Journal of the Autonomic Nervous System, 30 (1990) 209-220 Elsevier

JANS 01067

209

Quantitative analysis of spinally projecting adrenaline-synthesising neurons of Cl, C2 and C3 groups in rat medulla oblongata

Jane Minson \ Ida Llewellyn-Smith \ Angela Neville \ Peter Somogyi 2 and John Chalmers 1

I Department of Medicine and Centre for Neuroscience, .School of Medicine, Flinders University of South Australia, Bedford Park, South Australia, Australia, and 2 MRC Anatomical Neuropharmacology Unit, South Parks Road, Oxford, u.K.

(Received 28 January 1990) (Revision received 20 March 1990)

(Accepted 26 March 1990)

Key words: PNMT; Adrenaline-containing neurons; Bulbospinal pathways; Retrograde tracing; Medulla

Abstract

Spinal projections of the phenylethanolamine N-methyltransferase (PNMT) immunoreactive neurons of the medulla were investigated using a combination of immunohistochemistry and retrograde transport of colloidal gold particles conjugated to cholera toxin B subunit (CTB-gold). The PNMT-containing adrenergic neurons were localised in three groups, the Cl group in the rostral ventrolateral medulla, the C2 group in the nucleus tractus solitarius/ dorsal vagal motor complex in the dorsal medulla and the C3 group in the mediodorsal medulla. The Cl group contained 72% of the medullary PNMT-IR neurons, while C2 comprised 13% and C3 15% of the medullary PNMT-IR neuron population. CTB-gold was injected in the area of the intermediolateral cell column in either upper (T2-T4) or lower (T8-T9) thoracic spinal cord and retrogradely labelled cells were found in the areas of the Cl , C2 and C3 groups and in other regions of the medulla which did not contain PNMT-IR neurons. After tracer injections bilaterally at levels T2- T4, retrograde labelling suggested that at least 21 % of all medullary PNMT-IR neurons projected to these levels. As a proportion of each group, 26% of Cl, 9% of C2 and 33% of C3 neurons projected spin ally to T2-T4. After tracer injections bilaterally at levels T8-T9, retrograde labelling suggested that at least 17% of all medullary PNMT-IR neurons projected to these levels. As a proportion of each group, 16% of Cl, 9% of C2 and 30% of C3 neurons projected spin ally to T8-T9. These figures represent minimum numbers since it is impossible to ensure that every neuron has equal access to the tracer. The results demonstrate that contrary to previous belief, the PNMT-IR innervation of the spinal cord derives from PNMT-IR neurons in the dorsal medulla, as well as from the rostral ventrolateral medulla. Indeed 24% of the PNMT-IR neurons terminating at spinal cord levels T2~ T4, and 35 % of those terminating at levels T8-T9, derive from the dorsal (C2 and C3) medullary cell groups. Since the PNMT-IR projections are directed towards the intermediolateral cell column, it seems likely that all three groups of medullary adrenaline-containing neurons contribute to the regulation of sympathetic outflow.

Introduction

The enzyme phenylethanolamine N-methyl­transferase (PNMT) converts noradrenaline to adrenaline and is considered to be a marker for

Correspondence: l.P. Chalmers, Department of Medicine, Flinders Medical Centre, Bedford Park, South Australia 5042, Australia.

neurons that use adrenaline as a transmitter. PNMT-immunoreactive (PNMT-IR) neurons have been demonstrated in brain of several species, including rat [5,29,32], cat [21], monkey [12] and man [23,36].

I t is in the rat that the PNMT -IR neuronal system has been studied most extensively. Adren­aline cell groups in the brainstem have been loca­lised in three areas, the Cl, C2 and C3 groups [30,32]. Although other neurons displaying

0l65-1838/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

210

PNMT-IR have been localised in the hypothala­mus, these neurons do not contain the other en­zymes of catecholamine synthesis [19,51] and are not generally believed to contain adrenaline. The rostral ventrolateral medulla (RVLM), the area of the Cl adrenaline neurons, plays an important role in maintaining vasomotor tone and mediating vasomotor reflexes, probably via the direct pro­jections of the RVLM neurons to the intermedio­lateral cell column of the thoracic spinal cord. Although the Cl PNMT-IR neurons of the RVLM project to this area of spinal cord [49,50], the vasomotor functions of the RVLM cannot be at­tributed solely to these PMNT-IR neurons, since a variety of other putative neurotransmitters have also been localised in spinally projecting neurons of the RVLM. These include neuropeptide Y (NPY; [10]), serotonin, substance P and thyrotro­pin-releasing factor, in some neurons localised to­gether [33], enkephalin [31,41] and the enzyme of GABA synthesis, glutamic acid decarboxylase [42]. Studies have supported the hypothesis that the Cl adrenaline neurons of the RVLM play a major role in the regulation of sympathetic tone and blood pressure. Stimulation of the Cl area of the RVLM increases sympathetic activity and blood pressure [16,21,40,44,48,50,52,61], and electro­physiological studies have suggested that a popu­lation of spinally projecting, barosensitive neurons in RVLM [11,40,58] are likely to contain PNMT [26,45].

The first observations in the rat suggested that only the adrenaline-containing neurons of the RVLM Cl group projected to the spinal cord [49]. This was supported by experiments in cat [21] and monkey [12]. More recently some PNMT-IR neu­rons of the dorsal medulla have been observed to be labelled retrogradely from the spinal cord [25,39,46,54], but a detailed study of this dorsal bulbospinal projection has not been made. Since these dorsal medullary PNMT -IR neurons, to­gether with the ventrally located PNMT-IR neu­rons, might contribute to the regulation of sym­pathetic activity and blood pressure via a direct spinal projection, it is important that their loca­tion be precisely defined as an aid to further investigation. It was the aim therefore of the pre­sent study to map and quantitate the adrenaline-

containing neurons that project to the spinal cord from both the dorsal and the ventral medulla. We have used cholera toxin B sub unit conjugated to 7-nm colloidal gold particles (CTB-gold) to retro­gradely label spinally projecting neurons in the medulla in rats. In combination with PNMT im­munohistochemistry we have demonstrated that PNMT-containing neurons belonging to all three medullary adrenaline groups, that is Cl, C2 and C3, project to the intermediolateral cell column area of the thoracic spinal cord.

Materials and Methods

Surgical procedures Rats (Wistar Kyoto, male, 250-350 g) were

anaesthetised with a sodium pentobarbital (N em­butal, 30 mg/kg)/ chloral hydrate (100 mg/kg) mix i.p. After femoral artery cannulation (polyeth­ylene SP10 tubing), rats were intubated, and placed in a stereotaxic holder (David Kopf). Short-lasting ganglionic blockade was induced by suxa­methonium chloride (4 mg/kg i.a., Anectine, Wellcome) and rats were artificially ventilated (Harvard 608 ventilator). By dorsal incision the spinal column was exposed in either upper thoracic (T2-T4) or lower thoracic (T8-T9) levels. Partial laminectomy exposed the dorsal surface of the spinal cord. A glass pipette (Clay Adams), outside tip diameter 30-50 pm, was placed in the spinal cord to lie in the area of the intermediolateral cell column (0.6 mm lateral to midline, 0.7-0.9 mm below dorsal surface). The pipette was filled with cholera toxin B sub unit conjugated colloidal gold (CTB-gold [38]; now commercially available from Gilt Products, Flinders University Department of Medicine, Bedford Park, Australia). GTB-gold (250 nl) was injected over 10 min (25-nl injections each 60 s). The pipette was withdrawn and placed further along the cord using the same coordinates. In total 1 pI of CTB-gold was injected into each rat, either bilaterally (2 X 250 nl each side of spi­nal cord) or unilaterally (4 X 250 nl on one side of spinal cord). The CTB-gold injections were placed over a 2-mm section of spinal cord. After CTB-gold injection the exposed cord was covered with Gelfoam (Upjohn), the overlying muscle and skin was sutured, and the arterial cannula was

withdrawn. With resumption of spontaneous res­piration rats were extubated. 6-10 days after CTB-gold injections rats were anaesthetised (hexobarbitone sodium, Brietal, 40 mg/kg i.p.) and injected with colchicine (100 Mg in 10 MI 0.9% saline intracisternally). At 48 h after colchicine rats were deeply anaesthetised with sodium pentobarbital (60 mg/kg) and perfused transcar­dially with 150 ml 0.1 M phosphate buffered saline pH 7.2, followed by 1 1 fixative, formaldehyde (4%), glutaraldehyde (0.05%) and picric acid (about 0.2%) in phosphate buffer 0.1 M, pH 7.2 [56]. Brain and spinal cord were post-fixed at 4 0 C overnight.

Immunohistochemical procedures Transverse 50-Mm Vibratome (Oxford Instru­

ments) sections, cut perpendicular to the ventral surface of the brainstem, were taken from the level of the pyramidal decussation caudally to the level of the trapezoid body rostrally. Every fourth sec­tion was silver intensified to visualize the retro­gradely transported colloidal gold, borohydride treated and immunohistochemically processed to reveal PNMT-like immunoreactivity. Spinal cord segments were sectioned and silver intensified to localise injection sites.

Briefly, sections were washed 3 x 15 min in 0.1 M phosphate buffer pH 7.2. After 6 X 10 min in 0.1 M sodium citrate/ ammonium acetate buffer pH 5.5, sections were incubated in 2 X 20 min changes of IntenSE BL (Janssen). After washing in citrate-acetate buffer the silver processing was stabilised by 5 min incubation in 1 % sodium thio­sulphate. Sections were rinsed in citrate-acetate buffer, washed 3 X 10 min in 0.1 M phosphate buffer and incubated in 0.5% sodium borohydride in phosphate buffer for 30 min before being rinsed again in phosphate buffer. For immunohistoche­mistry sections were incubated in 10% normal sheep serum (NSS) in 10 mM Tris phosphate buffered saline pH 7.4 with 0.05% merthiolate (TPBS), containing 0.3% Triton X-lOO for 30 min. Sections were incubated in rabbit anti-PNMT serum (now commercially available from Incstar) diluted 1: 15000 in TPBS-Triton containing 10% NSS for 48 h. Sections were washed 3 X 20 min in 1 % NSS in TPBS-Triton, then incubated in bio-

211

tinylated sheep anti-rabbit immunoglobulin (Sig­ma) diluted 1: 200 in this buffer containing 1% NSS for 24 h. Sections were washed 3 X 20 min in TPBS-Triton and incubated in 1: 1000 Avidin­HRP (Sigma) in TPBS-Triton for 4 h. Sections were rinsed 3 X 10 min in Tris 0.05 M pH 7.6 and immunoreactive cells and fibres were visualised by a peroxidase reaction. Sections were preincubated in 0.5 mg/ml diaminobenzidine in Tris 0.05 M pH 7.6, 0.04% NH4CI and 0.2% D-glucose for 10 min and the reaction was started by adding 1 Ml of glucose oxidase (Type V-S in 0.1 M sodium acetate buffer, Sigma) per 1 ml of reaction mixture. Reac­tions were for 20 min, sections were then finally washed 3 X 10 min in phosphate buffer 0.1 M. Sections were mounted on chrome alum slides, dehydrated and mounted in Depex. Using this procedure there was complete antibody penetra­tion and immunoreactivity was observed throughout the 50-llm thick sections. Preabsorp­tion of this antiserum with purified bovine PNMT [2], or the substitution of normal for immune serum, resulted in the complete absence of specific neuronal staining in the medulla.

In the reacted sections all PNMT -IR neurons and all PNMT-IR neurons containing retrogradely transported gold (PNMT-IR/CTB-gold-Iabelled) were counted on both sides of the medulla. After unilateral tracer injection cell counts were re­corded ipsi- and contralateral to the spinal cord injection site. Cell counts were not adjusted to whole brainstem numbers. The rostro-caudal loca­tions of PNMT-IR neurons are given in reference to the obex, defined as the rostral extent of the area postrema in the midline. Cell counts were made independently by two observers, and the counts varied by less than 2%. Statistical analysis was by chi-square test using 2 X 2 tables, or by Fisher's exact test when one of the groups for analysis contained less than five neurons. Results are given as means ± SEM.

Results

Injection sites

CTB-gold injection sites were. localised in the white matter adjacent to the lateral horn spread-

212

90 CTB-goId in T2-T 4 CTB-goId in T8-T9

C1 80

70

60

50

40-

30

c 20 0

~ I

!!!. 10., (J)

~ 0 0

a: T I-~ Z a...

3°1 C2 C2 r~t-l-l 20 I ' .....

: '.,-...-l-__i 101 I ,

01 j ~

40J r1\ C3 30 / \~

::1 PNMT-IR

~ \ --.-- I \ I \

-0--

PNMT -IR/CTB-goid~. ~ /~/ \

...... l-

-to OBEX +1.0 +2.0 -3.0

Rostro-caudal axis (mm)

Fig. 1. Distribution of all PNMT-IR neurons and PNMT-IR neurons retrogradely labelled with CTB-gold after bilateral CTB-gold injection into the thoracic spinal cord at levels T2-T4 (n 4, left) or T8-T9 (n 4, right). Results are expressed as counts of cells from every fourth 50-J.Lm thick section through the medulla. The distribution of cells rostrocaudally from the obex is given in mm.

The distribution of cells in the groups Cl, C2 and C3 is described in the text.

ing into the lateral funiculus. Injections sites were less than 0.5 mm in diameter.

PNMT-IR cell profiles

PNMT -IR neurons, with amber-stained cyto­plasm and showing considerable dendritic detail, were observed over a 3-mm extent of medulla. The rostro-caudal distribution of the PNMT-IR neu­rons is shown in Fig. 1. Medullary PNMT -IR neurons in the examined series of sections (every fourth 50 J-Lm thick section) from each rat num­bered 959 ± 44 (n = 15). These PNMT -IR cells were localised in three cell groups, the ventro­lateral Cl group 686 ± 31 cells, the dorsal C2 group 130 ± 10 cells and the dorsomedial C3 group

143 ± 8 cells; respectively 72%, 13% and 15% of the total PNMT -IR neuron population.

The PNMT-IR neurons of the Cl group were located in the ventrolateral medulla in the area of the nucleus paragigantocellularis as defined by Andrezik [4]. Passing rostrally the PNMT-IR cell numbers in the Cl group increased from ap­proximately 1.4 mm caudal to obex in a steady almost linear pattern, to their highest density of 75-85 cells/section at 1.2-1.4 mm rostral to obex. From this level the cell numbers in the nucleus declined over another 1 mm, so that at the caudal edge of the facial nucleus few Cl PNMT cells were to be found (Figs. 1, 3). Throughout this group the Cl neurons were fusiform or multipolar with extensive dendritic branches.

The C2 PNMT -IR neurons extended from the middle of the area postrema caudally to a level approximately 2 mm rostral to obex. The number of cells per section in the C2 group was relatively constant rostro-caudally (Fig. 1). Caudally these cells were small and round, and were located in the dorsal strip area and the dorsal subnucleus of the nucleus of the solitary tract (NTS), and in the commissural subnucleus of NTS. At the rostral levels of the area postrema some PNMT -IR neurons appear in the medial portions of NTS. Further rostral these larger neurons were located in the medial and dorsal areas of the NTS and the dorsal motor nucleus of the vagus. These neurons appeared bipolar, with dendrites extending later­ally into the NTS and medially towards the mid­line.

The PNMT-IR neurons in the C3 group ex­tended rostrally from approximately 1 mm to 2.6 mm rostral to obex in the medial longitudinal fasciculus and dorsal to this structure, close to the ventricular surface. Other cells were seen lateral to the fasciculus in the dorsal medullary reticular formation and extending laterally in the preposi­tus hypoglossal nucleus to the NTS. PNMT-IR neurons in these regions extended rostrally be­yond the nucleus tractus solitarius. The C3 PNMT-IR neurons were large multipolar cells, with dendrites extending laterally and dorsovent­rally.

Spinally projecting PNMT-IR neurons

PNMT-IR neurons retrogradely labelled from T2-T4

Seven rats were injected with retrograde tracer at levels T2-T4 of the spinal cord. Retrogradely filled PNMT-IR neurons were observed in the areas of the Cl, C2 and C3 groups. These neurons contained dense, black particles with an amber­stained cytoplasm. The CTB-gold particles en­larged by silver deposit were visible in the soma and in proximal dendrites (Fig. 2). Other CTB­gold-labelled neurons in the regions of the Cl, C2 and C3 groups were not themselves PNMT -IR. CTB-gold-labelled cells were also found in areas of medulla devoid of PNMT-IR neurons - mainly

213

in the ventral midline area and extending laterally to the edges of the pyramidal tracts. In rats in­jected with tracer at the T2-T4 levels, 920 ± 69 (n = 7) PNMT-IR neurons were counted. 195 ± 16 (21 %) of these PNMT-IR cells were retrogradely labelled (Fig. 1). The retrogradely labelled cells, determined as the proportion of each PNMT -IR cell group, comprised 26 ± 2.2% of the Cl, 9 ± 1.6% of the C2 and 33 ± 5.1 % of the C3 groups, after bilateral injection of CTB-gold (n = 4). The PNMT-IR input to the T2-T4 levels of spinal cord derived largely (76%) from the Cl group, with 4% of input arising from the C2 group and 20% from the C3 group.

Retrogradely labelled PNMT -IR Cl neurons were observed rostrally from the level of the obex (Figs. 1, 2A). The numbers of these CTB-gold­labelled PNMT -IR cells per section increased to a maximum at approximately 1.4 mm rostral to obex, where they comprised 40-50% of the total Cl PNMT -IR population. Further rostral the CTB-gold-labelled PNMT-IR neurons accounted for an increasing proportion of the total PNMT-IR neurons. Virtually all of the most rostrally posi­tioned PNMT -IR Cl neurons were retrogradely labelled. In the C2 group the small PNMT -IR neurons in the dorsal NTS were not retrogradely labelled after spinal cord tracer injections. Only a small population of C2 PNMT -IR neurons were retrogradely labelled and these cells were found medially in the rostral NTS, where the NTS is separated laterally from the ventricle by the pre­positus hypoglossal nucleus (Figs. 1, 2B). Some of the most rostrally lying PNMT-IR cells in the C2 group were CTB-gold-Iabelled. CTB-gold-Iabelled neurons were found throughout the extent of the C3 group of PNMT -IR neurons (Fig. 1). These CTB-gold-Iabelled PNMT-IR cells were found scattered in the medial longitudinal fasciculus, some neurons lying adjacent to the ventricle and in the reticular formation 2C).

After unilateral injection of tracer at levels T2-T4 a stronger ipsilateral projection from the Cl cells was revealed, with 22 3.2% (n = 3) of PNMT -IR cells of the ipsilateral Cl nucleus being retrogradely labelled compared with 14 ± 0.9% (n = 3) of the contralateral Cl cell group (X2

6.535, P 0.0106, Table I). The C2 and C3 cell

\ ' . . A

'" .,. .. ~ ~(.

Fig. 2. Photomicrographs of PNMT-IR neurons in rat medulla localised in Cl (A), C2 (B) and C3 (C) areas. Higher power photomicrographs on the right are of the boxed areas and show PNMT-IR/CTB-gold-Iabelled neurons (single arrowheads). A. Cl neurons in RVLM at 1.4 mm rostral to obex, CTB-gold injection was at the T4level; in the low-power field there are 50 PNMT-IR neurons, of which 26 are CTB-gold-Iabelled. B. C2 neurons in the rostral NTS at 1.0 mm rostral to obex, CTB-gold injection was at the T8.1evel; in the. low power field there are six PNMT-IR neurons, one of which is CTB-gold-Iabelled. Three C3 neurons lie close to the ventricle. C. C3 neurons in the medial longitudinal fasciculus and prepositus hypoglossal nucleus at 1.6 mm rostral to the obex, CTB-gold injection was at the T8 level; in the low-power field there are 27 PMNT-IR neurons, 12 of which are CTB-gold labelled.

Calibration bar: low power 100 p.m; high power, 30 p.m.

groups did not reveal any lateral bias in their projection (Table I).

PNMT-IR neurons retrogradely labelled from T8-T9

Eight rats were injected with CTB-gold at levels T8-T9 of the spinal cord. Retrogradely labelled cells were found over a similar rostrocaudal extent of the medulla as observed after T2-T4 level injections (Fig. 1), and in similar medullary re­gions (Fig. 3). In these rats 993 ± 57 (n = 8) PNMT -IR cells were counted, 170 ± 9 (17%) of these PNMT-IR cells were retrogradely labelled

a proportion similar to that observed after tracer injection in upper thoracic (T2-T4) spinal cord. These CTB-gold-Iabelled PNMT-IR neurons displayed a similar rostrocaudal distribution to retrogradely labelled PNMT -IR neurons after tracer injection in T2-T4. CTB-gold-Iabelled PNMT-IR cells comprised 16 ± 1.0% of the Cl, 9 ± 3.1 % of the C2 and 30 ± 0.8% of the C3 groups after bilateral injection of CTB-gold (n = 4, Fig. 1), so that to the T8-T9 level of spinal cord the PNMT-IR input derived 65% from the Cl, 7%

+1.2

215

from the C2 and 28% from the C3 groups. A significantly lower proportion of the Cl PNMT-IR neurons were labelled from the T8-T9 level than from the T2-T4 level (16% compared to 26% X2 = 16.011, P = 0.0001, Table I). The numbers of spinally projecting C2 and C3 PNMT-IR neu­rons was similar after either T2-T4 or T8-T9 level tracer injections.

After unilateral injection of tracer at T8-T9 levels there were no differences in the two sides of the medulla in the proportions of retrogradely labelled PNMT-IR cells in the Cl, C2 or C3 groups.

Discussion

The present study demonstrates that in ad­dition to the established projection of the Cl PNMT-IR neurons of the RVLM to the inter­mediolateral cell column of the thoracic spinal cord, significant numbers of PNMT-IR neurons belonging to the C2 and C3 groups and localised in the dorsal medulla, also project to this region.

+2.0

Fig. 3. The distribution of PNMT-IR neurons (0) and PNMT-IR/crB-gold-labelled (e) neurons in one rat injected with CTB-gold bilaterally in the thoracic spinal cord at the T8 leveL Each eighth 50-p.m thick section through the medulla is presented. Distance is

given in mm from the obex.

216

Investigators have long sought evidence of direct connections of PNMT-IR fibres to the sym­pathetic preganglionic neurons in order to estab­lish anatomical evidence of a direct neural path­way which might modulate sympathetic outflow and blood pressure. In the light of the present findings it cannot be assumed that PNMT -IR fibres synapsing on identified sympathetic pregan­glionic neurons [43] arise only in the R VLM as originally suggested by Ross et al. [49]. Some PNMT -IR fibres in this area could arise from dorsal medullary regions - from either the C2 neurons of the NTS-dorsal vagal motor nucleus complex, or the C3 neurons in the mediallongitu­dinal fasciculus.

The topographical distribution of PNMT -IR cells in medulla has been found to be similar to that described previously [30,32,35]. The Cl cell group in the RVLM is the largest of the three medullary PNMT -IR groups of cells, PNMT -IR neurons being more numerous in the Cl group (72 % of total medullary PNMT -IR neurons), and also more extensively distributed rostrocaudally.

PNMT-IR projections to the spinal cord A population of PNMT -IR neurons extending

rostrally from the obex to the nucleus of the facial nerve projects to the spinal cord. Most of the

TABLE I

spin ally projecting PNMT-IR neurons are loca­lised in the rostral portions of the Cl nucleus of R VLM, supporting other retrograde transport studies [39,49,50,59] and electrophysiological stud­ies [26]. Approximately 65 % (after T8-T9 tracer injection) and 76% (after T2-T4 tracer injection) of retrogradely labelled PNMT -IR neurons were localised within the Cl cell group. This result is similar to the findings of a less detailed study [54], in which absolute cell numbers were not pre­sented, but where 69% of the retrograde1y labelled PNMT -IR neuron population were reported to be localised in Cl after large Tl-T2 spinal cord injections of the fluorescent tracer true blue. In the present study the proportion of Cl PNMT-IR neurons retrogradely labelled from the upper thoracic cord (T2-T4) was greater than from the lower thoracic cord (T8-T9), suggesting that the adrenergic innervation of the spinal cord is topo­graphically organised. Additionally, the projection of Cl PNMT -IR neurons to upper thoracic spinal cord displayed an ipsilateral predominance, but the projection to the lower thoracic spinal cord was comprised equally of PNMT-IR neurons on either side of the medulla. It is not yet known, however, whether any Cl PNMT-IR neurons pro­ject both to upper and lower thoracic levels. Some PNMT-IR neurons in the Cl area also project to

Numbers of PNMT-IR neurons and PNMT-IR/ CTE-gold labelled neurons in Cl. C2 and C3 groups after injection of CTE-gold into the spinal cord at levels T2-T4 or T8-T9

Tracer injection in spinal cord

TI-T4

T8-T9

PNMTcell group

Cl C2 C3

Cl C2 C3

Bilateral tracer

PNMT-IR PNMT-IR/ CTB-gold

585 54 148 ± 16 109 ± 16 8± 2 112 20 39± 10

(n =4)

706 ± 50 115 ± 9 * 159 24 12± 2 163 ± 9 49± 2

(n 4)

Unilateral tracer

Ipsilateral Contralateral

PNMT-IR PNMT-IR/ PNMT-IR PNMT-IR/ CTB-gold CTB-gold

386 ± 9 82 ± 10 393 9 56 ± 4 ** 73 ± 10 6± 1 77 6 2±1 74± 8 26± 6 70 14 20±2

(n 3)

368 ± 36 65 12 334 35 52 8 55 ± 11 3± 1 50 6 2±1 80± 4 26± 3 77±3 17 ±2

(n 4)

Results are means SEM. Number of rats is shown in parentheses. Analysis by X 2 test: * P 0.0001, comparing numbers of PNMT-IR/CTB-gold cells in Cl after tracer into T2-T4 vs. T8-T9; * * P = 0.0106, comparing ipsi- vs. contralateral distribution of PNMT-IR/CTB-gold neurons in Cl after unilateral tracer in TI-T4.

higher brain centres [26,47,53,54,55,59]. Anatomi­cal studies have suggested the Cl PNMT-IR neu­rons projecting to hypothalamus belong almost completely to a population of neurons indepen­dent of the spinally projecting Cl PNMT-IR neu­rons [59], although electrophysiologically-identi­fied, spinally projecting neurons in the Cl region, which are thought to correspond to the PNMT -IR cells, have been found to have collaterals ascend­ing through the central tegmental tract [26]. Possi­bly these collaterals might innervate the locus coeruleus [47].

Other PNMT-IR neurons projecting to the spi­nal cord were found in the dorsal medulla. These neurons were more numerous in the medial C3 group than in the C2 group. In contrast with the Cl spinal projection there were no topographical differences in the extent of innervation by the dorsal medullary neurons of the different thoracic levels of the spinal cord. Furthermore, the PNMT­IR innervation of the intermediolateral cell col­umn from the dorsal medulla derived equally from ipsi- and contralateral PNMT-IR neurons. Al­though spinally projecting neurons from dorsal medulla have been observed previously [1,9,49,52], only incidental observations had suggested that some of these neurons might be adrenergic. Lorenz et al. [39] in investigations of spinally projecting neurons from the RVLM, observed a small num­ber of spin ally projecting PNMT/substance P-IR cells along the medial edge of the NTS. Similarly, Hancock et al. noted spin ally projecting PNMT-IR neurons in the Cl, C2 and C3 groups [24,25,46]. In the more recent study of Sawchenko and Bohn [54], approximately 31% of the total number of PNMT-IR cells that could be retrogradely labelled following spinal injections have been observed localised in the C2 and C3 regions, a pattern of retrograde labelling similar to that demonstrated in this quantitative study_

Possible role of adrenergic spinal projections in cardiovascular control

The anatomical location of PNMT-IR fibres in the intermediolateral cell column has suggested a role for adrenergic neurons in the control of sym­pathetic spinal outflow. Projections of medullary PNMT -IR neurons observed in this study to up-

217

per and lower thoracic spinal cord are consistent with a cardiovascular role, since in the area of T2-T4 lie some of the cardiac and pulmonary sympathetic preganglionic neurons [57], as well as sympathetic preganglionic neurons projecting in the cervical sympathetic trunk. Lower thoracic cord (T8-T9) contains the highest numbers of sympathetic preganglionic neurons innervating the adrenal medulla (e.g. [3,7]).

The involvement of the rostral Cl PNMT -IR region in mediating changes in sympathetic activ­ity is well documented. Stimulation in this Cl region increases sympathetic nerve activity [16,40,48], circulating adrenaline levels [40,52] and blood pressure [16,21,28,40,44,48,50,52,61]. Fur­thermore, inhibition of the Cl region of the RVLM evokes a profound fall in arterial pressure and a reduction in basal sympathetic discharge [17]. The observation that spinally projecting presumed PNMT-IR neurons are barosensitive and pulse­modulated [26,45], suggests that the physiological responses are at least in part mediated by adren-

neurons in this region. That the PNMT-IR neurons in the Cl group might have a role in the reflex control of blood pressure is supported by anatomical evidence of a substantial direct input from the intermediate NTS, an area of barorecep­tor afferent input [14], to the PNMT-IR Cl neu­rons [24].

The dorsal medullary PNMT -IR neurons pro­jecting to spinal cord are also well located to have a role in the regulation of blood pressure, and some evidence suggests that a component of basal sympathetic nerve discharge originates in these areas of the dorsal medulla [8]. It is possible these spinally projecting C2 PNMT-IR neurons might directly mediate baro- or chemoreceptor reflex activity since these neurons are localised in the NTS, the termination site of the baroreceptor afferents [14,34,60]. Furthermore substance P, a putative neurotransmitter in afferent fibres of the IXth and Xth cranial nerves transmitting baro­and chemoreceptor information [20,27], has been observed in processes close to these bulbospinal C2 PNMT-IR neurons [46]. In addition, local cir­cuits within NTS might also mediate activity of PNMT-IR neurons of the C2 group, since the intermediate NTS also targets these neurons [24].

218

A cardiovascular regulatory role of the C3 PNMT-IR is consistent with the observations of Chai and co-workers [13,37], that chemical stimu­lation in the rostral dorsomedial medulla near the NTS and in the lateral tegmental field, elevated arterial pressure, possibly by the direct activation of spin ally projecting neurons. The region of the C3 PNMT-IR neurons, the prepositus hypoglossal nucleus, together with the Cl region of RVLM, is a major source of afferents to the locus coeruleus [6] and some of this input arises from the PNMT­IR neurons [47]. It has been observed that dis­charge of the locus coeruleus parallels sympathetic nerve activity [18]. Such co-regulation of locus coeruleus and sympathetic preganglionic neurons might arise, in the light of present observations, from the ,neurons of the C3 area in the dorsal medulla and not only from RVLM neurons as previously suggested [6]. Such an association of locus coeruleus function with sympathetic activity, and therefore blood pressure regulation, is con­sistent with the postulated role of the locus coeruleus in arousal and anxiety mediating sys­tems.

In addition to the observed increases in sym­pathetic activity that accompany stimulation of the PNMT-IR neuron areas, support for a sym­pathoexcitatory neurotransmitter role of adren­aline comes also from in vitro studies. Intracellu­lar recordings from sympatheti preganglionic neu­rons have shown an increase in neuronal excitabil­ity and a potential for repetitive discharge, after application of adrenaline [62]. Other studies have however suggested a sympathoinhibitory role for adrenaline, since iontophoresis of adrenaline onto sympathetic preganglionic neurons inhibited their discharge [15,22]. PNMT-IR synapses have now been demonstrated on sympathetic preganglionic neurons, both on perikarya and dendrites [43], and so it is possible that these observed excitatory or inhibitory effects of adrenaline ret1ect different sites of release of adrenaline onto the sympathetic preganglionic neuron.

In summary, the quantitative evaluation of PNMT -containing neuronal projections from the medulla to the intermediolateral cell column area of thoracic spinal cord suggests a significant role for the dorsal adrenergic cell groups. All the

adrenergic neurons are well situated in the brainstem to sub serve a sympathetic regulatory function.

Acknowledgements

This work was supported by the National Health and Medical Research Council of Australia, the National Heart Foundation of Australia and the Medical Research Council, U.K. We wish to thank Dr. P.R.C. Howe, CSIRO Division of Hu­man Nutrition, Adelaide, for his kind gift of anti­serum to PNMT and Ms. K. Honkoop for techni­cal assistance.

References

1 Amendt, K, Czarchurski, J., Dembowksy, K and Seller, H., Bulbospinal projections to the intermediolateral cell column: a neuroanatomical study, J. Auton. Nerv. Syst., 1 (1979) 103-117.

2 Anderson, CR. and Howe, P.R.C, Is phenylethanolamine­N-methyltransferase (PNMT) contained in rat hypo­thalamic neurons? Neurosci. Lett., 93 (1988) 164-169.

3 Anderson, CR., McLachlan, E.M. and Srb-Christie, 0., Distribution of sympathetic preganglionic neurons and mono-aminergic nerve terminals in the spinal cord of the rat, J. Comp. Neurol., 283 (1989) 269-284.

4 Andrezik, J.A, Chan-Palay, V., Palay, S.L., The nucleus para-gigantocellularis lateralis in the rat, Anat. Embryol., 161 (1981) 355-371.

5 Armstrong, D.M., Ross, C.A, Pickel, V.P., Joh, T.H. and Reis, D.J., Distribution of dopamine, noradrenaline and adrenaline-containing cell bodies in the rat medulla ob­longata: demonstration by immunocytochemical localiza­tion of catecholamine biosynthetic enzymes, J. Comp. Neu­rol., 212 (1982) 173-187.

6 Aston-Jones, G., Ennis, M., Pieribone, V.A, Nickell, W.T. and Shipley, M.T., The brain nucleus locus coeruleus: restricted afferent control of a broad efferent network, Science, 234 (1986) 734-737.

7 Bacon, S.J. and Smith, AD., Preganglionic neurons in­nervating the rat adrenal medulla: immunocytochemical evidence of synaptic input from nerve terminals containing substance P, GABA or 5-hydroxytryptamine, J. Auton. Nerv. Syst., 24 (1988) 97-122.

8 Barman, S.M. and Gebber, G.L., Sequence of activation of ventrolateral and dorsal medullary sympathetic neurons, Am. J. Physiol., 245 (1983) R438-R447.

9 Blessing, W.W., Goodchild, AK, Dampney, R.AL. and Chalmers, J.P., Cell groups in the lower brain stem of the

rabbit projecting to the spinal cord, with special reference to catecholamine-containing neurons, Brain Res., 221 (1981) 35-55.

10 Blessing, W.W., Oliver, J.R., Hodgson, AH., John, T.H. and Willoughby, J.O., Neuropeptide Y-like immunoreactive Cl neurons in the rostral ventrolateral medulla of the rabbit project to sympathetic preganglionic neurons in the spinal cord, J. Auton. Nerv. Syst., 18 (1987) 121-129.

11 Brown, D.L. and Guyenet, P.G., Electrophysiological study of cardiovascular neurons in the rostral ventrolateral medulla in rats, Circ. Res., 56 (1985) 359-369.

12 Carlton, S.M., Honda, CN., Denoroy, L. and Willis, W.D., Jr., Descending phenylethanolamine-N-methyltransferase projections to the monkey spinal cord: an immuno­histochemical double labelling study, Neurosci. Lelt., 76 (1987) 133-139.

13 Chai, C.Y., Lin, R.H., Lin, AM.Y., Pan, CM., Lee, E.H.Y. and Kuo, J.S., Pressor responses from electrical or gluta­mate stimulations of the dorsal or ventrolateral medulla, Am. J. Physiol., 255 (1988) R709-R717.

14 Ciriello, J., Brainstem projections of aortic baroreceptor afferent fibres in the rat, Neurosci. Lelt., 36 (1983) 37-42.

15 Coote, J.H., MacLeod, V.H., Fleetwood-Walker, S. and Gilbey, M.P., The response of individual sympathetic pre­ganglionic neurons to microiontophoretically applied en­dogenous monoamines, Brain Res., 215 (1981) 135-145.

16 Dampney, R.AL., Goodchild, A.K., Robertson, L.G. and Montgomery, W.,Role of ventrolateral medulla in vasomo­tor regulation: a correlative anatomical and physiological study, Brain Res., 249 (1982) 223-235.

17 Dampney, RAL. and Moon, E.A, Role of the ventro­lateral medulla in vasomotor response to cerebral ischemia, Am. J. Physio/., 239 (1980) H349-H358.

18 Elam, M., Yao, T., Svensson, T.H. and Thoren, P., Regu­lation of locus coeruleus neurons and splanchnic, sym­pathetic nerves by cardiovascular afferents, Brain Res., 290 (1984) 281-287.

19 Foster, G.A, Hokfelt, T., Coyle, J.T. and Goldstein, M., Immunohistochemical evidence for phenylethanolamine N-methyl-transferase positive/tyrosine hydroxylase-nega­tive neurons in the retina and the posterior hypothalamus of the rat, Brain Res., 330 (1985) 183-188.

20 Gillis, R.A, Helke, C.J., Hamilton, B.L., Norman, W.P. and Jacobowitz, D.M., Evidence that substance P is a neurotransmitter of baro- and chemoreceptor afferents in nucleus tractus solitarius, Brain Res., 181 (1980) 476-481.

21 Goodchild, AK., Moon, E.A, Dampney, R.AL. and Howe, P.R.C, Evidence that adrenaline neurons in the rostral ventrolateral medulla have a vasopressor function, Neuro­sci. Lelt., 45 (1984) 267-272.

22 Guyenet, P.G. and Stometta, RL., Inhibition of sym­pathetic preganglionic discharges by epinepbrine and a­methylepinephrine, Brain Res., 235 (1982) 271-283.

23 Halliday, G.M., Li, Y.-W., Joh, T.H., Cotton, RH., Howe, P.R.C, Geffen, L.B. and Blessing, W.W., Monoamine­synthesizing neurons in the medulla oblongata of the adult human, J. Comp. Neurol., 273 (1988) 301-317.

219

24 Hancock, M.B., Evidence for direct projections from the nucleus of the solitary tract onto medullary adrenaline cells, J. Comp. Neurol., 276 (1988) 460-467.

25 Hancock, M.B. and Nicholas, AP., Oxytocin-immunoreac­tive projections onto medullary adrenaline neurons, Brain Res. Bull., 18 (1987) 213-219.

26 Haselton, J.R. and Guyenet, P.G., Electrophysiological characterization of putative Cl adrenergic neurons in the rat, Neuroscience, 30 (1989) 199-214.

27 Helke, CJ., 0' Donohue, T.L. and Jacobowitz, D.M., Sub­stance P and a baro- and chemoreceptor afferent neuro­transmitter: immunocytochemical and neurochemical evi­dence in the rat, Peptides, 1 (1980) 1-9.

28 Hodgson, AJ., Minson, J.B., Chalmers, J.P. and Pilowsky, P.M., The use of microinjected colloidal gold and im­munocytochemistry to localise pressor sites in the rostral medulla oblongata of the rat, Neurosci. Lett., 77 (1987) 125-130.

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

30 Hokfelt, T., Martensson, R., Bjorklund, A, Kleinan, S. and Goldstein, M., Distributional maps of tyrosine-hydroxyl­ase-immunoreactive neurons in the rat brain, in Handbook of Chemical Neuroanatomy, Vol. 2. Classical Transmitters in the CNS, Part 1. A Bjorklund and T. Hokfelt, Eds. Else­vier, Amsterdam, 1984.

31 Hokfelt, T., Terenius, L., Kuypers, H.G.J.M. and Dann, 0., Evidence for enkephalin immunoreactive neurons in the medulla oblongata projecting to the spinal cord, Neurosci. Lett., 14 (1979) 55-60.

32 Howe, P.R.C, Costa, M., Fumess, J.B. and Chalmers, J.P., Simultaneous demonstration of phenylethanolamine N­methyl-transferase immunofluorescent and catecholamine fluorescent nerve cell bodies in rat medulla oblongata, Neuroscience, 5 (1980) 2229-2238.

33 Johansson, 0., Hokfelt, T., Pemow, B., Jeffcoate, S.L., White, N., Steinbusch, H.W.M., Verhofstad, A.AJ., Evin­son, P.e. and Spindel, E., Immunohistochemical support for three putative transmitters in one neuron: coexistence of 5-hydroxytryptamine-, substance P- and thyrotropin-re­leasing hormone-like immunoreactivity in medullary neu­rons projecting to the spinal cord, Neuroscience, 6 (1981) 1857-1881.

34 Kalia, M., Localization of aortic and carotid baroreceptor and chemoreceptor primary afferents in the brain stem, in Perspectives in Cardiovascular Research, Vol. VI. Central Neurons System Mechanisms in Hypertension, AM. Katz, J.P. Buckley and C.M. Ferrario, Eds., Raven Press, New York, 1981, pp. 9-24.

35 Kalia, M., Fuxe, K. and Goldstein, M., Rat medulla ob­longata Ill. Adrenergic (Cl and C2) neurons, nerve fibres and presumptive terminal processes,. J. Comp. Neurol., 233 (1985) 333-349.

36 Kitahama, K., Pearson, J., Denoroy, L., Kopp, N., Ulrich, J., Maeda, T. and Jouvet, M., Adrenergic neurons in human

220

brain demonstrated by immunohistochemistry with anti­bodies to phenylethanolamine N-methyltransferase (PNMT): discovery of a new group in the nucleus tractus solitarius. Neurosci. Lett., 53 (1985) 303-308.

37 Lin, AM.Y., Wang, Y., Kuo, l.S. and Chai, CY., Homo­cysteic acid elicits pressor responses from ventrolateral medulla and dorsomedial medulla, Brain Res. Bull., 22 (1989) 627-631.

38 Llewellyn-Smith, U., Minson, l.B., Wright, AP. and Hodg­son, Al., Cholera toxin B-gold, a retrograde tracer that can be used in light and electron microscopic immuno­cytochemical studies, J. Comp. Neurol., 294 (1990) 179-191.

39 Lorenz, R.G., Saper, CB., Wong, D.L., Ciaranello, R.D. and Loewy, AD., Co-localization of substance P- and phenylethanolamine-N-methyl-transferase-like immunore­activity in neurons of ventrolateral medulla that project to the spinal cord: potential role in control of vasomotor tone, Neurosci. Lett., 55 (1985) 255-260.

40 McAllen, R.M., Action and specificity of ventral medullary vasopressor neurones in the cat, Neuroscience, 18 (1986) 51-59.

41 Menetrey, D. and Basbaum, A.I.,)The distribution of sub­stance po, enkephalin- and dynorphin-immunoreactive neu­rons in the medulla of the rat and their contribution to bulbospinal pathways, Neuroscience, 23 (1987) 173-187.

42 Millhom, D.E., Hokfelt, T., Seroogy, K., Oertel, W., Verhofstad, AAJ. and Wu, J.-Y., Immunohistochemical evidence for colocalization of y-aminobutyric acid and serotonin in neurons of the ventral medulla oblongata projecting to the spinal cord, Brain Res., 410 (1987) 179-185.

43 Milner, T.A, Morrison, S.F., Abate, C and Reis, D.J., Phenylethanolamine N-methyltransferase-containing termi­nals synapse directly on sympathetic preganglionic neurons in the rat, Brain Res., 448 (1988) 205-222.

44 Minson, J.B., Chalmers, J.P., Caon, A.C. and Renaud, B., Separate areas of rat medulla oblongata with populations of serotonin- and adrenaline-containing neurons alter blood pressure after L-glutamate stimulation, J. Auton. Neru. Syst., 19 (1987) 39-50.

45 Morrison, S.F., Milner, TA., and Reis, D.J., Reticu!ospinal vasomotor neurons of the rat rostra! ventrolateral medulla: relationship to sympathetic nerve activity and the Cl adrenergic cell group, J. Neurosci., 8 (1988) 1286-1301.

46 Nicholas, AP. and Hancock, M.B., Immunocytochemical evidence for substance P and serotonin input to medullary bulbospinal adrenergic neurons, Synapse, 2 (1988) 569-576.

47 Pieribone, V.A, Aston-Jones, G. and Bohn, M.C, Adren­ergic and non-adrenergic neurons in the Cl and C3 areas project to locus coeruleus: a fluorescent double labelling study, Neurosci. Lett., 85 (1988) 297-303.

48 Pilowsky, P., West, M. and Chalmers, l.P., Renal sym­pathetic nerve responses to stimulation, inhibition and de­struction of the ventrolateral medulla in the rabbit, Neuro­sci. Lett., 60 (1985) 51-55.

49 Ross, C.A, Armstrong, D.M., Ruggiero, D.A, Pickel, V.A, Joh, T.H. and Reis, D.L., Adrenaline neurons in the rostral ventrolateral medulla innervate thoracic spinal cord: a

combined immunocytochemical and retrograde transport demonstration, Neurosci. Lett., 25 (1981) 257-262.

50 Ross, CA, Ruggiero, D.A, Joh, T.R., Park, D.H. and Reis, D.J., Adrenaline synthesizing neurons in the rostral ventrolateral medulla: a possible role in tonic vasomotor control, Brain Res., 273 (1983) 356-361.

51 Ross, CA, Ruggiero, D.A, Meeley, M.P., Park, D.H., Joh, T.H. and Reis, D.J., A new group of neurons in hypo­thalamus containing phenylethanolamine N-methyltrans­ferase (PNMT) but not tyrosine hydroxylase, Brain Res., 306 (1983) 349-353.

52 Ross, CA, Ruggiero, D.A, Park, D.H., Joh, T.H., Sved, A.F., Fernandez-Pardal, J., Saavedra, l.M. and Reis, D.J., Tonic vasomotor control by the rostral ventrolateral medulla: effect of electrical or chemical stimulation of the area containing Cl adrenaline neurons on arterial pressure, heart rate, and plasma catecholamines and vasopressin, J. Neurosci., 4 (1984) 474-494.

53 Saper, CB., Reis, D.J. and Joh, T., Medullary catechola­mine inputs to the anteroventral third ventricular cardio­vascular regulatory region in the rat, Neurosci. Lett., 42 (1983) 285-291.

54 Sawchenko, P.E. and Bohn, M.C., Glucocorticoid receptor­immunoreactivity in Cl, C2 and C3 adrenergic neurons that project to the hypothalamus or to the spinal cord in the rat, J. Comp. Neurol., 285 (1989) 107-116.

55 Sawchenko, P.E., Swanson, L.W., Grzanna, R., Howe, P.R.C, Bloom, S.R. and Polak, J.M., Co-localization of neuropeptide Y immunoreactivity in brainstem catecho­laminergic neurons that project to the paraventricular nucleus of the hypothalamus, J. Comp. Neurol., 241 (1985) 138-153.

56 Somogyi, P. and Takagi, H., A note on the use of picric acid-paraformaldehyde-glutaraldehyde for correlated light and electron microscopic immunocytochemistry, Neurosci­ence, 7 (1982) 1779-1783.

57 Strack, AM., Sawyer, W.B., Marubio, L.M. and Loewy, AD., Spinal origin of sympathetic preganglionic neurons in the rat, Brain Res., 455 (1988) 187-191.

58 Sun, M.-K. and Guyenet, P.G., Effect of c10nidine and y-aminobutyric acid on the discharges of medullo-spinal sympatho-excitatory neurons in the rat, Brain Res., 368 (1986) 1-17.

59 Tucker, D.C, Saper, CB., Ruggiero, D.A and Reis, D.J., Organization of central adrenergic pathways: L Relation­ships of ventrolateral medullary projections to the hypo­thalamus and spinal cord, J. Comp. Neurol., 259 (1987) 591-603.

60 Wallach, J.H. and Loewy, AD., Projections of the aortic nerve to the nucleus tractus solitarius in the rabbit, Brain Res., 188 (1980) 247-251.

61 Willette, R.N., Barcas, P.P., Krieger, AJ. and Sapru, H.N., Vasopressor and depressor areas in the rat medulla: identi­fication by L-glutamate microinjections. Neuropharmacol­ogy, 22 (1983) 10~1-1079.

62 Yoshimura, M., Polosa, C and Nishi, S., Noradrenaline modifies sympathetic preganglionic neuron spike and afterpotential, Brain Res., 362 (1986) 370-374.