Neuroscience Letters, 109 (1990) 23 29 Elsevier Scientific Publishers Ireland Ltd.
23
NSL 06614
Adrenergic innervation of noradrenergic locus coeruleus neurons. A dual labeling immunocytochemical study in the rat Philippe Kachidian 1, Bernadette Astier z, Bernard Renaud z and Olivier Bosler 1 IEquipe de Neuromorphologie Fonctionnelle, Laboratoire de Neurobiologie, CNRS, Marseilles (France) andZ Laboratoire de Neuropharmacologie, CNRS, UMR 105, Facultk de Pharmacie, Universitk Claude Bernard, Lyons (France) (Received 5 August 1989; Revised version received 25 September 1989; Accepted 25 September 1989) Key words:
Adrenaline; Noradrenaline; Locus ceruleus; Dual immunocytochemistry; Electron micro-
scopy By means of dual immunocytochemistry, synaptic associations between adrenergic terminals and noradrenergic neurons were directly demonstrated in the rat locus ceruleus (LC). It could be estimated that every adrenergic afferent contacts at least one noradrenergic dendrite in the nucleus. An adrenergic innervation of non-noradrenergic targets was also evidenced. These data add to our knowledge on the synaptic circuitry by which activation of the adrenergic input could affect central mechanisms known to be influenced by LC neurons.
The locus ceruleus (LC) which provides extensive noradrenergic fiber projections throughout the central nervous system (CNS) (for reviews, see refs 2, 7, 19) is under the influence of a restricted afferent control system [6, 8]. One of the afferent inputs to the LC is composed of adrenergic fibers as identified by their immunoreactivity toward phenylethanolamine-N-methyl transferase (PNMT), the specific biosynthetic enzyme of adrenaline [4, 13]. This adrenergic projection which forms dense terminal plexuses in the whole LC [13] is derived via the medullary longitudinal axon bundle [4] from the ipsilateral C1 cellular group of the ventrolateral medulla and, to a lesser extent, from its C3 homologous in the dorsomedial medulla (23 also see ref. 12). That this adrenergic supply is of critical importance in regulating the output of LC noradrenergic neurons has been suggested by electrophysiological and biochemical pieces of evidence [1, 3, 5, 10, 11, 25] and by the recent electron microscopic demonstration that as many as 30% of all identified LC terminals consisted of PNMT-immunoreacCorrespondence." 0. Bosler, Laboratoire de Neurobiologie, CNRS, B.P. 71, 13402 Marseille c6dex 09, France. 0304-3940/90/$ 03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd.
25
tive axonal varicosities forming numerous synaptic junctions [18]. In the present paper, we report ultrastructural evidence from a dual labeling immunocytochemical study that LC adrenergic terminals innervate noradrenergic as well as non-noradrenergic targets. Adult male Sprague-Dawley rats were deeply anesthetized with chloral hydrate and perfused through the ascending aorta with 200 ml of 3.75% acroleine-2% paraformaldehyde mixtures in 0.12 M phosphate buffer and with 500 ml of 2% paraformaldehyde in the same buffer [20]. After overnight post-fixation of the brains, transverse sections through the entire LC were cut on a vibratome at 50/zm thickness. They were sequentially immunostained using a polyclonal rabbit anti-PNMT antiserum kindly provided by Dr. L. Denoroy (see ref. 15 for its production and specificity criteria), and a monoclonal anti-tyrosine hydroxylase (TH) antiserum from mouse-mouse hybrid cells (Boehringer Mannheim Biochemica). Dual labeling was achieved by combination of the peroxidase-antiperoxidase (PAP) method (PNMT staining) with a radioimmunocytochemical method which involved the use of a ~25Ilabeled secondary antiserum (TH labeling). Briefly, groups of three adjacent sections were pre-incubated in 1% sodium borohydride (30 min) and 0.3% hydrogen peroxide (30 min), and processed for PNMT staining as previously described [9]. After reaction with 0.05% 3,3'-diaminobenzidine (DAB) and 0.01% hydrogen peroxide added with 0.025% cobalt chloride and 0.02% nickel ammonium sulfate, they were incubated in a 1:300 dilution of the anti-TH antiserum (16-18 h) and in a solution of 1251-labeled anti-mouse IgG from sheep (spec. act.: 28-111 TBq/mmol, Amersham) diluted 1:50 from a 3700 MBq/ml radioactive solution. Two sections from each group were kept for light microscopic examination of PNMT staining and TH radiolabeling, respectively. Other sections which served as controls were run through the dual labeling sequence with omission of the anti-TH antiserum. As expected from a lack of cross-reactions between IgGs used in the first and second stages of the procedure, they did not show any nerve cell body 125I-labeiing at the light microscopic level. The remaining sections in each group were
Fig. 1. An axonal varicosity showing dual labeling for TH (silver grains) and PNMT (peroxidase staining) is directly apposed to an unlabeled soma. No synaptic membrane differentiation is visible on this plane of section. Fig. 2. A PNMT-positive terminal is synaptically linked to a TH-radiolabeled soma with an asymmetrical junctional complex. Fig. 3. A PNMT-positive terminal synaptically contacts both a TH-radiolabeled dendrite (arrow) and an unlabeled dendritic spine (*). Note that the spine is completely surrounded by its immunostained partner. Fig, 4. Other examples of PNMT-stained synaptic terminals surrounding dendritic spines (*). Figs. 5 and 6. Serial sections showing two PNMT/TH-positive terminals that simultaneously contact one same unlabeled dendrite with visible junctional complexes. Note that the superimposed TH-labeling of the terminals is only significant on Fig. 6. Another large, TH-labeled dendritic shaft is also contacted by one of the immunostained varicosities.
26
Fig. 7. A TH radiolabeled dendritic shaft, largely apposed to the basal lamina of a capillary (c) receives two synaptic terminals. One of these is PNMT-positiveand forms asymmetricaljunctional complexes.The other one does not show any significant staining and is involved in a symmetricaljunction I*). Fig. 8. The synaptic PNMT-stained terminal shown here is engaged in asymmetricaljunctional complexes with three dendritic elements, one of which is a large TH-radiolabeled proximal dendrite.
immersed for 1 h in 2% osmium tetroxide, stained "en bloc" with uranyl acetate and flat-embedded in Epon between siliconized glass slides. Ribbons of ultrathin sections were cut from the surface of selected sections from 3 rostrocaudal levels of the LC showing good peroxidase P N M T staining. After processing for radioautography by means of standard dipping techniques, they were exposed for 2 3 months and developed with K o d a k microdol X (1 2 min at 18 "C). PNMT-positive structures stained with the dense DAB endproduct were found to exclusively consist of axons and, most of all, of axonal varicosities whose ultrastructural features were similar to those previously reported by Milner et al. [18] (Figs. 1-8). Immunoreactive varicose profiles ranged from 0.7 to 2 / t m in transverse diameter, being then somewhat larger than those previously described in other parts of the brain [9, 17, 21]. They were frequently engaged in axo-dendritic synapses (Figs. 3--8) and only occasionally contacted somatic profiles (Figs. 1, 2). The junctional complexes were of either the symmetrical or the asymmetrical type and usually occupied relatively large areas of the apposed plasma membranes.
27 From a sampling of 151 positively stained varicosities, not less than 47% showed synaptic densities in single sections, which tends to indicate that the adrenergic innervation of the LC could well be entirely junctional. This is in contrast with the figure reported for another important monoaminergic afferent system to the LC, namely the serotoninergic input which has been previously shown to be essentially of the non-junctional type [16, 22]. Even a greater number of PNMT-immunoreactive profiles (about 85% for a total of 141) was found to be engaged in synaptic junctions by Milner et al. [18], but these authors have reported that the majority of the synaptic contacts was found in tissue that had been sectioned in the sagittal rather than the coronal plane. The fact that our sampling only involved coronal sections could then explain the discrepancy. In any event, the important synaptic investment of LC neurons by adrenergic afferents provides a morphological substrate for the previously reported inhibitory control of the LC by adrenergic neurons [3, 5, 10, 11, 25]. Surprisingly, only a small fraction (< 10%) of the PNMT-positive profiles examined also exhibited significant TH-radiolabeling (Figs. 1, 6), an observation which is in keeping with the initial results of Pickel et al. [22] who did not describe TH-immunoreactive terminals in the LC after staining with the PAP technique. This could be accounted for by a low sensitivity of electron microscopic immunocytochemical procedures for localizing axonal TH immunoreactivity in the LC, as also reported in the hypothalamus [24]. From 71 PNMT-positive synaptic terminals, 26 of which were examined on at least two consecutive sections, 46 innervated TH-immunolabeled targets. This indicates that, in addition to the noradrenergic population, neurons of a non-noradrenergic identity would also constitute important targets for the synaptic action of adrenaline in the LC. Both labeled and unlabeled post-synaptic processes essentially consisted of proximal dendritic shafts and medium to small branches which could be contacted by more than one immunoreactive terminal (Figs. 5, 6) and usually received other unreactive afferents (Fig. 7), as well as of distal branchlets and/or spines (Figs. 3, 4). Only one TH-positive (Fig. 2) and three TH-negative (Fig. 1) post-synaptic targets were identifed as nerve cell bodies, confirming that the adrenergic regulation of LC neurons essentially takes place at the level of the dendritic tree of these neurons [18]. The TH-immunolabeled processes often directly apposed each other and, in favorable planes of section, single PNMT-positive afferent fibers could be seen synapsing simultaneously upon two such processes that were contiguous or not. A given immunoreactive terminal could also synapse simultaneously on TH-labeled and unlabeled dendritic processes (Fig. 8). Given that not each PNMT-immunoreactive varicosity engaged in a synapse could be expected to show the differentiated active zone(s) on a single etectron microscopic section, the actual incidence of synaptic junctions established by adrenergic terminals in the LC was obviously underestimated. Then, the fact that as much as 65% of the PNMT-immunoreactive profiles were identified as presynaptic to TH-positive elements could indicate that every adrenergic afferent to the LC actually makes a synapse with a noradrenergic target. This figure would imply that the non-noradrenergic synaptic partners of the adrenergic afferents to the LC are always contacted
28
by these afferents together with at least one noradrenergic dendrite. Taken together, these data should be of potential value in determining the cellular mechanisms by which activation of the adrenergic input could affect central mechanisms known to be influenced by the LC (for reviews, see refs. 2, 8). As concerns the possible chemical identity of post-synaptic dendrites that did not show TH immunolabeling, it is interesting to note that medium-sized and small neurons immunopositive for GABA and presumed to constitute inhibitory interneurons have recently been described within the rat LC [14]. This study was supported by Grants 856003 (O.B.) and 856017 (B.R.) from the INSERM. We are particularly grateful to Dr. L. Denoroy for providing us with the anti-PNMT antiserum. We also thank Mrs. B. Besson for secretarial assistance. 1 Aghajanian, G.K. and Vandermaelen, C.P., Alphaz-adrenoceptor-mediatedhyperpolarization of locus coeruleus neurons: intracellular studies in vivo, Science, 215 (1982 1394~ 1396. 2 Amaral, D.G. and Sinnamon, H.M., The locus coeruleus: neurobiology of a central noradrenergic nucleus, Progr. Neurobiol., 9 (1977) 147 196. 3 Astier, B. and Aston-Jones, G., Electrophysiological evidence for medullary adrenergic inhibition of rat locus coeruleus, Soc. Neurosci. Abstr., 15 (1989) in press. 4 Astier, B., Kitahama, K., Denoroy, L., Jouvet, M. and Renaud, B., lmmunohistochemical evidence for the adrenergic medullary longitudinal bundle as a major ascending pathway to the locus coeruleus, Neurosci. Lett., 74 (1987) 132 138. 5 Astier, B., Kitahama, K., Denoroy, L., B6rod, A., Jouvet, M. and Renaud, B., Biochemical evidence for an interaction between adrenaline and noradrenaline neurons in the rat brainstem, Brain Res., 397 (1986) 333- 340. 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 Aston-Jones, G., Foote, S.L. and Bloom, F.E., Anatomy and physiology of locus coeruleus neurons: functional implications. In M. Ziegler and C. Lake (Eds.), Frontiers in Clinical Neuroscience: Vol. 2, Norepinephrine, Williams and Wilkins, Baltimore, 1984, pp. 92-116. 8 Aston-Jones, G., Shipley, M.T., Ennis, M., Williams, J.T. and Pieribone, V.A., Restricted afferent control of locus coeruleus neurons revealed by anatomic, physiologic and pharmacologic studies. In C.A. Marsden and D.J. Heal (Eds./, The Pharmacology of Noradrenaline in the Central Nervous System, Oxford University Press, Oxford, in press. 9 Bosler, O., Beaudet, A. and Denoroy, L., Electron-microscopic characterization of adrenergic axon terminals in the diencephalon of the rat, Cell Tissue Res., 248 (1987) 393 398. l0 Cedarbaum, J.M. and Aghajanian, G.K., Noradrenergic neurons of the locus coeruleus: inhibition by epinephrine and activation by the :~-antagonist piperoxane, Brain Res., 112 (1976) 413 419. 11 Engberg, G., Elam, M. and Svensson, T.H., Effects of adrenaline synthesis inhibition on brain noradrenaline neurons in locus coeruleus, Brain Res., 223 (1981) 49-58. 12 Guyenet, P.G. and Young, B.S., Projection of nucleus paragigantocellularis lateralis to locus coeruleus and other structures in rat, Brain Res., 406 (1987) 171-184. 13 H6kfelt, T., Johansson, O. and Goldstein, M., Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS: Part l, Elsevier, Amsterdam, 1984, pp. 157 276. 14 Ijima, K. and Ohtomo, K., Immunocytochemical study using a GABA antiserum for the demonstration of inhibitory neurons in the rat locus coeruleus, Am. J. Anat., 181 (1988) 43 52. 15 Kitahama, K., Pearson, J., Denoroy, L., Kopp, N., Ulrich, J., Maeda, T. and Jouvet, M., Adrenergic neurons in human brain demonstrated by immunocytochemistry with antibodies to phenylethanola-
29 mine-N-methyltransferase (PNMT). Discovery of a new group in the nucleus tractus solitarius, Neurosci. Lett., 53 (1985) 303 308. 16 L6ger, L. and Descarries, L., Serotonin nerve terminals in the locus coeruleus of adult rat: a radioautographic study, Brain Res., 145 (1978) 1 13. 17 Liposits Zs., Phelix, C. and Paull, W.K., Electron microscopic analysis of tyrosine hydroxylase, dopamine fl-hydroxylase and phenylethanolamine-N-methyltransferase immunoreactive innervation of the hypothalamic paraventricular nucleus in the rat, Histochemistry, 84 (1986) 105-I 20. 18 Milner, T.A., Abate, C., Reis, D.J. and Pickel, V.M., Ultrastructural localization of phenylethanotamine-N-methyltransferase-like immunoreactivity in the rat locus coeruleus, Brain Res.,(1989) 1 15. 19 Moore, R.Y. and Card, J.P., Noradrenaline-containing neuron systems. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS: part I, Elsevier, Amsterdam, 1984, pp. 123 156. 20 Pickel, V.M., Chart, J. and Milner, T.A., Autoradiographic detection of ~2Sl-secondary antiserum: a sensitive light and electron microsope labeling method compatible with peroxidase immunocytochemistry for dual localization of neuronal antigens, J. Histochem. Cytochem., 34 (1986) 707 718. 21 Pickel, V,M., Chan, J., Park, D.H., Joh, T.H. and Milner, T.A., Ultrastructural localization of phenylethanolamine N-methyltransferase in sensory and motor nuclei of the vagus nerve, J. Neurosci. Res., 15 (1986) 43~456. 22 Pickel, V.M., Joh, T.H. and Reis, D.J., A serotonergic innervation of noradrenergic neurons in nucleus locus coeruleus: demonstration by immunocytochemical localization of the transmitter specific enzymes tyrosine and tryptophan hydroxylase, Brain Res. 131 (1977) 197 214. 23 Pieribone, V.A.. Aston-Jones, G. and Bohn, M.C., Adrenergic and non-adrenergic neurons in the CI and C3 area project to locus coeruleus: a fluorescent double label study, Neurosci. Lett., 85 (1988) 297-303. 24 Piotte, M., Beaudet, A., Joh, T.H. and Brawer, J.R., The fine structural organization of tyrosine hydroxylase immunoreactive neurons in rat arcuate nucleus, J. Comp, Neurol., 239 (1985) 44-53. 25 Stolk, J.M., Vantini, G., Perry, B.D., Guchhait, R.B. and U'Prichard, D.C., Assessment of the functional role of brain adrenergic neurons: chronic effects of phenylethanolamine-N-methyltransferase inhibitors and alpha adrenergic receptor antagonists on brain norepinephrine metabolism, J. Pharmacol. Exp. Ther., 230 (1984) 577-586.
23
NSL 06614
Adrenergic innervation of noradrenergic locus coeruleus neurons. A dual labeling immunocytochemical study in the rat Philippe Kachidian 1, Bernadette Astier z, Bernard Renaud z and Olivier Bosler 1 IEquipe de Neuromorphologie Fonctionnelle, Laboratoire de Neurobiologie, CNRS, Marseilles (France) andZ Laboratoire de Neuropharmacologie, CNRS, UMR 105, Facultk de Pharmacie, Universitk Claude Bernard, Lyons (France) (Received 5 August 1989; Revised version received 25 September 1989; Accepted 25 September 1989) Key words:
Adrenaline; Noradrenaline; Locus ceruleus; Dual immunocytochemistry; Electron micro-
scopy By means of dual immunocytochemistry, synaptic associations between adrenergic terminals and noradrenergic neurons were directly demonstrated in the rat locus ceruleus (LC). It could be estimated that every adrenergic afferent contacts at least one noradrenergic dendrite in the nucleus. An adrenergic innervation of non-noradrenergic targets was also evidenced. These data add to our knowledge on the synaptic circuitry by which activation of the adrenergic input could affect central mechanisms known to be influenced by LC neurons.
The locus ceruleus (LC) which provides extensive noradrenergic fiber projections throughout the central nervous system (CNS) (for reviews, see refs 2, 7, 19) is under the influence of a restricted afferent control system [6, 8]. One of the afferent inputs to the LC is composed of adrenergic fibers as identified by their immunoreactivity toward phenylethanolamine-N-methyl transferase (PNMT), the specific biosynthetic enzyme of adrenaline [4, 13]. This adrenergic projection which forms dense terminal plexuses in the whole LC [13] is derived via the medullary longitudinal axon bundle [4] from the ipsilateral C1 cellular group of the ventrolateral medulla and, to a lesser extent, from its C3 homologous in the dorsomedial medulla (23 also see ref. 12). That this adrenergic supply is of critical importance in regulating the output of LC noradrenergic neurons has been suggested by electrophysiological and biochemical pieces of evidence [1, 3, 5, 10, 11, 25] and by the recent electron microscopic demonstration that as many as 30% of all identified LC terminals consisted of PNMT-immunoreacCorrespondence." 0. Bosler, Laboratoire de Neurobiologie, CNRS, B.P. 71, 13402 Marseille c6dex 09, France. 0304-3940/90/$ 03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd.
25
tive axonal varicosities forming numerous synaptic junctions [18]. In the present paper, we report ultrastructural evidence from a dual labeling immunocytochemical study that LC adrenergic terminals innervate noradrenergic as well as non-noradrenergic targets. Adult male Sprague-Dawley rats were deeply anesthetized with chloral hydrate and perfused through the ascending aorta with 200 ml of 3.75% acroleine-2% paraformaldehyde mixtures in 0.12 M phosphate buffer and with 500 ml of 2% paraformaldehyde in the same buffer [20]. After overnight post-fixation of the brains, transverse sections through the entire LC were cut on a vibratome at 50/zm thickness. They were sequentially immunostained using a polyclonal rabbit anti-PNMT antiserum kindly provided by Dr. L. Denoroy (see ref. 15 for its production and specificity criteria), and a monoclonal anti-tyrosine hydroxylase (TH) antiserum from mouse-mouse hybrid cells (Boehringer Mannheim Biochemica). Dual labeling was achieved by combination of the peroxidase-antiperoxidase (PAP) method (PNMT staining) with a radioimmunocytochemical method which involved the use of a ~25Ilabeled secondary antiserum (TH labeling). Briefly, groups of three adjacent sections were pre-incubated in 1% sodium borohydride (30 min) and 0.3% hydrogen peroxide (30 min), and processed for PNMT staining as previously described [9]. After reaction with 0.05% 3,3'-diaminobenzidine (DAB) and 0.01% hydrogen peroxide added with 0.025% cobalt chloride and 0.02% nickel ammonium sulfate, they were incubated in a 1:300 dilution of the anti-TH antiserum (16-18 h) and in a solution of 1251-labeled anti-mouse IgG from sheep (spec. act.: 28-111 TBq/mmol, Amersham) diluted 1:50 from a 3700 MBq/ml radioactive solution. Two sections from each group were kept for light microscopic examination of PNMT staining and TH radiolabeling, respectively. Other sections which served as controls were run through the dual labeling sequence with omission of the anti-TH antiserum. As expected from a lack of cross-reactions between IgGs used in the first and second stages of the procedure, they did not show any nerve cell body 125I-labeiing at the light microscopic level. The remaining sections in each group were
Fig. 1. An axonal varicosity showing dual labeling for TH (silver grains) and PNMT (peroxidase staining) is directly apposed to an unlabeled soma. No synaptic membrane differentiation is visible on this plane of section. Fig. 2. A PNMT-positive terminal is synaptically linked to a TH-radiolabeled soma with an asymmetrical junctional complex. Fig. 3. A PNMT-positive terminal synaptically contacts both a TH-radiolabeled dendrite (arrow) and an unlabeled dendritic spine (*). Note that the spine is completely surrounded by its immunostained partner. Fig, 4. Other examples of PNMT-stained synaptic terminals surrounding dendritic spines (*). Figs. 5 and 6. Serial sections showing two PNMT/TH-positive terminals that simultaneously contact one same unlabeled dendrite with visible junctional complexes. Note that the superimposed TH-labeling of the terminals is only significant on Fig. 6. Another large, TH-labeled dendritic shaft is also contacted by one of the immunostained varicosities.
26
Fig. 7. A TH radiolabeled dendritic shaft, largely apposed to the basal lamina of a capillary (c) receives two synaptic terminals. One of these is PNMT-positiveand forms asymmetricaljunctional complexes.The other one does not show any significant staining and is involved in a symmetricaljunction I*). Fig. 8. The synaptic PNMT-stained terminal shown here is engaged in asymmetricaljunctional complexes with three dendritic elements, one of which is a large TH-radiolabeled proximal dendrite.
immersed for 1 h in 2% osmium tetroxide, stained "en bloc" with uranyl acetate and flat-embedded in Epon between siliconized glass slides. Ribbons of ultrathin sections were cut from the surface of selected sections from 3 rostrocaudal levels of the LC showing good peroxidase P N M T staining. After processing for radioautography by means of standard dipping techniques, they were exposed for 2 3 months and developed with K o d a k microdol X (1 2 min at 18 "C). PNMT-positive structures stained with the dense DAB endproduct were found to exclusively consist of axons and, most of all, of axonal varicosities whose ultrastructural features were similar to those previously reported by Milner et al. [18] (Figs. 1-8). Immunoreactive varicose profiles ranged from 0.7 to 2 / t m in transverse diameter, being then somewhat larger than those previously described in other parts of the brain [9, 17, 21]. They were frequently engaged in axo-dendritic synapses (Figs. 3--8) and only occasionally contacted somatic profiles (Figs. 1, 2). The junctional complexes were of either the symmetrical or the asymmetrical type and usually occupied relatively large areas of the apposed plasma membranes.
27 From a sampling of 151 positively stained varicosities, not less than 47% showed synaptic densities in single sections, which tends to indicate that the adrenergic innervation of the LC could well be entirely junctional. This is in contrast with the figure reported for another important monoaminergic afferent system to the LC, namely the serotoninergic input which has been previously shown to be essentially of the non-junctional type [16, 22]. Even a greater number of PNMT-immunoreactive profiles (about 85% for a total of 141) was found to be engaged in synaptic junctions by Milner et al. [18], but these authors have reported that the majority of the synaptic contacts was found in tissue that had been sectioned in the sagittal rather than the coronal plane. The fact that our sampling only involved coronal sections could then explain the discrepancy. In any event, the important synaptic investment of LC neurons by adrenergic afferents provides a morphological substrate for the previously reported inhibitory control of the LC by adrenergic neurons [3, 5, 10, 11, 25]. Surprisingly, only a small fraction (< 10%) of the PNMT-positive profiles examined also exhibited significant TH-radiolabeling (Figs. 1, 6), an observation which is in keeping with the initial results of Pickel et al. [22] who did not describe TH-immunoreactive terminals in the LC after staining with the PAP technique. This could be accounted for by a low sensitivity of electron microscopic immunocytochemical procedures for localizing axonal TH immunoreactivity in the LC, as also reported in the hypothalamus [24]. From 71 PNMT-positive synaptic terminals, 26 of which were examined on at least two consecutive sections, 46 innervated TH-immunolabeled targets. This indicates that, in addition to the noradrenergic population, neurons of a non-noradrenergic identity would also constitute important targets for the synaptic action of adrenaline in the LC. Both labeled and unlabeled post-synaptic processes essentially consisted of proximal dendritic shafts and medium to small branches which could be contacted by more than one immunoreactive terminal (Figs. 5, 6) and usually received other unreactive afferents (Fig. 7), as well as of distal branchlets and/or spines (Figs. 3, 4). Only one TH-positive (Fig. 2) and three TH-negative (Fig. 1) post-synaptic targets were identifed as nerve cell bodies, confirming that the adrenergic regulation of LC neurons essentially takes place at the level of the dendritic tree of these neurons [18]. The TH-immunolabeled processes often directly apposed each other and, in favorable planes of section, single PNMT-positive afferent fibers could be seen synapsing simultaneously upon two such processes that were contiguous or not. A given immunoreactive terminal could also synapse simultaneously on TH-labeled and unlabeled dendritic processes (Fig. 8). Given that not each PNMT-immunoreactive varicosity engaged in a synapse could be expected to show the differentiated active zone(s) on a single etectron microscopic section, the actual incidence of synaptic junctions established by adrenergic terminals in the LC was obviously underestimated. Then, the fact that as much as 65% of the PNMT-immunoreactive profiles were identified as presynaptic to TH-positive elements could indicate that every adrenergic afferent to the LC actually makes a synapse with a noradrenergic target. This figure would imply that the non-noradrenergic synaptic partners of the adrenergic afferents to the LC are always contacted
28
by these afferents together with at least one noradrenergic dendrite. Taken together, these data should be of potential value in determining the cellular mechanisms by which activation of the adrenergic input could affect central mechanisms known to be influenced by the LC (for reviews, see refs. 2, 8). As concerns the possible chemical identity of post-synaptic dendrites that did not show TH immunolabeling, it is interesting to note that medium-sized and small neurons immunopositive for GABA and presumed to constitute inhibitory interneurons have recently been described within the rat LC [14]. This study was supported by Grants 856003 (O.B.) and 856017 (B.R.) from the INSERM. We are particularly grateful to Dr. L. Denoroy for providing us with the anti-PNMT antiserum. We also thank Mrs. B. Besson for secretarial assistance. 1 Aghajanian, G.K. and Vandermaelen, C.P., Alphaz-adrenoceptor-mediatedhyperpolarization of locus coeruleus neurons: intracellular studies in vivo, Science, 215 (1982 1394~ 1396. 2 Amaral, D.G. and Sinnamon, H.M., The locus coeruleus: neurobiology of a central noradrenergic nucleus, Progr. Neurobiol., 9 (1977) 147 196. 3 Astier, B. and Aston-Jones, G., Electrophysiological evidence for medullary adrenergic inhibition of rat locus coeruleus, Soc. Neurosci. Abstr., 15 (1989) in press. 4 Astier, B., Kitahama, K., Denoroy, L., Jouvet, M. and Renaud, B., lmmunohistochemical evidence for the adrenergic medullary longitudinal bundle as a major ascending pathway to the locus coeruleus, Neurosci. Lett., 74 (1987) 132 138. 5 Astier, B., Kitahama, K., Denoroy, L., B6rod, A., Jouvet, M. and Renaud, B., Biochemical evidence for an interaction between adrenaline and noradrenaline neurons in the rat brainstem, Brain Res., 397 (1986) 333- 340. 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 Aston-Jones, G., Foote, S.L. and Bloom, F.E., Anatomy and physiology of locus coeruleus neurons: functional implications. In M. Ziegler and C. Lake (Eds.), Frontiers in Clinical Neuroscience: Vol. 2, Norepinephrine, Williams and Wilkins, Baltimore, 1984, pp. 92-116. 8 Aston-Jones, G., Shipley, M.T., Ennis, M., Williams, J.T. and Pieribone, V.A., Restricted afferent control of locus coeruleus neurons revealed by anatomic, physiologic and pharmacologic studies. In C.A. Marsden and D.J. Heal (Eds./, The Pharmacology of Noradrenaline in the Central Nervous System, Oxford University Press, Oxford, in press. 9 Bosler, O., Beaudet, A. and Denoroy, L., Electron-microscopic characterization of adrenergic axon terminals in the diencephalon of the rat, Cell Tissue Res., 248 (1987) 393 398. l0 Cedarbaum, J.M. and Aghajanian, G.K., Noradrenergic neurons of the locus coeruleus: inhibition by epinephrine and activation by the :~-antagonist piperoxane, Brain Res., 112 (1976) 413 419. 11 Engberg, G., Elam, M. and Svensson, T.H., Effects of adrenaline synthesis inhibition on brain noradrenaline neurons in locus coeruleus, Brain Res., 223 (1981) 49-58. 12 Guyenet, P.G. and Young, B.S., Projection of nucleus paragigantocellularis lateralis to locus coeruleus and other structures in rat, Brain Res., 406 (1987) 171-184. 13 H6kfelt, T., Johansson, O. and Goldstein, M., Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS: Part l, Elsevier, Amsterdam, 1984, pp. 157 276. 14 Ijima, K. and Ohtomo, K., Immunocytochemical study using a GABA antiserum for the demonstration of inhibitory neurons in the rat locus coeruleus, Am. J. Anat., 181 (1988) 43 52. 15 Kitahama, K., Pearson, J., Denoroy, L., Kopp, N., Ulrich, J., Maeda, T. and Jouvet, M., Adrenergic neurons in human brain demonstrated by immunocytochemistry with antibodies to phenylethanola-
29 mine-N-methyltransferase (PNMT). Discovery of a new group in the nucleus tractus solitarius, Neurosci. Lett., 53 (1985) 303 308. 16 L6ger, L. and Descarries, L., Serotonin nerve terminals in the locus coeruleus of adult rat: a radioautographic study, Brain Res., 145 (1978) 1 13. 17 Liposits Zs., Phelix, C. and Paull, W.K., Electron microscopic analysis of tyrosine hydroxylase, dopamine fl-hydroxylase and phenylethanolamine-N-methyltransferase immunoreactive innervation of the hypothalamic paraventricular nucleus in the rat, Histochemistry, 84 (1986) 105-I 20. 18 Milner, T.A., Abate, C., Reis, D.J. and Pickel, V.M., Ultrastructural localization of phenylethanotamine-N-methyltransferase-like immunoreactivity in the rat locus coeruleus, Brain Res.,(1989) 1 15. 19 Moore, R.Y. and Card, J.P., Noradrenaline-containing neuron systems. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2, Classical Transmitters in the CNS: part I, Elsevier, Amsterdam, 1984, pp. 123 156. 20 Pickel, V.M., Chart, J. and Milner, T.A., Autoradiographic detection of ~2Sl-secondary antiserum: a sensitive light and electron microsope labeling method compatible with peroxidase immunocytochemistry for dual localization of neuronal antigens, J. Histochem. Cytochem., 34 (1986) 707 718. 21 Pickel, V,M., Chan, J., Park, D.H., Joh, T.H. and Milner, T.A., Ultrastructural localization of phenylethanolamine N-methyltransferase in sensory and motor nuclei of the vagus nerve, J. Neurosci. Res., 15 (1986) 43~456. 22 Pickel, V.M., Joh, T.H. and Reis, D.J., A serotonergic innervation of noradrenergic neurons in nucleus locus coeruleus: demonstration by immunocytochemical localization of the transmitter specific enzymes tyrosine and tryptophan hydroxylase, Brain Res. 131 (1977) 197 214. 23 Pieribone, V.A.. Aston-Jones, G. and Bohn, M.C., Adrenergic and non-adrenergic neurons in the CI and C3 area project to locus coeruleus: a fluorescent double label study, Neurosci. Lett., 85 (1988) 297-303. 24 Piotte, M., Beaudet, A., Joh, T.H. and Brawer, J.R., The fine structural organization of tyrosine hydroxylase immunoreactive neurons in rat arcuate nucleus, J. Comp, Neurol., 239 (1985) 44-53. 25 Stolk, J.M., Vantini, G., Perry, B.D., Guchhait, R.B. and U'Prichard, D.C., Assessment of the functional role of brain adrenergic neurons: chronic effects of phenylethanolamine-N-methyltransferase inhibitors and alpha adrenergic receptor antagonists on brain norepinephrine metabolism, J. Pharmacol. Exp. Ther., 230 (1984) 577-586.
