Coexisting Neurotransmitters in Central Noradrenergic Neurons

Philip V. Holmes and Jacqueline N. Crawley

The noradrenergic neurons in the mammalian central nervous system (CNS) are localized in the nucleus locus coeruleus (LC), or A6 cell group, and in medullary nuclei designated A1, A2, A5, and A7. As diagrammed in , reviewed in ref. 14, and described further in chapters by Valentino and Aston-Jones and by Robbins and Everitt in this volume, the LC is a bilateral nucleus in the dorsal tegmentum, just ventral and lateral to the fourth ventricle. Each LC contains approximately 1500 densely packed neurons, with extensively branching axons forming five major noradrenergic tracts. The ascending projections of the LC comprise the dorsal noradrenergic bundle, the central gray dorsal longitudinal facsiculus, and the ventral tegmental–medial forebrain bundle, which project to the hypothalamus, thalamus, and telencephalon. Additional projections innervate the cerebellar cortex and the ventrolateral spinal cord. The A7 noradrenergic cell group is located in the lateral tegmental area ventral to the LC. The A1, A2, and A5 cell groups are situated more caudally in the medulla. The ascending projections of the A1, A2, A5, and A7 cell groups comprise the ventral noradrenergic bundle, innervating forebrain areas including the septum and hypothalamus. The LC and the A1 and A2 cell groups are the most extensively studied of the noradrenergic cell groups with respect to coexistence of transmitters and will therefore be the focus of the present review.

The noradrenergic neurons of the LC contain two neuropeptides, galanin (GAL) and neuropeptide Y (NPY) (26, 34). Neuropeptides are short sequences of amino acids, localized in high concentrations in neurons. The majority of neuropeptides were discovered in the brain during the 1970s, with the advent of radioimmunoassay and immunohistochemical techniques developed for applications to the nervous system (25). At least 50 neuropeptides have been conclusively identified in the mammalian CNS (25; also see General Overview of Neuropeptides). Functional studies are in progress for many peptides, to determine whether the criteria (i.e., neuronal synthesis, vesicular storage, release, specific receptors, effectors, and physiological actions) for a neurotransmitter are satisfied for a candidate neuropeptide (14). Following closely upon the discovery of the many putative peptide transmitters was the discovery of coexistence, the phenomenon of two or more neurotransmitters synthesized within the same neuron (25). Histochemical techniques, including double-labeling, tract-tracing, and lesion studies, demonstrate that neuropeptides are often localized within the same neuron as a well-known "classical" neurotransmitter (25; also see General Overview of Neuropeptides). The coexistence of GAL and NPY with norepinephrine in neurons of the LC provides an excellent model for investigating the functional interactions of a triple coexistence (see also Colocalization in Dopamine Neurons and General Overview of Neuropeptides).

Current histochemical evidence suggests that GAL is the predominant neuropeptide coexisting with norepinephrine (NE) in LC neurons in the rat brain (26,37; see also Galanin: A Neuropeptide with Important Central Nervous System Actions). GAL is a 29-amino-acid peptide that is regionally distributed throughout the mammalian CNS (37, 51). GAL apparently belongs to a distinct family of peptides, because it contains no significant sequence homologies with other known peptides (27, 58). The cDNA encoding this putative neurotransmitter has been cloned and sequenced (27, 58). GAL-like-immunoreactive (GAL-LI) and GAL mRNA-containing cell bodies are densely clustered in the hypothalamic paraventricular, arcuate, and dorsomedial nuclei, the bed nuclei of the stria terminalis, and the nucleus tractus solitarius in addition to the LC (37, 51). GAL coexists with other neurotransmitters, including cholinergic neurons of the basal forebrain and serotonergic neurons of the dorsal raphe (25, 26).

GAL-LI is present in approximately 80% of noradrenergic LC neurons in the rat (26). The coexistence of GAL with NE was demonstrated by immunocytochemical double-staining techniques using the synthetic enzymes tyrosine hydroxylase (26) and dopamine-b-hydroxylase (34) as markers for noradrenergic neurons. GAL-LI is not typically observed in the absence of dopamine-b-hydroxylase or tyrosine hydroxylase, suggesting that GAL is present only within NE-containing neurons of the LC (26, 34). GAL-LI is present throughout the LC, with a slightly higher density in the dorsal aspect of the LC (26, 34). The distribution of GAL mRNA detected by in situ hybridization matches well with the immunocytochemical distribution. GAL mRNA is present throughout the LC, showing highest levels in the dorsal LC (3, 15).

NPY coexists with NE in LC neurons of the rat, to a lesser extent than GAL. NPY is a 36-amino-acid peptide that is widely distributed throughout the mammalian CNS 19, 32; see Neuropeptide Y and Related Peptides). NPY-LI neurons are present in high concentrations in the rat cerebral cortex, hippocampus, and arcuate nucleus of the hypothalamus (19, 32, with densely clustered NPY-LI cell bodies in the brainstem catecholaminergic neurons (19, 47). NPY-like immunoreactivity is present in moderate concentrations in the LC (26, 47, 55), where double-labeling studies indicate that approximately 20–40% of noradrenergic neurons in the LC contain NPY-LI (19, 26). NPY, like GAL, appears to be present in the LC only in tyrosine hydroxylase (TH)-immunoreactive neurons (47).

diagrams the major noradrenergic pathways containing GAL and NPY in the rat brain. Combined immunocytochemistry and axonal transport studies have demonstrated that approximately 30% of GAL-LI LC neurons project to the hypothalamus, primarily terminating in the parvocellular subdivision of the paraventricular nucleus (PVN) of the hypothalamus (26, 34), corresponding closely with the location of preganglionic autonomic neurons (34). Electron microscopic analysis indicates that GAL-LI terminals in the PVN, containing dense-core vesicles characteristic of peptide storage, synapse predominantly on those parvocellular cells lacking in well-developed Golgi cisternae and secretory granules (34). Noradrenergic LC neurons containing GAL-LI also project to the medial thalamus, terminating on thalamic neurons that process and relay ascending nociceptive input from the spinothalamic tract ( 8,33). LC neurons containing GAL-LI and dopamine-b-hydroxylase-LI project to the cerebral cortex, hippocampus, and spinal cord (26, 37). Holets et al. (26) estimated that approximately 14% of GAL-LI noradrenergic neurons in the LC project to ipsilateral cerebral cortex and 3% project to either ipsilateral or contralateral spinal cord.

Few NPY-LI neurons project to the hypothalamus, and these projections do not terminate in the PVN (76), indicating that NPY in the hypothalamus arises from sources other than the LC (46, 47). NPY projections to thalamus appear to terminate exclusively in the medial division (33). The coexistence of NE and NPY in this coeruleo-thalamic projection has not been established by double-labeling techniques, but the extent of NPY-LI found in this pathway is almost 100% (33). NPY-LI fibers comprise a small percentage of the ipsilateral projection from LC to the entorhinal, medial, and lateral cortical areas and to the ventral hippocampus (26, 62). Approximately 3% of NPY-LI LC neurons project either ipsilaterally or contralaterally to the spinal cord (26).

GAL coexists with NE in neurons of the A1 cell group, but to a much lesser degree than in the LC, with approximately 15% of A1 neurons immunoreactive for both dopamine-b-hydroxylase and GAL (34). A high density of GAL-LI neurons are located in the A2 noradrenergic region of the caudal nucleus tractus solitarius (24, 34, 37, 51). However, double-labeling studies reveal that GAL does not coexist with NE in this cell group (34). GAL-immunoreactive fibers from the A1 group are widely scattered throughout the PVN, projecting to the parvocellular and magnocellular subdivisions of the PVN (34).

NPY-LI coexists with NE in a high proportion of neurons in the rostral and ventral portion of the A1 cell group. Estimates of the degree of coexistence range from 50% to 90% (19, 47). A1 cells expressing immunoreactivity for NPY without dopamine-b-hydroxylase are rare, suggesting that NPY is present only as a coexisting transmitter in the A1 cell group. Retrograde tracing and double-labeling studies indicate that about 60% of A1 noradrenergic neurons sending axons to the PVN contain NPY-LI, densest in the parvocellular division, which contains high levels of hypophyseal hormones (47). In the A2 cell group, approximately 10–30% of noradrenergic neurons contain NPY-LI (19, 47).

Neurons immunoreactive for atrial natriuretic factor, bombesin, calcitonin gene-related peptide, enkephalin, neurotensin, substance P, and vasoactive intestinal peptide have been observed in the LC (33, 55). However, there is no evidence to date that these peptides coexist with NE in LC neurons. Basic fibroblast growth factor (BDNF), a 24-amino-acid peptide, was recently discovered to coexist with NE in LC, A1, A5, and A7 perikarya of the rat, which will be of interest in light of the trophic actions of BDNF at other brain sites (10).

A neuropeptide coexisting with a "classical" neurotransmitter may act as a primary transmitter producing independent actions, may interact with the "classical" transmitter as a facilitatory or inhibitory modulator, or may serve a minor function only under specialized physiological conditions or during discrete developmental stages. Investigations of the role of a coexisting neuropeptide include: (a) studies of synthesis, using in situ hybridization and Northern blot analysis of messenger RNA (mRNA) levels; (b) studies of storage in synaptic vesicles, using electron microscopy combined with immunocytochemistry; (c) studies of release, using tissue slice preparations and in vivo microdialysis; (d) studies of receptors, using high-affinity binding assays and quantitative autoradiography; (e) studies of effector mechanisms, using biochemical assays for adenylate cyclase, phosphoinositide hydrolysis, potassium and chloride channels; (f) studies of neurophysiology, using tissue slice and in vivo recording techniques; and (g) studies of behavior, using animal behavior paradigms and central microinjections. Functional interactions of GAL and NPY with NE in the LC have not been extensively characterized, despite the fact that the percentage of LC neurons containing GAL is higher than any other reported coexistence (26, 37). Preliminary findings published to date are described below.

Coexisting neurotransmitters may be synthesized at the same rate, or their synthesis may be regulated differentially. Quantitative analysis of neuropeptide synthesis has been limited by the unavailability of selective agents that uniquely stimulate or inhibit synthesis of a specific peptide. Unlike NE, for example, in which synthesis can be inhibited through the rate-limiting synthetic enzyme TH, both GAL and NPY appear to be cleaved from larger, pre-pro peptides by relatively nonselective enzymes (27, 58). The methods in current use to quantitate peptide synthesis are Northern blot and in situ hybridization histochemistry for the mRNA specific to the peptide (see ). Employing quantitative in situ hybridization, GAL mRNA was found to increase concomitantly with TH mRNA in the LC, after treatment with the catecholamine-depleting agent reserpine (3). Treatments that did not increase TH mRNA in the LC (e.g., cold water swim stress) did not increase GAL mRNA in the LC (3). Desmethylimipramine blocked the reserpine-induced increase in TH mRNA, but only partially attenuated the increase in GAL mRNA, suggesting a selective interaction of the antidepressant desmethylimipramine with reserpine on the regulation of TH, but not GAL, gene expression (48).

Similarly, coexisting neurotransmitters may be released simultaneously or differentially, depending on neuronal activity. Analysis of the release of endogenous GAL and NPY in vivo has been limited by the sensitivity of existing radioimmunoassays. Investigations have begun on the ability of GAL and NPY to modulate NE release. GAL and NPY both inhibited 3H-NE release in tissue slices from rat hypothalamus, a terminal field region of the LC (22, 57). However, in vivo microdialysis studies in the paraventricular nucleus of the hypothalamus of freely moving rats reported that GAL microinjected into the paraventricular nucleus significantly increased NE levels in the microdialysate in either the presence or the absence of food (30). NPY increased NE release in this paradigm when food was present, but decreased NE release when food was absent (30). These contradictory data suggest that GAL has opposite actions on NE release in vitro versus in vivo.

Distinct receptors for NE, GAL, and NPY are well established (see Signal Transduction Pathways for Catecholamine Receptors and General Overview of Neuropeptides). GAL binds to a pertussis-toxin-sensitive, ADP-ribosylated Gi/Go protein-coupled high-affinity receptor (4, 31). GAL receptor activation may involve opening of an ATPsensitive or -insensitive potassium channel, reduction in intracellular calcium, stimulation of phosphatidyl inositol hydrolysis, or inhibition of adenylate cyclase activity (4,12, 41). Specific, high-affinity binding sites for 125I-GAL are regionally distributed in the locus coeruleus, hypothalamus, thalamus, amygdala, hippocampus, septum, striatum, and cerebral cortex, demonstrating a relatively good match between the distribution of GAL-LI terminals and GAL receptors (31, 38, 52). Two subtypes of high-affinity NPY receptors have been identified: (i) NPY-Y1, linked to intracellular calcium mobilization, and (ii) NPY-Y2, with high affinity for the C-terminal NPY 13–36 sequence (59). Both receptor subtypes are associated with inhibition of adenylate cyclase, and they may interact functionally with alpha-2-adrenergic receptors (59). Specific, high-affinity binding sites for 125I-NPY are regionally distributed in the cerebral cortex, hippocampus, thalamus, hypothalamus, septum, striatum, and brainstem, sites which receive NPY innervation (59).

GAL (10-9 to 10-7 M) inhibits the firing rate of locus coeruleus neurons in tissue slices from the rat hindbrain (49, 50). This inhibition is similar to that induced by alpha-adrenergic and m-opiate receptor agonists such as NE and [Met5]enkephalin, respectively. The mechanism for the inhibitory action of GAL on LC firing rate may involve an indirect interaction between GAL receptors and m-opiate receptors, but not between GAL receptors and alpha-adrenergic receptors. GAL-induced hyperpolarization was unaffected by the alpha-adrenergic antagonist, idazoxan, but was potentiated by the m-opiate receptor antagonist, naloxone (50). NPY depressed noradrenergic inhibitory postsynaptic potentials of LC neurons in rat pontine slices (20). In vivo studies of the neurophysiological actions of GAL and NPY in the LC have not been performed to date.

GAL and NPY both stimulate feeding (16, 17, 29, 54). Microinjected into the PVN, both GAL and NPY increase food consumption in satiated rats (17, 29, 54). GAL also increased food consumption when microinjected into the amygdala (16). These effects of GAL may be mediated through interactions with NE in terminal fields of LC projections. The ability of GAL to preferentially increase consumption of a high-fat diet was blocked by treatment with alpha-2-adrenergic receptor antagonists and by depletion of endogenous norepinephrine with alphamethyl-para-tyrosine, whereas NPY-induced feeding was independent of noradrenergic antagonists (29). GAL has been implicated in pain transmission. Intrathecally administered GAL potentiated the analgesic effects of morphine in unanesthetized rats (61). GAL directly antagonized the spinal reflex in decerebrate rats (63), suggesting that the mechanism of action of GAL is via primary sensory neurons, not through descending spinal projections from the LC. GAL administered into the lateral ventricle or into the ventral hippocampus has an inhibitory effect on spatial memory tasks in rats, including T-maze delayed alternation, delayed non-matching to sample, the sunburst maze, and the Morris water maze (35), 36, 42, 44). However, it is not known whether these performance deficits induced by GAL are mediated through cholinergic pathways, noradrenergic pathways, and/or other mechanisms (18). NPY has been reported to improve retention of a step-down passive avoidance task and a T-maze active avoidance task when administered intraventricularly to mice (21). The projections of NPY-containing neurons of the LC are one possible set of sites for this NPY action on a memory paradigm. NPY administered into the lateral ventricle of rats produced an anxiolytic-like action on punished conflict responding (23). Intraventricular administration of an NPY antisense oligodeoxynucleotide to rats produced an anxiolytic-like action on the elevated plus maze (60), suggesting a function for endogenous NPY, arising from the LC and/or from other sources, in the reduction of anxiety-related behaviors. GAL administered intraventricularly produced inhibitory effects on sexual behaviors in male rats (43), whereas GAL microinjected into the preoptic nucleus facilitated copulatory behavior in male rats (9). NPY administered into the third ventricle stimulated the release of luteinizing-hormone-releasing hormone (an NE-like effect), thereby initiating ovulation (45). Hypothalamic receptors innervated by GAL- and NPY-containing LC projections to the hypothalamus (26) could be mediating these sexual and reproductive behaviors.

Clinical investigations of GAL and NPY levels in cerebrospinal fluid and in postmortem brain tissue, including terminal fields for LC projections, have been conducted for several disease states. In cerebrospinal fluid of patients with anorexia and bulimia, NPY levels were elevated whereas GAL levels were normal (7). GAL levels were normal in cerebrospinal fluid from Alzheimer's disease patients, as compared to age-matched controls (7). In postmortem brain tissue from Alzheimer's patients, GAL-like immunoreactivity was found in the senile plaques (28). Tissue assays from Alzheimer's disease detected normal levels of GAL in cerebral cortex and hippocampal regions, in the same samples in which choline acetyltransferase levels were significantly reduced (5). Conversely, increased concentrations of GAL were found in the cholinergic cell body regions of the nucleus basalis of Meynert in Alzheimer's brains as compared to age-matched controls, whereas in the same samples NPY levels were normal and choline acetyltransferase was reduced (6). Histological analysis revealed that small, GAL-immunoreactive interneurons, which innervate the large cholinergic nucleus basalis of Meynert neurons in the human brain, and additional extrinsic GAL-LI axons, possibly from the LC, hyperinnervate the nucleus basalis of Meynert neurons in Alzheimer's brains (11, 39). This finding of greatly increased numbers of GAL-LI terminals on cholinergic basal forebrain neurons does not appear to be a nonspecific, or space-filling, artifact resulting from the progressive loss of cholinergic neurons, because GAL hyperinnervation did not occur in Down's syndrome, in which cholinergic neurons also degenerate (39).

The coexistence of GAL and NPY with NE in the LC, along with the inhibitory actions of both peptides to reduce the firing rate of LC neurons in vitro, suggests that these two coexisting neuropeptides may be inhibitory modulators, serving as inhibitory feedback to noradrenergic actions mediated by the LC. Noradrenergic pathways of the LC are thought to mediate aspects of depression, anxiety, withdrawal from drug addiction, arousal, attention, response to novelty, learning, memory, and feeding (1, 2, 13, 23, 40, 53, 60; also see Signal Transduction Pathways for Catecholamine Receptors, Pharmacology and Physiology of Central Noradrenergic Systems, Central Norepinephrine Neurons and Behavior, Physiological and Anatomical Determinants of Locus Coeruleus Discharge: Behavior and Clinical Implications, and Intracellular Messenger Pathways as Mediators of Neural Plasticity). The few functional studies performed to date and described above, which investigate the actions of GAL and NPY in relevant animal behavior models and clinical conditions, indicate that GAL has inhibitory effects on memory (35, 36, 42, 44) and hyperinnervates the cholinergic neurons remaining after degeneration in Alzheimer's disease (6, 11, 39). NPY may be elevated in a human feeding disorder (7), and it acts as an anxiolytic in an animal model (23). Both peptides stimulate feeding (16, 17, 29, 54). The recent availability of GAL antagonists (4), NPY antagonists (56), and NPY receptor antisense (60) will enable direct studies of the contributions of endogenous GAL and NPY to LC functions in animal behavior models. Future experiments will determine whether endogenous GAL and/or NPY are functionally important in these animal paradigms and whether GAL and/or NPY are implicated in the etiology of neuropsychiatric disorders. It is interesting to speculate that GAL and NPY agonists may be useful as treatments for anxiety, or withdrawal from drug addiction, and that GAL and NPY antagonists, alone or together with NE agonists and cholinergic therapies, may be useful as treatments for feeding disorders, depression, and Alzheimer's disease (see also Biological Markers in Alzheimer’s Disease, Psychopharmacology of Anorexia Nervosa, Bulimia Nervosa, and Binge Eating, and Basic Biological Overview of Eating Disorders).

published 2000