediated by autonomic and neuroendocrine relay nuclei in the nearby nucleus of the solitary tract (NTS) and ventrolaterai medulla. The most likely candidates in this regard are the medullary and pontine catecholaminergic cell groups, including the noradrenergic Al, A2, A5 and A6 (locus coeruleus) &I groups and the adrenergic Cl, C2, and C3 ceil groups, since each of these gives rise to substantial and topo~ph~lly distinctive inputs to the hypothalamic neurosecreto~ cell group~‘~~*‘*~’ and/or to s~pathetic p~g~~onic ne~ons in the thoracolumbar spinal ~0n.i.~’None~eless, studies of the dis~bu~on of AP efferents have failed to achieve a consensus as to how widespread or restricted these may be, and the relationship of AP projections to vagaf motor neurons and to brainstem catechotamine-containing cell groups remains to be systematically explored.

63.5

636

E. T. CUNNINGHAM cr ~1.

In this study, iontophoretic application of the high-resolution axonally transported tracers cholera toxin-horseradish peroxidase (CT-HRP) and Phaseolus vulgarj.~-leuco~gg~utinin (PHA-L) was used to chart the distribution of efferent projections from the AP in the rat. The PHA-L anterograde tracing technique was also combined with retrograde labeling of vagal motor neurons and immunohisto~hemical labeling of medullary cat~holamine-containing cells with an antiserum against the catechofaminesynthesizing enzyme tyrosine hydroxylase (TH) to determine the relationship of AP efferents to vagal motor and catecholamine-containing relay nuclei in the cauda! medulla. EXPERKMENTAL

PROCEDURES

AI1 experiments were carried out with adult albino rats of the Sprague-Dawley strain (Harlan, Indianapolis, IN) in accordance with the Declaration of Helsinki and the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health. Animals were housed in a light- and temperature-controlled environment with food and water freely available. All surgeries were performed under deep methoxyflurane anesthesia. Tracer deposits into the AP were made with a stereotaxic frame under direct visualir;ltion of the dorsal medulla following reflection of the atlanto-occipital membrane. Phaseolus vulgaris-leucau~gltinin

~ter~gr~de

tracing

Iontophoretic deposits of a 2.5% solution of PHA-L (Vector Laboratories) in 0.1 M sodium phosphate buffer, pH 7.4, were made inta the center of the AP. Deposits were delivered over 5 -1Omin through IO-am (id.) glass micropipettes, using a constant-current device (Model &S-3: Transkinetics, Inc.) set to deliver +3-/rA p&es, with %s pufse and inter-pulse durations. After survival times of five to seven days, animais were an~thetized and perfused for immunohist~~emistr~ as described previously.3’ Cholera t#~~n-~lorser~d~~~peroxidase ~nt~r~gr~de tritci~g The CT-HRP used in these experiments was conjugated and diluted in 0.1 M sodium phosphate-buffered saline, pH 7.4, to a final concentration of 0.2-0.346 as previously described.‘,69 ~o~tophoretic deposits of CT-HRP were made into the center of the AP as described above for PHA-L. After survival times of two to three days, animals were perfused and processed for HRP histochemistry according to the tetramethyl benzidine protocol of Mesulam.‘O

indirect immunohistochemical techniques were used to localize the catecholamine-synthesizing enzyme TH and the anterograde tracer PHA-L. The rabbit anti-PHA-L (Vector Laboratories, 1:2000) and rabbit anti-TH (Pel-Freeze Laboratories. t : IOW)have each been c~racterized previously in similariy processed tissue.s2 Avidin-biotin immunoperoxidase localization of rabbit-derived primary antisera was pa-formed with a biotinylated goat anti-rabbit IgG and an avidinbiotin complex (Vector Laboratories). as described previously.” Immunofluore~nce focalization of rabbit-derived primary antisera was performed with either a rhodamine- or fluorescein-conjugated goat anti-rabbit 1gG (Tago, Inc., 1:200), using a protocol described eisewhere.“,“’ C~~bj~e~ Phaseoius v~lga:aris-l~e~~g~uli~~j3~ anterogrude trueing-immunohisto~mi~tr~

Ioniophoretic deposits of a 2.5% solution of PHA-L were aimed at the center of the AP and the tissue was processed

as described previously. ” Most animals that received a deposit of PHA-L into the AP, also received a I-1.11 injection of a 5.0% suspension of True Blue (Dr filing GmbH & Co. KG) in distilled water into the sheath surrounding the cervical vagus nerve. True Blue deposits were made at, or slightly distal to, the levbl of the nodose ganglion, and gave rise to extensive retrograde labeling of ceIIs within the dorsal motor nucleus of the vagus nerve and the nucleus arnbjguus.~?.~~ In order to investigate the relationship of antero~radely transports PRA-L to TH-containing catecholaminergic celfs and True Blue-labeled v&gal motor neurons, rabbit anti-TH [I : 101X3) and guinea-pig anti-PHA-L (I: 1500) were applied concurrently to complete series of sections through the medulla. The guinea-pig anti-PHA-L used in these experiments was raised in our laboratory, and has been characterized in previous studies.2z”XA mixture of rhndam~ne-~njuga~ed goat anti-rabbit IgG (see above) and fluorescein-conjugated goat anti-guinea-pig IgG (Cooper Biomedical, West Chester, PA, I :200) was used to localize the primary antisera. Control procedures to assure that the secondary antisera did nat cross-react with an inappropriate primary antiserum or with each other were negative.” KtESULTS

Phaseolus vulgaris-ieuc.oag~lutinin

~nt~r~grade

trues-

ing

A total of 18 animals received deposits of WA-L that were centered in the AP. In some of these cases a small amount of peroxidase reaction product extended into the subjacent NTS, but in no instance were neurons labeled outside of the AP (Fig. 1). The projection pattern for each of these experiments was virtually identical (Fig. 2). Heavy inputs were directed from the AP to the subjacent caudaf NTS, most prominently its commissurai, medial, and dorsal su~ivisions (Fig. 20-c). Sparse label was also observed as far caudal as the spinomeduliar~ junction, and along the medial edge of the NTS to its rostralmost extension (Fig. 2D, E). A particular?y dense 500-700~pm column of labeled fibers was observed in the dorsal subdivision of the NTS, mid-way between the solitary tract and the AP {Figs 1, 2F). Few fibers were found in the iateral division of the NTS or in the dorsal motor nucleus of the vagus nerve (Fig. 2M; see below). In general, fibers within the NTS arborized extensively, suggesting functional interaction with neuronal somata and/or dendrites in this region. Area postrema efferents exited the NTS approximately 500 pm rostra1 to the caudalmost aspect of the AP (Fig. 2F), traversed the medu~lary reticular formation obliquely, and turned rostrally just dorsal to the region of the lateral reticular nucleus, a region known to contain the noradrenergic Al and adrenergic Cl cell groups (Fig. D-F, see below). Fibers then ascended through the medulla ventrolateral to both the nucleus ambiguus and the facial motor nucleus (Fig. 2C, D), turned dorsally at the level of the superior oiivary complex, in proximity to the region of the noradrena~ine-synthesizing AS ceil group {Fig. 2B), and ended in a dense terminal arbori~ation in the external lateral su~ivisi~n of the PB (Pig. 2A;

Area postrema efferents

637

Abbreviations used in figures AP

Bar bc bp cc cl CG CU dc DCN DMX dsct DTN EC Gr gVII IO LC LDT LRN LV KF MeV mlf MoV MV NA,, NA,, NA,, NA,,

area postrema Barrington’s nucleus brachium conjunctivum brachium pontis central canal compact formation central gray cuneate nucleus dorsal column nucleus dorsal cochlear nucleus dorsal motor nucleus of the vagus nerve dorsal spin~e~bellar tract dorsal tegmental nucleus external cuneate nucleus gracile nucleus genu of the seventh nerve inferior olive locus coeruleus lateral dorsal tegmental nucleus lateral reticular nucleus lateral vestibular nucleus Kiilliker-Fuse nucleus mesencephalic nucleus of the trigeminal nerve medial longitudinal fasciculus motor nucleus of the trigeminal nerve medial vestibular nucleus nucleus ambiguus, compact formation nucleus ambiguus, external forma~on nucleus ambiguus, loose formation nucleus ambiguus, semicompact formation

NTB NTS NTS, NT%,, NTS, NTS, NT% PB,, PB, PB, PH PHA-L PY Rob rb RPa SNV sot stv sv tb TB-X FRN VI VII VIIIn VLL XII

nucleus of the trapezoid body nucleus of the solitary tract nucleus of the solitary tract, central subnucleus nucleus of the solitary tract, commissural subnucleus nucleus of the solitary tract, dorsal subnucleus nucleus of the solitary tract, medial subnucleus nucleus of the solitary tract, ventrolateral subnucleus parabrachial nucleus, external lateral subnucleus parabrachial nucleus, lateral subdivision parabrachial nucleus, medial subdivision nucleus preposnus hypoglossi Phase&s Bulgaria-leucoa~lutinin py~midal tract nucleus raphe obscurus restiform body nucleus raphc pallidus spinal nucleus of the trigeminal nucleus superior olivary complex spinal tract of the trigeminal nucleus superior vestibular nucleus trapezoid body True Blue in the cervical vagus solitary tract tegmental reticular nucleus abducens nucleus facial nucleus eighth nerve ventral nucleus of the laterat lemniscus hypoglossal nucleus

Fig. 1. B~~t~eld photomi~ographs of adjacent coronal sections through the dorsal vagal complex prepared with avidin-biotin-immunoperoxidase (A) and thionin (B) to show the placement of a representative PHA-L tracer deposit centered in the AP. Anterogradely labeled fibers and terminals are evident within the medial part of the NTS (NTS,,,),particularly in its most dorsal aspect (arrow), which we term the dorsal NTS (NTS,).

Fig. 2. A series of schematic line-drawings through the caudal brainstem to show the distribution of anterogradely labeled fibers and terminals following a discrete iontophoretic deposit of PHA-L centered in the AP (F). Note the dense terminal fields in the medial and commissural parts of the NTS (D-G), and in the external lateral portion of the ~rabrachial nucleus (PB,: section A). Nate also that fibers leaving the dorsal vagal complex traverse the caudai medufla at and slightly rostra1 to the level of the AP (E, F), ascend through the ventrolateral medulla and pons (B-D), and turn dorsally to innervate the external lateral portion of the parabrachial nucleus just rostra1 to the level of the facial nucleus (VII; B, C).

639

Area postremaefferents

Fig. 3. Low-powerfluorescence(A) and darEeM (A’)photomicrograp~sof the same sectionto showthe distribution of MA-L labeled fibers and terminalsin ttreexternallateral part of the parochial nucleus (P&J folfoting an i&j&ion of PEW-Lcentered in the area postrema.Only a few scatteredfibers were observedrostra1to this level in the lateral part of the parabrachial nucleus. Fig. 3). Only isolated fibers were observed in other portions of the PB. No fibers were observed in the A6 cell group, in the paramedian aspects of the medulla occupied by the C3 adrenergic cell group, or in any structure rostral to the PB. With the exception of the terminal ar~rjzatio~ observed in the external lateral part of the PB, as~nding projections from the AP were poorly arborized, suggesting little functional contact with neurons in the ventrolaterai medulla and pons (see below). A total of six animals received deposits of PHA-L that were centered in the AP but which labeled cells in the subjacent NTS. In addition to the abovedescribed projections from the AP proper, each of these deposits labeled densely ramifying terminal fields in the regions of the Al and Cl catecholaminecontaining cell groups, and in several aspects of both the medial and lateral parts of the PB (not shown). In general, this latter pattern of brainstem projections was similar to that reported previously foIlo~ng tracer deposits in the caudomedial and commissural parts of the NTS.2a*22,63

A total of five animals received deposits of CTHRP that were centered in and confined to the AP.

Individually labeled cells at the injection site were difficult to discern with this technique because of the diffuseness and intensity of the peroxidase reaction product. Three animals received deposits of CT-HRP that were centered in the AP but which showed reaction product exten~ng into the subjacent media1 and commissural portions of the NTS. In all respects the pattern and density of anterograde labeling observed with the CT-HRP deposits was identical to that described above for PHA-L (not shown), with those deposits involving the NTS again labeling more widespread projections. The relationship motor neurons

of area

postrema eferents

to vagal

Twelve of the 18 animais described above with PHA-I.. deposits centered in the AP also received deposits of the retrograde tracer True Blue into the sheath surrounding the vagus nerve.” These deposits resulted in retrograde labeling of neurons in all su~ivisions of the dorsal motor nucleus’ and the ambiguual complex.8 Sections were then processed for PHA-L immuno~uores~n~ to allow simultaneous localization of the anterograde tracer and retrogradely labeled vagal motor neurons (Figs 4, 5). Few anterogradely labeled fibers were seen near neurons within the dorsal motor nucleus (Fig. 4),

640

E. T.

CUNNINGHAM er ai.

Fig. 4. Pairs of fluorescence photomicrographs of the same sections through the dorsal saga1 complex approximately 0.5 mm rostra1 to (A,A’), at (B,B’), and OS mm caudal to (CC’) the level of the AP to show the distribution of fibers and terminals anterogradely labeled with PHA-L (left) in the medial (NTS,) and ~ommissural (NTS,,) parts of the NTS following an injection of PHA-L centered in the area postrema. Sparse labeling is present around neurons in the dorsal motor nucleus of the vagus nerve (DMX), labeled here with True Blue following application of the tracer around the cervical portion of the vagus nerve (TB-X, rig&). although contact with unlabeled dorsal vagal motor neuron dendrites extending into the NTS is possible. although

the possibility for contact with distal den-

drites of vagal motor neurons that extend into the adjacent NTS could not be excluded. Anterogradely labeled fibers were observed passing through the caudal portion of the nucleus ambigu~ en route to

the ascending path of AP axons near the ventrolateral surface of the medulla, but few arborized. and virtually none were found at more rostra1 levels of the

complex (Fig. 5), thus providing little basis for interaction of AP projections with ventral vagal motor neurons.

ambiguual

The re~utjo~shi~ of area postrema e$erents to brain

sle~n ~atechol~mi~e-~o~tuini~g ceil groups Five of the 18 animals described above with PHA-L deposits centered in the AP were processed

Area postrema efferents

641

Fig. 5. A seriesof fluorescencepbotomicrographsof the same sectionsthrough the rostra1(A,A’),mid-level

tB,B’I,and caudal Kc’) aspects of the ventrolatcral medulla to show the ~aucitvof PRA-L ileft) labeled projections in the cornpact-( semicompact (scf), external (ef), and loose fdrmations (lf) within the nucleus ambiguussfoilowing an injection of PHA-L,centered in the area postrema. Arn~~~~ motor neurons are labeled here with True Blue foLIowingapplication of the tracer around the cervicalporrion of the vagus nerve FRX. right). Straight, non~arborizingPHA-L labeled fibers are seen traversing the caudal ventrolateral medulla (arrowheads: C,C’).

for dual immuno~uor~s~n~ l~ali~ation of anterograd&y ~ansport~ PHA-L and the catechoolaminesynthesizing enzyme TH. Within the NTS, dense terminal ar~~~t~ons were present around THimmunoreactive ceil bodies and dendrites in regions of both caudalty situated A2 noradrenergic and the rostromedially disposed C2 adrenergic neurons (Figs 6,7). As mentioned above, a particularly dense

te~inal field was found in the dorsal subdivision of the PITS, mid-way betw~n the AP and the solitary tract, an area shown by dual staining ta contain a cluster of TV-immunore~ctive neurons (Fig. 6). These To-immunoreactive cells have been identified by HBkfelt and colleagues as adrenergic,” and have been termed the C2d, or dorsal strip, cell group.36*4’@ These C2d neurons tend to be smaller and with

642

E. T. CUNNDJGHAMet d.

Fig. 6. Two fluorescence photornicrographs of the same section through the dorsal aspect of the medial part of the NTS (NTS,) to show the distribution of PHA-L-labeled terminals (A, arrow) in relation to known adrenergic neurons in the C2d cell group (A’, arrow) in this region following an injection of PHA-L centered in the AP. Catecholamine-containing neurons are labeled using an antiserum directed against TH. Note the correspondence between the location of the PHA-L-labeled terminal field in the dorsal NTS and the C2d cell cluster. shorter dendrites than the more rostrally-situated C2 cells, and may therefore be distinguished on morphological grounds, at least. Numerous PHA-L-immunoreactive fibers were seen in the vicinity of TH-containing neurons in the Al and Cl catecholaminergic cell groups in the ventrolateral medulla (Figs 8, 9). However, these fibers invariably were poorly arborized, and contained few terminal bouton-like specializations. While fibers ascended through the substance of the Al cell group (Fig. 8) they maintained a distinctly lateral position relative to the major clusters of Cl neurons (Fig. 9) and on past the seventh nerve nucleus and A5 region en route to the PB (Fig. 2A-C). Fibers in the region of the A5 cell group also made few arborizations (not shown). No anterogradely labeled fibers were found in the immediate vicinity of either the A6 or C3 cell groups. DISCUSSION

The results suggest that the targets of AP projections are limited to a few circumscribed regions. Our observations are consistent with previous descrip-

tions of dense projections from the AP to the caudal NTS, most notably its commissural, medial, and dorsal subdivisions.39,69~75,77,78In addition, we have shown that AP inputs to the caudal NTS are closely associated with catecholamine-containing neurons of the A2 and C2 cell groups, with a particularly dense input directed toward known C2d adrenergic neurons in the dorsal NTS. This study also supports previous reports of a dense projection from the AP to the parabrachial nucleus (PB), ,6.34,35,44.45.47 49.51.69.7?,75 specifically to its external lateral subdivision.34.“5,h” On the other hand, we adduced little evidence to support functional interactions between AP efferents and either vagal motor neurons or catecholaminecontaining neurons in the Al, Cl, or A5 cell groups of the ventrolateral medulla or pans.‘,“’ Rather, efferent projections from the AP, when present, appeared to traverse these regions making few arborizations, more typical of fibers of passage. In addition, we found no evidence to support previous descriptions of light AP projections to the superior or inferior colliculi,77,78 central gray,69,77.78dorsal tegmental nuclei75,77.78mesencephalic, spinal or para-trigeminal nuclei,69.77.78 hypoglossal or perihypoglossal

postrema efferents

suggestedby Herbert ef ai,35 since OeIis Iabeied by our deposits were never su~~en~ly &cumscribed to enable such a distinction. In fact, neurons in all parts of the AP were labeled with virtually every PHA-L deposit, and the CT-HRP deposits tended to fill the entire AF. Slightly larger deposits extending beyond the borders of the AP frequently labeled neurons in the nearby NTS, dorsal column nuclei, caudal vestibular nuclei, and adjacent reticular formation, perhaps accounting for previous reports of more widespread AP terminal fields. No inputs were observed to either the A6t6~69~77~78 or C3 cell groups. The residence of neurons within the AP, as well as the existence of a dense fiber bundle linking the AP with the caudal NTS, has been bogus since the original apphcation of silver-based fiber staining techniques to this region over 40 years ag~.~‘t*‘~*~~~” However, recognition that the fiber bundle between the AP and the NTS represented postremofugai, or efferent, fibers awaited the advent of tract tracing techniques based on axonal transport.39*54,55~69J5~77,78 With the exception of the study by van der Kooy and Koday’ which failed to detect fibers in the area of the NTS immediately adjacent to the AP, the pattern of anterograde label obsetved in the NTS following AP injections was ~rna~ab~y comparabie for ait of these ex~~rnen~s~ and parallels the ~st~bution observed in the present study. In each case, heaviest projections were observed in the commissural, medial, and dorsal subdivisions of the NTS at the level of the AP, regions known to receive dense vagaf tierents, most prominently from the cardiovascular’2,?6~sS and gastrointestina11.~58,6’systems. Shapiro and Miselis69 were the first to call attention to the particularly prominent input from the AP to the dorsal aspect of the medial NTS, a region we refer to as the dorsal NTS, and which Kalia and Sulkvat? termed the dorsolateral NTS. In addition, sparse projections were observed as far caudal as the spinom~ulla~ junction, and along the medial most edge of the NTS to the rostra1 Iimit of the nucleus, We have shown here that tfie inputs from the AP to the caudal NTS are largely co-~st~but~ with cat~holamiue-containing cells in the A2 and C2 cell groups. This includes a particularly prominent input to C2d neurons in the dorsal NTS. As mentioned above, these C2d neurons tend to be smaller and ovoid, in contrast to the larger, multipolar C2 neurons found at more rostra1 levels.36”64 Recently, Kachidian and Picke139examined the relationship of AP efferents to TV-immunoreactive neurons within the caudal NTS at the u~tras~u~tura~ levei, and provided evidence for direct monosynaptic inputs. These inputs were found to arise almost entirely from non~at~holamjner~~ neurons within the AP, and formed mainly symmetric, presumed inhibitory, synaptic contacts with TH-positive neurons in the A2 and C2 cell groups. It should be mentioned, however, that many NTS neurons extend dendrites well into the parenchyma of the AP,57and that the study by recently

Fig. 7. Three high-power fluorescence photomicrographs of the same section through the commissural part of the NTS (NTS,,) to show the distribution of PHA-L labeled terminals (A) in relation to known noradrenergic neurons in the A2 cell group (A’) and neurons within the dorsal motor nucleus of the vagus nerve (DMX, A”) in this region following an injection of PHA-L centered in the AP. ~at~hoIarnin~on~~ng neurons are labagedhere with an indirect ~unohi~t~he~~ technique using an antiserum directed against the ~t~ho~~~ne s~~~i~~ enzyme TH. Neurons in the DMX are labeled with True Blue following application of the tracer around the cervical portion of the vagus nerve. Note the preferential association of AP projections with catecholamine-containing A2 cells. nuclei,“*‘* or the cerehellum.69 We are unable to comment on diffe~ntia~ pattern of projections to PB from the “core” verses the “rim” of the AP as

643

E. T. CUNNINGHAM et al

Fig. 8. Fluorescence photomicrographs of the same sections through the caudal (A,A’) and rostra1 (B&I’) ventrolateral medulla to show the distribution of PHA-L-labeled fibers (left) in relation to known catecholamine-containing neurons (right) of the AI and Cl cell groups (B’) following an injection of PHA-L centered in the AP. Note the paucity of terminal arborizations made by PHA-L-labeled fibers traversing these regions.

Kachidian and Pickelj’ also demonstrated a large number of AP inputs to non-TH-immunoreactive cell bodies and dendrites in the caudal NTS, thus providing ample opportunity for AP efferents to directly contact non-catecholaminergic NTS neurons and neuronal processes. Efferent projections from the AP to the external lateral subdivision of the PB have been reported previously, although these studies described additional projections to nearby subdivisions within the PB.34x35,69 As mentioned, we too observed more widespread projections to the lateral PB whenever our injection sites extended beyond the border of the AP to involve portions of the subjacent NTS (see above). Previous reports have shown that only a small percentage of AP neurons labeled following an injection of retrograde tracer into the lateral PB co-localize TH.45.5’ Interestingly, Herbert and Saper34 recently demonstrated that more than one-third of the cholecystokinin-immunoreactive neurons within the AP project to the external lateral part of the PB. There were relatively few cholecystokinin-immunoreactive cells within the AP, however, such that 90% of neurons labeled in the AP following an injection of

retrograde tracer into the PB were not cholecystokinin-immunoreactive. Others have also reported that small subsets of PB-projecting AP neurons contain immunoreactive serotonin4’ or angiotensin II.M However, the neurochemical signature for the majority of PB inputs from the AP remains unknown. Very few fibers labeled following AP injections were found within the cytoarchitectonic boundaries of the dorsal motor nucleus, although these neurons do have extensive dendrites which penetrate well into the caudal NTS.‘.‘2 Fibers were observed traversing the region of the caudal NA, as well as the area of the Al, Cl and A5 cell groups. However, these were invariably few in number and minimally arborized, more suggestive of fibers of passage. The possibility of en passant contacts with vagal motor neurons and catecholamine-containing cell groups in these regions cannot, however, be excluded. The observed pattern of projections from the AP to the NTS and the PB fits well with the concept of the AP as a monitor of fluid and nutrient homeostasis,‘x,s’,54 a role supported by studies implicating the AP in the control of emesis,‘0,‘4 conditioned taste aversion learning,4.30,60ingestive behaviors,24,5Y.62 heart rate.3,27.28and the blood pressure.‘7x2729.43.73

Area postrema efferents

Fig. 9, A low-power fluorescencephoto~crog~~~ of ~~ho~amjne~nt~ini~g neurons in the Cf cett group iccatedin the rostra1ventrolateral meduiki. A schematic iine-drakng of this region of the brainstem is shown for reference (A’). Two higher-power photomicrographs (boxes)show the preferentialcourse of ascending PHA-L, (B’,C’)-labeledfibers through the lateral (right), as opposed to the medial (left), aspect of the Cl cell group following an injection of PHA-L centered in the area postrema. Catecholaminecontaining neurons are labeled using an antiserum directed against TH (B,C). Note the paucity of terminal arborizations made by PHA-L-labeled fibers traversing this regions.

release of the ~euro~y~ph~seai hormones vasopressin3337*76 and ~xytociin.~~~~~ Its poorly developed blood-brain barrier,” afong with the presence of trigeminal, ~ossopbarynge~, and vagal aRerents,” provide the AP with an array of possible interoceptive inputs. The AP must then influence autonomic preganglionic and hypothalamic neurosecretory neurons to effect behavioral changes. While direct inputs from the AP to autonomic and ~ypoth~amic effector neurons appear sparse at best, the adjacent caudal NTS does provide subs~nt~al projections

to these cell groups.‘263The noradrenergic A2 and rostra1 adrenergic C2 cells, for exampie, densely innervate the neurosecretory h~othaiamus, including regions known to contain co~~~tropin-releasing factor- and vasop~ssio-p~ducing cell~,‘~,~‘“’and/or sympathetic preganglionic neurons in the thoracic spinal cord.6S Non-catecholaminergic projections arising from the caudal NTS also innervate the neurosecretory hypothalamus, including a dense, p~marily jnhibin~activin-containing input to oxy tocinergic rna~n~ll~ar neurons,~,s’

Another major target of AP projections, the C2d cell group in the dorsal NTS, appears to project to the PB, with a particularly dense input to the external lateral subdivision.42 Interestingly, Miselis and colleagues? have shown that cells in the dorsal NTS directly innervate cardiac preganglionic neurons in the ventrolateral medulla, although it remains to be determined whether the cells of origin of this pathway are catecholaminergic. Other aspects of the caudal NTS abo project heavily to the region of cardiac pre~nglioni~ neurons, and to the area containing AI and Cl cells in the ventrolateral meduIla.20~22~63 The primary target of the Al cell group is magnocellular neurosecretory vasopressin-containing neurons, while projections from the Cl cell group are directed toward both the hypothalamus2’ and sympathetic preganghonic neurons in the spinal cord.65 In addition, gastric and ceeal motor neurons in the dorsal motor nucleus project dendrites into the caudal NTS, well into the terminal field of the AP.‘B~’ Thus, numerous mono-, di-, and oligosynaptic pathways exist through which the AP might inlkence cardiovascular, gastrointestinal, and neumendocrine reffexes. Interestin~y, recent studies have provided evidence that g~trointesti~l fluid absorption is dy namically regulated in response to vagal afferent activity,70,7’,79providing at least one possible adaptive advantage to having cardiovascular, gastrointestinal, and neuroendocrine reflexes coordinated from a common pivotahy placed center, such as the caudal NTS. The external lateral su~ivision of the PB projects primarily to the limbic forebrain, including a dense input to the central nucleus of the amygdala516 and a

moderate projection to the perinuclear region of the hypothalamic supraoptic nucleus, a pure magnocellular neurosecretory cell gro~p.~~ Physiological studies have suggested that the pathway to the amygdala may convey nociceptive5 and/or peripheral metabolic21*25information, while the input to the perinuclear region of the supraoptic nucleus may indirectly effect vasopressin release in response to hypovolemia and/or hypotension.38 This latter pathway may therefore compliment the above described indirect input from the AP to ma~~llular vasop~ssin-producing cells by way of the Al cell group. CONCLUSlON

In summary, the present results indicate that AP projections are directed primarily to discrete aspects of the NTS and PB nuclear wmplexes. Within the NTS, targets of AP axonal projections appear to include, but are not restricted to, catecholaminergic neurons of the A2, C2 and C2d cell groups. These findings serve to focus attention on a limited number of simple circuits through which documented AP influences on the neuroendocrine, cardiovascular and gastrointestinal systems may be brought to bear. Acknowledgemenrs-This work was supported by NIH grant HL-35137 (PES), and was conducted in part by the Foundation for Medical Research. P.E.S. is an Investigator of the Foundation for Medical Research. We thank Belle Wamsley, Kris Trulock and Carlos Arias for excellent editorial, photographic and technical assistance, respectively.

R~ERENCES 1. Altschuler S. M., Ferenci D. A., Lynn R. B. and Misehs R. R. (1991) Representation of the eecum in lateral dorsal motor nucleus of the vaguenerve and commissural subnucleus of the nucleus of the tractus solitarii in the rat. J. camp. Neural. 283, 248-268. 2. Barnes K. L. and Ferrario C. M. (1981) Anatomical and physiological characterization of the sympatho-facilitative area postrema pathways in the dog. In Central Nervous System Mechanisms in Hypertension (eds Buckley J. P. and Ferrario C. M.), pp. 25-35. Raven Press., New York. Barnes K. L., Ferrario C. M. and Conomy J. P. (1979) Comparison of the h~~ynamic changes produced by electrical stimulation of the area postrema and nucleus tractus solitarii in the dog. ~irc~lurjon Res. 45, 136143. Berger B. D., Wise C. D. and Stein L, (1973) Area postrema damage and bait shyness. J. camp. Physiot. Psycho!. 82, 415479.

7. 8. 9. 10. 11.

Bernard J.-F., Carroue J. and Besson J.-M. (1991) Efferent projections from the external parabrachial area to the forebrain: a phaseolus vulgaris leucoagglutinin study in the rat. Neurosci. Lect. 122, 257-260. Bernard J.-F., Alden M. and Besson J.-M. (1993) The organization of the efferent projections from the pontine parabrachial area to the amygdaloid complex: A phaseolus uutgaris leucoagglutinin (PHA-L) study in the rat. J. camp. Neurof. 329, 201-229. Berthoud H.-R. and Powiey T. L. (1992) Vagal afferent innervation of the rat fimdic stomach: morphological characterization of the gastric tension receptor. J. camp. Neural. 319, 261-276. Bieger D. and Hopkins D. A. (1987) Viscerotopic representation of the upper alimentary tract in the medulla oblongata in the rat: the nucleus ambiguus. J. camp. Neurol. 262, 546-562. Blessing W. W., Hedger S. C., Joh T. H. and Willoughby J. 0. (1987) Neurons in the area postrema are the only catecholamine-synthesizing cells in the medulla or pons with projections to the rostra1 ventrolateral medulla (Cl-area) in the rabbit. Brain Res. 419, 336340. Borison H. L. (1989) Area postrema: chemoreceptor cir~umvent~~ular organ of the medulla oblongata. Prog. Neu~obio~. 32, 351-390. Borison H. L. and Brizzee K. R. (1951) Morphology of emetic chemoreceptor trigger zone in cat medulla obiongata. Proc. Sot. exp. Biol. Med. 77, 38-42.

12. Bradd J., Dubin J., Due B., Miselis R. R., Montor S., Rogers W. T., Spyer K. M. and Schwaber J. S. (1989) Mapping of carotid sinus inputs and vagal cardiac outputs in the rat. Sot. Neurosci. Abstr. 15, 593.

Area postrema efferents

641

13. Brizzee K. R. and Neal L. M, (1954)A re-evaluation of the cellular mo~hology of the area postrema in view of recent evidence for a chemo-receptor function. J. cotnp. Neural. 100, 41-61. 14. Carpenter D. 0. (1990) Neural mechanisms of emesis. Can. J. Physiol. Pharmac. 68, 230-236. 15. Carter D. A. and Lightman S. L. (1987)A role for the area postrema in mediating cholecystokinin-stimulated oxytocin secretion. Brain Res. 435, 327-330. 16. Cedarbaum J. M. and Aghajanian G. K. (1978) Afferent projections to the rat locus coeruleus determined by a retrograde tracing technique. J. camp. Neural. 178, 1-16. 17. Chan R. K. W, and Sawchenko P. E. (1993) Spatially and temporally differentiated patterns of c-fos expression in brains&em~t~ol~iner~c cell ~oups’indu~~by cardiovascular challenges in the rat. J. camp. Neuro!. (submitted). 18. Conmeras R. J., Kosten T. and Bird E. (1984) Area postrema: part of the autonomic circuitry of caloric homeostasis. Fe& Proc. 43, 2966-2968.

19. Cunningham E. T. Jr and Sawchenko P. E. (1988)Anatomical specificity of noradrenergic inputs to the paraventrictdar and supraoptic nuclei of the rat h~othalamus. J. corn@Nearoi. 274, 6076. 20. Cunningham E. T. Jr. and Sawchenko P. E. (1989) A circumscribed projection from the nucleus of the solitary tract to the nucleus ambiguus in the rat: anatomical evidence for somatostatin-28-immunoreactive interneurons subserving reflex control of esophageal motility. J. Neurosci. 9, 1668-1682. 21. Cunningham E. T. Jr, Bohn M. C. and Sawchenko P. E. (1990)Organization of adrenergic inputs to the paraventricular and supraoptic nuclei of the hypothalamus in the rat. J. camp. Neural. 292, 651667. 22. Cunningham E. T. Jr., Simmons D. M., Swanson L. W. and Sawchenko P. E. (1991) Enkephalin immunoreactivity and messenger RNA in a discrete projection from the nucleus of the solitary tract to the nucleus ambiguus in the rat. J. camp. Neural. 307, l-16. 23. Edwards G. L. and Johnson A. K. (1991) Enhanced drinking after excitotoxic lesions of the parabrachial nucleus in the rat, Am. J. Physiol. 261, R1039-R1044. 24. Edwards G. L. and Ritter R. C. (1981) Ablation of the area postrema causes exaggerated ~nsumption of preferred foods in the rat. &uin Res. 211, 265-273. 25. Edwards G. L and Ritter R. C. (1989) Lateral parabrachial lesions attenuate ingestive effects of area postrema lesions. Am. J. Physiol. 256, R306-R312. 26. Escardo J. A., Schwaber J. S., Paton J. F. R. and Miselis R. R. (1991) Rostra-cau~ topography of cardiac vagal innervation in the rat. Neurosci. Abstr. 17, 993. 27. Ferguson A. V. and Smith P. (1991)Autonomic mechanisms underlying area postrema stimulation-induced cardiovascular responses in rats. Am. J. Physioi. 261, Rl-R8. 28. Ferguson A. V. and Marcus P. (1988) Area postrema stimulation induced cardiovascular changes in the rat. Am. J. Physiol. 255, RS%R86O. 29. Ferrario C. M., Barnes K. L., Diz D. I., Block C. H. and Averill D. B. (1987)Role of area postrama pressor mechanisms in the regulation of arterial pressure. Can. J. Physiot. Pharmac. 65, 1591-1597. 30. Gallo M., Arnedo M., Aguero A. and Puerto A. (1990)The functional relevance of the area postrema in drug-induced aversion learning. Pharmac. Biochem. Behau. 35, 543-M. 31. Gerfen C. R. and Sawchenko P. E. (1984) An anterograde neuroanatomical tracing method that shows the detailed morphology of neurons, their axons and terminals: Immuno~stochemical localization of an anterogradely transported plant &tin, Phuseoius u~~g~r~-leu~agglutin~ (PHA-L). Brain ffes. 290, 219-238. 32. Gerfen C. R. and Sawchenko P. E. (1985)A method for auterograde tracing of chemically specified circuits in the central nervous system: combined Phaseolus vulgaris-leucoagglutinin tracing and immunohistochemistry. Brain Res. 343, I44150. 33. Hawthorn J., Andrews P. L. R., Ang V. T. Y. and Jenkins J. S. (1988) Differential release of vasopressin and oxytocin in response to abdominal vagal aRerent stimulation or apomorphine in the ferret. Brain Res. 438, 193-198. 34. Herbert H. and Saper C. B. (1990) Cholecystokinin-, galanin-, and corticotropin-releasing factor-like immunoreactive projections from the nucleus of the solitary tract to the parabrachial nucleus in the rat. J. camp. Neural. 293,581-598. 35. Herbert H., Moga M. M. and Saper C. B. (1990)Connections of the parabrachial nucleus with the nucleus of the solitary tract and the medullary reticular formation in the rat. J. camp. Neurol. 293, 540-580. 36. Hiikfelt T., Johansson 0. and Goldstein M. (1984)Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In Hark of Ckem~caZ N~ro~ut~y, Vol. 2: C~~sical Transmitters in the CNS, Part 1. (eds Bjijrklund A. and Hokfelt T.), pp. 157-276. Elsevier, Amsterdam. 37. Iovino M., Papa M., Monteleone P. and Steardo L. (1988) Neuroanatomical and biochemical evidence for the involvement of the area postrema in the regulation of vasopressin release in rats. Brain Res. 447, 178-182. 38, Jh~~das I. G., Harris K. H. and Krukoff T. L. (1991) Pa~~achial nucleus projection towards the h~~alamic supraoptic nucleus: electrophysiological and anatomical observations in the rat. J. camp. Neural. 3Og, 42-50. 39. Kachidian P. and Pickel V. M. (1993)Localization of tyrosine hydroxylase in neuronal targets and efferents of the area postrema in the nucleus tractus solitarii of the rat. J. camp. Neuroi. 329, 337-353. 40. Kalia M. and Sullivan J. M. (1982) Brain stem projections of sensory and motor ~mponents of the vagus nerve in the rat. J. camp. Neural. 211, 248-264. 41. Kalia M., Fuxe K. and Goldstein M. (1985) Rat medulla oblongata. II. Dopaminergic, noradrenergic (Al and A2) and adrenergic neurons, nerve fibers, and presumptive terminal processes. J. camp. Neurof. 233, 308-332. 42. Kawai Y., Takagi H., Yanai K. and Tohyama M. (1988) Adrenergic projection from the caudal part of the nucleus of the tractus solitarius to the parabrachial nucleus in the rat: immunocytcehemieal study combined with a retrograde tracing method. Brain Res. 459, 369-372. 43. Kawano Y., Barnes K. L. and Ferrario C. M. (1991)Pressor effect of centrally administered sodium chloride: role of the ventral third ventricle region and the area postrema. J. uuton. nerv. SYS. 35, 153-160. 44. KdcdTT. L., VU T., Harris K. H., Aippersbach S. and Jhamandas J. H. (1992)Neurons in the rat medulla oblongata containing neuropeptide Y-, angiotensin II-, or galanin-like immunoreactivity project to the parabrachial nucleus. Neuroscience 47, 175-184. 45. Lanca A. J. and van der Kooy D. (1985) A seroto~n~on~ining pathway from the area postrema to the parabmchial nucleus in the rat. Neuroscience 14, I117-1126. 46. Leslie R. A. and Gwyn D. G. (1984) Neuronal connections of the area postrema. Fe&r. Proc. 43, 2941-2943.

648

E. T. CUNNINGHAM et rrl

47 Loewy A. D. and Burton H. (1978) Nuclei of the solitary tract: etTerent projections to the lower brain stem and spinal cord of the cat. J. camp. Neural. 181, 421450.

48 McLean J. H. and Hopkins D. A. (1982) Ultrastructural identification of labeled neurons in the dorsal motor nucleus of the vagus nerve following injections of horseradish peroxidase into the vagus nerve and brainstem. J. camp. Neural. 206, 243-252. 49. Mehler W. R. (1983) Observations on the connectivity of the parvicellular reticular formation with respect to a vomiting center. Brain Behau. Evol. 23, 63-80. 50. Me&am M.-M. (1982) Principles of horseradish peroxidase neurochemistry and their applications for tracing neural pathways-axonal transport, enzyme histochemistry and light microscopic analysis. In Tracing Neural Connections wirh horseradish Ferox~~e (ed. Mesulam M.-M.), pp. l-t 51. John Wiley, New York. 51. Milner T. A., Joh T. H. and Pickel V. M. (1986) Tyrosine hydroxyiase in the rat parabrachia~ region: UItrastru~tural localization and extrinsic sources of immunoreacti~ty, J. Neurosci. 6, 2585-2603. 52. Miselis R. R. and Shapiro R. E. (1983) Dorsal motor nucleus neurons have extensive dendrites penetrating the nucleus of the solitary tract. Fedn. Proc. 42, f125. 53. Miselis R. R., Hyde T. M. and Shapiro R. E. (1984) Area postrema and adjacent solitary nucleus in water and energy balance. Fedn. Proc. 43, 2969-2971. 54. Miselis R. R., Shapiro R. E. and Hyde T. M. (1986) The area postrema. In Circumvenfricuk~r Organs and Body Fluids. Vol. II (ed. Gross P. M.), pp. 185-207. CRC Press, Florida. 55. Miselis R. R., Rogers W. T., Schwaber J. S. and Spyer K. M. (1989) Localization of cardiomotor neurones in the anaesthetized rat; cholera toxin HRP conjugate and pseudorabies labelling. J. Physiol., Land. 416, 63P. 56. Morest D. K. (1967) Experimental study of the projections of the nucleus of the tractus solitarius and the area postrema in the cat. J. camp. Neural. 130, 277-300. 57. Morest D. K. (1960) A study of the structure of the area postrema with Golgi methods. Am. J. Anaf. 107, 291~-303. vagal branches in the rat. J. camp. Neural. 58. Norgren R. and Smith G. P. (1988) Central distribution of su~iaphra~ati~ 273, 2077223. 59. Olson B. R., Freilino M., Hoffman G. E., Stricker E. M., Sved A. F. and Verbalis J. G. (1993) c-Fos expression in rat brain and brainstem nuciei in response to treatments that alter food intake and gastric motility. MO&C.cell. Neur~),~~i. 4 93-106. 60. Ossenkopp K. P., Sutherland C. and Ladowsky R. L. (1986) Motor activity changes and conditioned taste aversion induced by administration of scopolamine in rats: role of the area postrema. Pharmac. Biochem. Behav. 25, 269~-276. 61. Plotsky P. M., Cunningham E. T. Jr and Widmaier E. P. (1989) Catecholaminergic modulation of corticotropinreleasing factor and adrenocorticotropin secretion. Endocrine Rev. 10, 437-458. 62. Ritter R. C. and Edwards G. L. (1986) Dorsomedial hindbrain participation in control of food intake. ln Feeding Behavior: Neural wd Humoral Confrols (eds Ritter R. C., Ritter S. and Barnes C. D.). pp. 131-161. Academic Press, Florida. 63. Ross C. A., Ruggiero D. A. and Reis D. J. (1985) Projections from the nucleus tractus solitarii to the rostra1 ventrolateral medulla. J. camp. Neurol. 242, 51 l-534. 64. Ruggiero D. A., Ross C. A., Anwar M., Park D. H., Joh T. H. and Reis D. J. (1985) Distribution of neurons containing PhenylethanoIamine N-methyltransferase in medulla and hypothalamus of rat. J. camp. Neural. 239, 1277154. 65. Sawchenko P. E. and Bohn M. C. (1989) Glucocorticoid receptor imm~orea~tivity in Cl, C2 and C3 adrenergic neurons that project to the hypothalamus or to the spinal cord in the rat. J. camp. Neural. 285, 107- 116. 66. Sawchenko P. E., Arias C. and Bittencourt J. C. (1990) Inhibin b-, somatostatin-, and enkephalin-~mmunoreactivities coexist in caudal medullary neurons that project to the paraventricular nucleus of the hypothalamus. J. camp. Neural. 291, 269-280. 67. Sawchenko P. E., Plotsky P. M., Pfeiffer S. W., Cunningham E. T. Jr, Vaughan J., Rivier J. and Vale W. (1988) Inhibin fl in central neural pathways involved in the control of oxytocin secretion. Nature 334, 615-617. 68. Shapiro R. E. and Miselis R. R. (1985) The central organization of the vagus nerve innervating the stomach of the rat. J. camp. Neural. 238, 473488. 69. Shapiro R. E. and Miselis R. R. (1985) The central neural connections of the area postrema of the rat. J. camp. Neural. 234, 344-364. 70. Smi C.-K. and Baertschi A. J. (1991) Splanchnic osmosensation and vasopressin: mechanisms and neural pathways. Am. J. Physioi. 261, E18-E25. 71. Smi C.-K., McCarty R. and Baertschi A. J. (1990) Splanchnic control of vasopressin secretion in conscious rats. Am. J. Physiol. 259, E19E26. 72. Strain S. M., Gwyn D. G., Rutherford J. G. and Losier B. J. (1990) Direct vagal input to neurons in the area postrema which project to the p~abrachial nucleus: an electron micro~opic-HRP study in the cat. Brain Res. BuII. t4,457-463. 73. Sun K.-M. and Spyer K. M. (1991) GABA-m~iated inhibition of medullary vasomotor neurones by area postrema stimuiation in rats. J. Physiol., Land. 436, 669-684. 74. Torack R. M. and Finke E. H. (1971) Evidence for a sequestration of function within the area postrema based on scanning electron microscopy and the penetration of horseradish peroxidase. 2. Ze&%rsch. 118, 85-96. 75. van der Kooy D., Koda L. Y. (1983) Organization of the projections of a circumventricutar organ: the area postrema in the rat. J. camp. Neurol. 219, 328-338. 76. Verbalis J. G., McHale C. M., Gardiner T. W. and Stricker E. M. (1986) Oxytocin and vasopressin secretion in response to stimuli producing learned taste aversions in rats. Behau. Neurosci. 100, 466-475. 71. Vigier D., RouviereA. (1979) Afferent and efferent connections of the area postrema demonstrated by the horseradish peroxidase method. Archs. ital. Biol. 177, 325--339. 78 Viaier D.. Portaher P. (1979) Efferent projections of the area postrema demonstrated by autoradiography. A&s. i&i. _ _ Bill. 177,‘308-324. . 79 Zhang X., Fogei R. and Renehan W. E. (1992) Physiology and morphology of neurons in the dorsal motor nucleus of the vazus and the nucleus of the solitary tract that are sensitive to distension of the small intestine. J. camp. Neurol. 323,432-448. (Accepted 13 September 1993)