The Neurobiology of Neurotensin

Garth Bissette and Charles B. Nemeroff

INTRODUCTION

The previous version of this chapter, The Neurobiology of Neurotensin, reviewed the literature on the central nervous system effects of neurotensin (NT) through 1993. This CD-ROM version of the 4th Generation of Progress provides an opportunity to update this review for the five intervening years to the present. The maturation of molecular biologic techniques and the advances in imaging technology have made the cell biology of NT the research arena where the most progress has been made, particularly with regard to neurotensin’s receptors. Further evidence of neuroendocrine and neuroimmune effects of NT and new physiological effects during development and in some cancers have also been revealed. The roles of NT in human psychopathology and psychopharmacology continue to be active research fronts, but progress in these areas has been more gradual due to constraints of clinical research and the logistical complications such research entails. The strategy for this updated review is to build on the previous chapters sections with the most current references. As before, when a current and comprehensive review is available, this will be cited instead of the original research papers to allow interested readers access to a more thorough bibliography than can be provided here. The references from the previous version of this chapter are not repeated here, only the newer references are cited. Those interested in the prior references should consult the chapter published in the original Psychopharmacology, 4th Generation of Progress volume.

Discovery

Although the history of NT only covers some 25 years since Carraway and Leeman’s initial isolation of NT from bovine hypothalamic extracts, it has proven to be one of the more remarkable neuropeptide discoveries outside of the classic hypothalamic releasing factors. The tridecapeptide structure of NT (pGlu-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-Leu-OH) places it among the intermediate-sized neuropeptides. The midportion of the NT molecules, being highly enriched in basic amino acids, likely confers some of its unusual properties. Once synthetic NT became commercially available in 1973, a variety of NT-induced effects were discovered in addition to the early description of hypotension and cyanosis after NT was injected intravenously in mice. Most of the panoply of CNS effects of exogenously administered NT were discovered only after direct CNS application and, like the peripheral effects, require integrity of the C-terminal region of NT. Thus, one of the first CNS effects of NT was discovered during routine screening of a variety of peptides to determine their effects on the sedation and hypothermia produced by a fixed-dose of pentobarbital. Neurotensin was the only one of 48 neuropeptides tested that markedly potentiated the narcosis and hypothermia produced by pentobarbital, and this was associated with decreased metabolism of pentobarbital in brain, liver, and blood of mice. This effect was only observed after intracistemal (IC) injection of NT and led to the discovery of the potent hypothermic effect of NT in a variety of mammals which is accentuated by low ambient temperatures. The inability of drugs acting on cholinergic, serotonergic, and noradrenergic systems to alter this effect, in contrast to its potentiation by agents reducing dopaminergic neurotransmission, served as the initial impetus for launching an intensive investigation aimed at testing the hypothesis that NT acts as an endogenous neuroleptic .

Shortly after the discovery of the hypothermic effects of NT, several investigators noted the potent analgesic effects of centrally administered NT which, along with the hypothermic effects, were regionally mapped. This NT-induced analgesia is not blocked by opiate receptor antagonists. Direct injection of NT into the central periaqueductal gray (PAG) region produces a long-lasting analgesia which is associated with increases in the firing rate of he majority of PAG neurons to which it is applied. These, in turn, inhibit the response of spinal cord neurons to noxious stimuli using a neuronal circuit which includes cells of the rostral ventral medulla. Williams and Beitz reported induction of NT—neuromedin-N (NMN) mRNA expression in the caudal portion of the PAG in the acute pain phase followed by induction in adjacent midbrain tegmental nuclei in the adaptive phase of three models of chronic pain (arthritis, sciatic nerve ligation, and paw inflammation). More recently, the release of NT in the PAG of freely moving rats after administration of mu opioid receptor agonists has been measured using microdialysis techniques and a contribution to this NT release by opioid receptor inhibition of GABA circuits has been ruled out (89). The rostral ventromedial medulla has also been identified as a region where microinjection higher (nanomolar) doses of NT can block spinal cord transmission of noxious stimuli (97) but lower (picomolar) doses facilitate pain perception and can be potentiated by NT receptor antagonists (87).

The ontogeny of NT has been studied in the rat. Neurotensin immunoreactivity is detectable by gestational day 18 and generally exhibits increases in regional CNS concentrations during the first 10—20 days of life before returning to adult levels by 60 days of age. Aged rats (22 months) have greatly reduced concentrations of NT in the striatum and substantia nigra compared to 3-month-old controls, whereas the relative concentrations in the nucleus accumbens of aged rats are only mildly decreased. Evidence for a trophic effect of NT in CNS growth and development include the fact that NT and NT receptors have been reported to be transiently expressed in some rat brain regions, such as hippocampus (10), during development, that NT receptors are present in meningioma’s, and that NT is found in high concentrations in cerebrospinal fluid (CSF) from human infants. Interestingly, NT-like immunoreactivity has been described in all species of mammalian CNS examined thus far, and C-terminally directed NT antisera have shown positive immunoreactivity to extracts of lobster and all other animal phyla, including species of Porifera and Protozoa. A recent paper has extended the phyolgenetic distribution of NT in the gut to include the Burmese python (20) and the role of NT in the growth of cells in prostate (82), pancreatic (46) (68), liver (23) and lung (83) cancer has recently begun to be appreciated and exploited in treatments (58).

NMN is a hexapeptide (H-Lys-Ile-Pro-Tyr-Ile-Leu-OH) which has significant C-terminal sequence similarity with NT and is contained in the pro-hormone precursor NT molecule, where it is separated from the NT sequence by a single pair of dibasic amino acids. NMN shares many properties with NT and binds to the NT receptor with similar affinity, but is less potent and apparently more easily degraded than NT. The post-translational processing of NMN exhibits some regional differences, but, thus far, agents that increase NT synthesis also increase NMN synthesis. Another neuropeptide derived from extracts of frog skin that shares some sequence homology with NT is xenopsin (pGlu-Gly-Lys-Arg-Pro-Trp-Ile-Leu-OH), and two NT-related, structurally homologous peptides

(pGlu-Leu-His-Val-Asn-Lys-Ala-Arg-Arg-ProTyr-Ile-Leu-OH) and (H-Lys-Asn-Pro-Tyr-Ile-Leu-OH) from chicken intestinal extracts have also been described. More recently, a new, 25 amino acid gut peptide named xenin has been isolated from human gastric mucosa and shown to bind to gut NT receptors (34) thus joining the xenopsin/neurotensin gene family.

Molecular Biology

The tools of molecular biology have now been applied to NT systems in the CNS. Both the NT—NMN prohormone and the high and low affinity NT—NMN receptor genes have now been successfully cloned and sequenced. The rat gene for NT spans 10.2 kb of DNA and consists of four exon sequences separated by three intervening introns. Exon 4 contains one copy of the sequence for NT and one for NMN, separated by a sequence coding for a single pair of basic amino acids. Production of the 170-amino acid pro-hormone from either 1.5- or 1.0-kb messenger RNAs is regulated by elements that are cis-responsive to nerve growth factor, cyclic nucleotides, glucocorticoids, and, possibly, lithium, upstream of the pro-hormone sequence. This NT—NMN gene sequence is highly conserved, as comparison between dog, bovine, rat, and human species has demonstrated. Expression of NT mRNA in neurons where NT message was previously absent is seen after either pharmacologic stimuli such as antipsychotic drugs or exposure to chronic pain. Other known inducers of NT—NMN messenger RNA are caffeine, staurosporine, sigma receptor antagonists and osmotic stimulation. Administration of antisense oligonucleotide to c-fos in the striatum prevents the caffeine-induced increase in NT mRNA expression (91). However, antisense to D3 dopamine receptor mRNA can reduce the expression of NT mRNA in the nucleus accumbens, indicating potential regulation of accumbens NT by this dopamine receptor (92). Drugs such as haloperidol that induce NT synthesis also induce NMN in the same regions and in similar proportions (see Fig. 1). The reported regional disparities in the molar ratio of NT to NMN are thought to be caused by post-translational processing differences in prohormone cleavage, as well as by regional product degradation differences. The promoter region of the NT/NMN gene has been shown to be activated by Src kinase (8) and the prohormone convertase PC-5 has been shown to process cleavage of the NT/NMN prohormone in some cell cultures, including human colon cancer cells (73) and rat pheochromocytoma PC-12 cells (9), whereas the central nervous system processing of the pro-hormone has been credited to PC-1 (96) and PC-2 (64).

The rat NT/NMN receptor is comprised of a 424-amino-acid protein with the seven-membrane-spanning domains characteristic of G-protein-coupled receptors. It has a somewhat lower-than-expected sequence homology (18—24%) with other members of this receptor superfamily . The human NT/NMN receptor has also now been cloned and found to contain 418 amino acids with 84% sequence homology to the rat NT—NMN receptor and a tetranucleotide repeat polymorphism of the human NT receptor gene has been identified (49,50). The distribution of this rat NT receptor mRNA has been mapped with quantitative RT-PCR (57) and two possible initiation sites for NT receptor mRNA expression leading to expression of 47 and 44 kDa proteins have been shown to be potentially regulated by the 5’ untranslated end of the mRNA (16). Both the aforementioned rat and human NT receptors are thought to represent the high-affinity form of the NT receptor (NTR1) that is revealed in the presence of the antihistamine agent, levocabastine, which binds to an apparent low-affinity (NTR2) site in brain membrane receptor preparations. The levocabastine-sensitive NT receptor has recently been cloned from rat (18) and mouse (55) and it’s mRNA mapped by in situ hybridization to hypothalamic, forebrain and brainstem structures (80). Recent evidence using antisense oligodeoxynucleotides in mice indicates that the levocabastine sensitive, low affinity NTR-2 receptor mediates the analgesia seen after central injection of NT (24). Both the rat and human NT—NMN-receptor clones exhibit sequence similarities with other G-protein receptors that use inositol phosphate as a second messenger transducer. The rat NT-receptor sequence, when expressed in oocytes, is linked to this second messenger, but the human NT-receptor sequence has multiple base-pair differences compared to the rat in the third cytoplasmic loop sequence known to mediate second messenger linkage. When expressed in Chinese hamster ovary cells, stimulation of the human NT receptor activates two isoforms of the mitogen-activated protein kinase through effects on both pertussis-sensitive and pertussis-insensitive G-proteins and protein kinase C (66). Recently, the alanine residue at the 302 position in the third cytoplasmic domain of the NT and cholecystokinin B receptor was required for full activation by NT, but not cholecystokinin, of the associated G-protein and phospholipase C response to receptor binding (102). Evidence for NT receptor linkage to a G protein which stimulates adenylate cyclase in the rat ventral tegmental area (VTA) is also available (see also Cholinergic Transduction, Signal Transduction Pathways for Catecholamine Receptors, and Serotonin Receptors: Signal Transduction Pathways) and evidence that this signal transduction mechanism mediates part of the high affinity NT receptor population has recently been obtained (41). The existence of still further variants of NT receptors has been adumbrated by evidence that certain analogs of NT possess hypothermic effects without potency in analgesic responses (93) and analgesic responses are not blocked by the potent histamine antagonist, pyrilamine, although other less potent histamine antagonists such as levocabastine and diphenhydramine effectively antagonize NT-induced analgesia (94). Biphasic regulation of NT receptor mRNA by a NT agonist in colonic adenocarcinoma cells has been reported with an initial increase in NT receptor mRNA synthesis after internalization of the NT-receptor complex followed by a decrease in NT receptor mRNA half-life after prolonged exposure to the agonist (88). Prolonged (15 days) exposure to the NT receptor antagonist SR48692 (1.0 mg/kg, i.p.) has been shown to greatly increase the expression NT receptor mRNA in regions where it is found in basal conditions and to reveal several regions where the receptor mRNA is normally undetectable (6). The Arg residues at position 327 and 328 in the first extracellular loop of the NT receptor have been shown to be crucial for binding NT (14) .

Localization

Many immunohistochemical and radioimmunoassay distribution studies in a variety of species have revealed that NT is distributed heterogeneously among the various anatomic regions of the brain. More recently, in situ hybridization studies of NT prohormone mRNA have been added to the growing list of techniques used to visualize endogenous NT neurons. Highest brain levels of NT are found in the hypothalamus, with the greatest concentration in the posterior hypothalamus and mammillary body regions. The substantia nigra, ventral tegmental area, and central nucleus of the amygdala, as well as the dorsal hippocampus, nucleus accumbens, septum, and globus pallidus, also contain significant amounts of NT in micro-punched rat brain regions. Neuronal cell bodies containing NT are found in most of these regions, along with high densities of NT-containing nerve terminals. Interestingly, few NT cell bodies are found in the striatum of normal, non-drug-treated animals. A particularly wide band of NT cells arises from the medial anterior hypothalamus and courses through the medial preoptic area to the bed nucleus of the stria terminalis/diagonal band area and on to the lateral aspects of the anterior septum. Immunohistochemical studies have also shown NT and NT mRNA to be present within the anterior pituitary gland. The distribution of NT-like immunoreactivity in the brain of chickens has been reported recently by two groups, one of which also compared chickens with pigeons. In one report, chicken NT is reminiscent of the distribution in mammals, but with immunoreactive cell bodies mainly restricted to the hypothalamus and NT-containing fibers in the limbic forebrain and midbrain (29). In the other report (4), NT cell bodies are also seen in dorsolateral and entorhinal cortex of chickens and major differences with pigeons noted in the median eminence, optic tectum and cerebellum where NT was present in pigeons but not in chickens. The sensitivity and specificity of the different NT antisera used in such studies is most likely the reason for the reported differences in NT distribution in the chicken.

Hokfelt and colleagues originally described NT colocalization in tyrosine hydroxylase-containing dopamine (DA) neurons of the VTA and the arcuate nucleus of the hypothalamus. Since then, NT in rat brain has been shown to be colocalized with cholecystokinin in VTA DA neurons, within subsets of corticotropin-releasing factor (CRF) neurons in the hypothalamus, with growth-hormone-releasing factor in the arcuate nucleus of the hypothalamus , and with calcitonin gene-related peptide in the bed nucleus of the stria terminalis as well as in the central nucleus of the amygdala. In the central nucleus of the rat amygdala, most NT is not co-localized in dopamine afferents while most of the NT afferents in the rat basal amygdalar complex contain dopamine (3). In the cat, low densities of NT-containing nerve fibers were observed in the anterior amygdala with low and moderate fiber density in the lateral and medial parts of the central nucleus of the amygdala (53). The most recent evidence indicates that NT is colocalized with somatostatin and substance P in ventrolateral hypothalamic neurons bearing progesterone receptors (25). In an elegant study using acute administration of haloperidol to maximize NT/NMN mRNA presence, Emson and colleagues demonstrated that within the rat striatum, NT/NMN mRNA was often co-localized within neurons also containing mRNA’s for preproenkephalin and the dopamine and adenylate cyclase phosphoprotein-32 (5). The caveat of species-specificity in these findings is emphasized by the apparent lack of NT colocalization within human midbrain DA neurons that project to the cerebral cortex. The anatomic colocalization of NT and DA has considerable implications for the understanding of NT interactions with DA systems. Although NT / DA interactions have received the most attention, the interaction of NT with serotonin (5-HT) neurons has recently become a target of investigation due to the abundent NT-containing neurons, fibers and terminals of the raphe complex, the presence of NTR1 receptors in raphe nuclei and the concentration dependent increase in firing rate of a sub-population of predominently ventral raphe serotonin neurons to NT iontophoresis (see 46 for review). A functional role for NT in the raphe has been postulated due to the ability of intraventricular NT to block the stress-induced activation of serotonergic raphe neurons in rats exposed to random inescapable sound (23).

Several putative NT pathways have been described in rat brain. Projections from cell bodies in the arcuate nucleus to the median eminence provide NT to the hypothalamic hypophyseal portal system. Projections from NT cell bodies in the central nucleus of the amygdala to the bed nucleus of the stria terminalis and midbrain parabrachial nucleus have also been mapped. A hippocampal NT pathway from the subiculum to the alveus and fimbria regions and from the dorsal hippocampus to the adjacent cortex have been described. A projection from the nucleus tractus solitarii to the nucleus accumbens has now been reported. Other NT pathways include a projection from piriform cortex to the anterior olfactory nucleus and from cell bodies in the PAG region and parabrachial nucleus to terminals in the nucleus raphe magnus. Beaudet’s group has demonstrated that the NT innervation of cholinergic neurons in the anterior aspects of the substantia innominata arises from the lateral septum, medial preoptic area, anterior hypothalamic region, nucleus accumbens and the rostral basal forebrain with the posterior substantia innominata receiving NT projections from the nucleus accumbens, lateral septum, hypothalamus, bed nucleus of the stria terminalis, ventral tegmental area and raphe complex (59). The caudal aspects of the basal forebrain chonlinergic neurons was shown to have more NT-containing axonal varicosities in close approximation to cholinergic perikarya than are present in the rostral regions of the basal forebrain although this asymmetry is not reflected in NT receptor distribution or cholinergic neuronal responses to applied NT (60). Based upon this distribution and the neuronal burst generating properties of NT applied to cholinergic neurons in this region, these researchers postulate a role for NT in the regulation of hippocampal theta activity through septal projections. Using cultured newborn rat nucleus basalis neurons, NT has been shown to inhibit both low- and high-voltage-activated calcium channels and this mechanism may underly the hippocampal regulation (54). A potential projection from the vagal nucleus to the parabrachial nucleus has been postulated based upon the continued release of NT in the parabrachial nucleus after up to two hours of continuous vagal nerve stimulation (77). When NT is microinjected into the caudal region of the nucleus of the solitary tract in anesthetized rats decreases in heart rate and arterial pressure are produced indicating a modulatory role for NT in central cardiovascular relex pathways (19). Further evidence of possible autonomic effects of NT are derived from it’s location alongside CRF in neurons of the external cuneate nucleus of the gerbil (48).

However, the most extensively characterized NT pathway is a mesolimbicocortical projection of NT-containing neurons from the VTA to the frontal cortex and nucleus accumbens. The former target contains an appreciable component of colocalized NT and dopamine and some cholecystokinin, whereas the latter is thought to be predominantly NT without dopamine. Recently improved techniques have allowed electron microscopic visualization of NT and tyrosine hydroxylase co-localization in cell bodies and nerve terminal which should greatly increase the resolution and anatomic detail of dopamine and NT containing neuronal pathways (52). Rats subjected to cold water restraint form gastric ulcers and exogenous NT administration into the nucleus accumbens attenuates this gastric injury. The concentration of endogenous NT and NT mRNA in the nucleus accumbens is transiently reduced by cold water restraint while NT receptor number and receptor mRNA were decreased in rats (105).

The pattern of distribution of NT receptors depends partly upon whether [³H]-NT or [¹²⁵I]-NT is used, though NT receptor distribution generally agrees well with immunohistochemical localization of NT-containing nerve terminals. Autoradiographic studies of brain NT binding sites have been conducted in rat, pigeon, guinea pig, monkey, and human brain as well as in transgenic mice over-expressing the gene for superoxide dismutase. High concentrations of putative NT receptors are found in the substantia nigra (pars compacta) and VTA and their terminal projections to the caudate, nucleus accumbens, and cortex. NT numbers are greatly diminished by 6-hydroxy-dopamine (6-OHDA) destruction of the DA neurons, where these receptors are predominately located. However, many NT terminals synapse on non-DA neurons in these regions. Moderate densities of NT receptors are also seen in the other components of the striatum, ventral hippocampus, PAG matter, superior colliculus, and dorsal raphe nucleus. Cerebral cortex in nonhuman primates and in humans contains higher densities of NT binding sites in the deep cortical layers than does rat brain. Recent evidence indicates that after NT binds to the NT receptor, a portion of the resulting receptor—NT complexes is internalized through the post-synaptic membrane and is subsequently transported to the postsynaptic neuron’s nucleus, where it may regulate tyrosine hydroxylase activation via protein kinase C. Internalization of the NT/receptor complex has now been shown to down-regulate NT receptor synthesis in cultured neurons and re-appearance of NT receptors on the cell surface has been demonstrated to require new receptor synthesis rather than recycling (42). The postulated tyrosine hydroxylase regulation by NT has been confirmed by demonstration of NT immunoreactivity in 16% of hypothalamic tuberinfundibular tyrosine hydroxylase-containing neurons which in turn represent over half of the neurons containing NT (see Figure 2). About 6% of the total number of tyrosine hydroxylase containing neurons had NT receptors and a population of apparent NT autoreceptors was also reported (7). These data provide other mechanisms by which NT can regulate DA neurons. This process of receptor internalization has now been shown by confocal microscopy to be a component of the high-affinity NT receptor response on neurons whereas low affinity NT receptors on astroglia do not internalize after binding NT (63). There is good agreement that the predominant second messenger systems linked to the activated NT receptor are the inisitol triphosphate/diacyl glycerol and the cyclic guanosine monophosphate (cGMP) signal transduction systems. The rapid desensitization of NT receptors after continued application of NT has been demonstrated in a neuroblastoma cell line by Richelson and co-workers. This phenomena is apparently due to NT-receptor down-regulation associated with a decrease in receptor number without changes in affinity.

Peptidases are the enzymes directed toward specific amino acid sequences that degrade neuropeptides. Several such enzymes have been identified that cleave NT, including endopeptidases 24.11, 24.15, and 24.16. Recent evidence indicates that 24.16 and 24.15 are the predominate NT degrading peptidases in the CNS (99,100). Their biological significance can be appreciated by the profound potentiation of CNS effects of NT observed after administration of specific peptidase inhibitors. Thiorphan, which inhibits endopeptidase 24.11 cleavage of NT on either side of the Tyr¹¹ moiety, and bestatin, which inhibits the aminopeptidase that cleaves NMN between Lys¹ and Ile², both potentiate the hypothermia and analgesia observed after intracerebral injection of NT or NMN, respectively. Differential processing of the NT—NMN precursor protein by trypsin-like enzymes have been implicated in the regional differences in NT/NMN ratios reported by various research groups, while the reduced potency of NMN in evoking NT-like responses is thought to be due to its relatively unprotected amino terminus. The hypothalamic concentrations of NT and NMN vary in a biphasic rhythm across the year in rats without exposure to photoperiod or temperature cues for season and the peaks and troughs of NT and NMN are not in register (see figure 3), indicating circannual changes in pro-hormone processing or peptidase degradation (13).

Endocrine

The effects of NT on the release of anterior pituitary hormones depend upon whether NT is injected peripherally or centrally, whether NT is applied alone or after stimulation by other peptides or pharmacologic agents, whether the experimental animal is awake or anesthetized, male or female, and, if female, whether they are ovariectomized (OVX) and/or whether they are estrogen-primed. Thus, NT injected into the third ventricle of conscious rats reduces plasma concentrations of prolactin and luteinizing hormone, increases plasma growth hormone, and has no effect on thyrotropin (TSH) or follicle-stimulating hormone plasma levels. This direct hypothalamic effect of NT in reducing prolactin secretion is thought to be mediated by DA release into the portal system, because blockade of dopamine receptors in lactotrophs prevents NT from acting. However, NT, when injected intravenously in conscious rats and acting directly on the adenohypophsis, elevates plasma TSH and prolactin while not having much effect on the other pituitary hormones. Intravenous injection of antiserum to NT in conscious, OVX female rats, which would remove endogenous NT from pituitary circulation, decreases plasma levels of prolactin and TSH and increases the amount of growth hormone in plasma. Thus NT may play a physiologic role in the secretory regulation of these anterior pituitary hormones. The regulatory neuroendocrine effects of NT occur at both the hypothalamic and pituitary levels and this fascinating field has been recently reviewed by (71)

Alexander and colleagues have shown that NT neurons in the medial preoptic anterior hypothalamus contain estrogen receptors and that NT may physiologically regulate the preovulatory surge of luteinizing hormone in female rats. Direct injection of NT antiserum to OVX, estrogen-primed rats, in the afternoon before a proestrus-like surge in the secretion of luteinizing hormone, prevented this surge. Watters and Dorsa have demonstrated that the effects of estrogen on NT/NMN gene expression are mediated via a cyclic AMP/protein kinase A dependent mechanism (103). Others have shown these neurons to be devoid of tyrosine hydroxylase, cholecystokinin, or luteinizinghormone-releasing hormone.

More recently, NT has been shown to have regulatory effects on CRF and ACTH release from adrenal medulla (56) and the ACTH release by NT was blocked by alpha-helical CRF, indicating NT release of CRF was responsible for the subsequent ACTH release. Blockade of NT receptors with SR48692 delivered into the paraventricular nucleus of the hypothalamus blocks the diurnal increase in plasma ACTH and corticosterone in the evening phase of the cycle (74) and blocks the stress-induced increase in ACTH but not corticosterone with little effect on basal release of either ACTH or corticosterone (62). The effects of centrally administered NT on corticosterone release are mediated through induction of the immediate early gene c-fos in the basolateral and central nucleus of the amygdala (47) and could be blocked by the NT receptor antagonist, SR48692. This was in contrast to the hypothermic effect which was correlated with activation of the immediate early gene, zif268, in the paraventricular nucleus of the hypothalamus and was not blocked by the antagonist. Further interactions of CRF and NT in gut have been postulated in the colonic response to immobilization stress in rats (67).

Interactions of NT systems with the immune system have also been documented. Cultured human thymic epithelial cells contain full-length NT and NT fragments and co-eluted from an affinity column that retained major histocompatability complex class 1-related proteins (98), indicating that NT may bind to these molecules. In pregnant rats exposed to a microwave heat source, NT concentrations are increased in the hypothalamus along with beta endorphin and ACTH while NT concentration decreased in the nucleus accumbens along with splenic natural killer cell activity (61). These data indicate that both central and peripheral sources of NT are modulated by conditions that alter immune system parameters.

Interactions with, Centrally Active Agents

Interactions of NT with CNS-active drugs can be divided into two components: (i) effects of exogenous NT on the actions of certain drugs and (ii) the effects of certain classes of drugs on endogenous NT systems. We have already described the ability of centrally administered NT to potentiate the sedation and hypothermia induced by pentobarbital in rodents. This potentiation of sedative effects extends to other CNS depressants such as alcohol in both mice and rats. While reduction of the rate of pentobarbital metabolism may be partly responsible for the effects of NT on barbiturate, this is not the case for ethanol. The site within the CNS where NT elicits these effects remains unknown. However, Erwin and Jones have reported that, in mice bred for long (LS) or short (SS) sleep after ethanol treatment, SS mice had higher densities of high-affinity NT receptors in the entorhinal and frontal cortex and striatum and exhibited greater sensitivity to the NT-induced potentiation of ethanol narcosis. Several quantitative trait gene locations have now been associated with the locomotor (28) and hypnotic and hypothermic (27) effects of alcohol in these two strains. This reseach group has also bred rats for high and low sensitivity to alcohol and have demonstrated ventral midbrain increases in high affinity NT receptors in high sensitivity males (26). This line of research has been extended to other mouse strains selectively bred for voluntary ethanol consumption and the frontal cortex concentrations of low affinity NT receptors were increased in the strains with increased ethanol consumption (35).

This potentiation of sedatives, along with the hypothermic and analgesic effects of centrally administered NT, led to our original "endogenous neuroleptic" hypothesis. Soon after this hypothesis was first promulgated, Govoni and colleagues demonstrated the ability of clinically efficacious antipsychotic drugs to increase the concentrations of endogenous NT in brain regions where projections of midbrain DA neurons terminated—that is, the nucleus accumbens, caudate nucleus, and olfactory tubercles. This effect is seen within 16 hr after a single injection of haloperidol and does not exhibit tachyphalaxis to repeated doses over several months. A wide variety of classical and atypical antipsychotic drugs have now been studied. Classical antipsychotics such as haloperidol, pimozide, chloropromazine, and butaclamol increase NT concentrations in both the caudate nucleus and nucleus accumbens. In contrast, atypical antipsychotic drugs, such as clozapine, which are devoid of extrapyramidal side effects or tardive dyskinesia liability, produce increases in NT concentrations in the nucleus accumbens, but not in the striatum. This field has recently been reviewed (51). Thus NT increases in the striatum may predict extrapyramidal liability of antipsychotic drugs. Another recent dichotomy between the effects of haloperidol and clozapine is the report that haloperidol, but not clozapine, increases NT-receptor mRNA levels in the substantia nigra. Recent work has shown that NT mRNA is increased by antipsychotic drugs such as haloperidol within 2 hr after a single dose in the dorsolateral striatum and the shell region of the nucleus accumbens of rats. In contrast, neurotensin mRNA expression is increased only in the nucleus accumbens after clozapine treatment. The dorsolateral striatal NT—NMN mRNA increase is immediately preceded by c-fos gene activation, but this is not seen in the accumbal shell region’s NT—NMN mRNA induction (see Fig. 4). Recently, both Robertson (personal communication) and Dorsa (personal communication) have shown that injection of antisense probes to c-fos blocks the antipsychotic drug-induced increases in NT synthesis in adults. However, others have shown some decrease in haloperidol-induced NT production in the striatum of 10 and 15 day old rats with c-fos genetic knockouts but report no effect on haloperidol NT responses in c-fos knockout adults (85). The increase in NT synthesis following the administration of antipsychotic drugs is not altered by blockade or stimulation of cholinergic muscarinic receptors and the striatal induction of NT/NMN mRNA by haloperidol is not effectively blocked by the irreversible monoaminergic receptor antagonis EEDQ (2). Dorsa and colleagues have also shown that gene knockout of protein kinase A in mice prevents the catalepsy and NT mRNA induction in the striatum seen after haloperidol (1) and untreated protein kinase A knockout mice do not exhibit alterations of endogenous NT concentrations (G. Bissette, personal communication). In a recent study using microdialysis, haloperidol increased the extracellular concentration of NT in the nucleus accumbens but decreased in the caudate region surrounding the microdialysis probe, although tissue concentrations of NT were increased in both regions 24 hours after the single haloperidol injection. The concentration of NT in the ventral tegmental area decreased while substantia nigra NT content was unchanged by the haloperidol injection (44). Thus, antipsychotic drugs increase NT tissue concentrations by increasing the synthesis of NT relative to effects on release and degradation, though these processes require further study. The steroisomers of butaclamol, (+)- and (—)-butaclamol, which share all effects on adrenergic, serotonergic, and histamninergic systems, have also been investigated for NT-system interactions. Only (+)-butaclamol is able to increase NT concentrations in the caudate nucleus and nucleus accumbens, and only the (+)-isomer blocks D₂ receptors and has antipsychotic efficacy. A variety of other drugs have also been studied for their effects on NT systems. Tricyclic antidepressants, benzodiazepines, and antihistamines do not alter NT concentrations in the striatum or nucleus accumbens.

Surprisingly, the psychostimulant indirect DA agonists, such as cocaine, methylphenidate, and amphetamine, are able to reverse NT-induced hypothermia, yet in high doses these drugs increase NT concentrations in DA terminal regions. Within 8—24 hr after several high doses of cocaine or after chronic cocaine administration in moderate doses, NT production is apparently induced in these DA terminal regions. This results in increased NT concentrations and decreased NT-receptor binding in the VTA, during both chronic treatment and 10 days after withdrawal of chronic cocaine. The latest evidence along these lines indicates that 10 days of cocaine injections at moderate doses (15 mg/kg, i.p) induces NT mRNA expression in the shell of the nucleus accumbens and posterior dorsomedial and ventrolateral areas of the striatum (12). Other work in our laboratory has extended this effect to the relatively specific D₂-receptor agonist, quinelerone, which increased NT concentrations in caudate and nucleus accumbens, similar to haloperidol. This avenue of investigation has been further extended using the response of NT systems to low and high doses of methamphetamine with the finding that low doses increase NT release by a D2 receptor mediated mechanism relative to NT synthesis while high doses increase NT synthesis in a D1 receptor mediated but does not induce release of the newly synthesized NT (39). In a further development, Rompre has recently shown that repeated stimulation of NT receptors through intraventricular administration of NT produces sensitization to the behavioral effects of amphetamine (70) and that pre-frontal cortex may be the site where this occurs (69). The ability of centrally administered NT to release dopamine from mesolimbic terminal regions has been extensively documented using increasingly elegant techniques, such as microdialysis and in vivo chronoamperometry, and is thought to be due to direct NT effects on mesencephalic DA neurons. The ability of these drugs to release NT in colocalized prefrontal cortical DA terminal regions has recently been documented by During and colleagues, while Rivest and co-workers reported that the effect of NT on dopamine release is of greater magnitude in the nucleus accumbens than the striatal response. More recently, NT has been shown to release DA above basal release levels in the prefrontal cortex, nucleus accumbens and striatum and to significantly enhance the release of DA by the glutamate receptor agonist, NMDA in these brain regions (104).

However, using the specific D₂ and D₁ DA-receptor agonists and antagonists, differential regional CNS effects on NT systems have been observed. Singh and colleagues reported that D₂ antagonists increase striatal NT concentrations whereas D₂ agonists decrease NT concentrations in this region. Both nucleus accumbens and striatal NT concentrations are increased by D₁ agonists, and N-methyl-D-aspartate (NMDA) receptor blockade completely blocks this D₁ agonist effect (101) while blockade of NT receptors with SR48692 does not block the vacuous chewing response of rats treated with the D1 agonist SKF38393 (90). Taylor and colleagues have reported that D₁ antagonists decrease NT concentrations in the striatum. Fuxe and co-workers ascribe these striatal effects to an NT—DA-receptor interaction that reduces the affinity of D₂ pre- and postsynaptic receptors while enhancing D₁-receptor sensitivity; cholecystokinin exerts a synergistic effect. Recent theories have postulated that NT and dopamine may form a chemical complex, thereby attenuating dopamine’s ability to activate dopamine receptors. Attempts to replicate the initial evidence for this complex have not been successful, however.

Haloperidol not only has a high affinity for D₂ DA receptors, but also has a high affinity for sigma receptors. In fact, extant data suggest that the ability of antipsychotic drugs to increase NT concentrations may reside in the ability to act at both D₂- and sigma-receptor sites. Selective sigma antagonists, such as BMY-14802, increase NT—NMN mRNA and NT concentrations in the nucleus accumbens and the caudate nucleus, but produce a decrease in frontal cortical NT concentrations. Previous studies have only observed a decrease in NT in the micropunch-dissected medial prefrontal cortex after chronic haloperidol and clozapine. Conceivably, sigma-receptor and D₂-redeptor blockade may regulate NT neurons together. However, the inability of (—)-butaclamol, a sigma, but not D₂ DA, receptor antagonist, to alter NT concentrations indicates that more than sigma-receptor blockade is required to induce NT production. Another sigma-receptor ligand, SR 31742A, has been reported to selectively increase NT in the nucleus accumbens, but not in the striatum, which is similar to the anatomic selectivity of clozapine for producing NT effects.

Currently, two schools of thought exist about the anatomic relationship of these DA and sigma receptors and the NT receptive neurons mediating the neuroleptic-like effects of NT. Destruction of DA neurons by 6-OHDA does not prevent haloperidol from increasing NT concentrations in the nucleus accumbens and caudate nucleus, indicating a postsynaptic location of the relevant D₂ DA or sigma receptor on, or immediately adjacent to, the resident NT neurons. However, Fuxe’s group argue for a presynaptic effect of NT in decreasing the affinity of D₂ DA receptors, which is postulated to result in decreased DA activity.

Electrophysiology

The electrophysiological effects of exogenously applied NT depends upon the regional cell population under study, whether brain slices or in vivo recording sites are used and the concentration of NT applied. Until recently, the lack of specific NT receptor antagonists precluded any definitive conclusion as to whether these effects of NT were due to the direct activation of NT receptors or due to indirect effects. Most of the brain regions where NT has been applied respond with excitation of the majority of neurons tested. Two exceptions to this general statement are the locus coeruleus and the nucleus accumbens. The application of NT or NMN to midbrain DA neurons produces an increase in firing rates and DA release at the terminal DA projections. DA-induced inhibition of the firing rate of DA neurons through autoreceptor activation is also attenuated by NT application and this effect has recently been shown to be due to opposite effects of NT and DA D2 receptors on the same potassium conductance (30). This effect on DA neuronal firing rateis not produced by cholecystokinin, which, like NT, is colocalized in a sizable proportion of midbrain DA neurons. The opposite effects of NT and cholecystokinin application to parabrachial neurons has been recently reported, with NT dose-dependently increasing and cholecystokinin decreasing parabrachial neuron firing rates (76).

Fuxe and coworkers have conducted a series of experiments which present convincing evidence that these effects of NT are mediated by NT-receptor-induced disinhibition of DA autoreceptor activation through a direct NT receptor-DA receptor interaction and such interactions are further supported by microdialysis evidence indicating NT receptor regulation of the D2 dopamine receptor -mediated release of GABA in the globus pallidus (33). Using the currently available NT receptor antagonists, the electrically induced release of DA seems to be mediated more by low affinity NT recptors whereas the K+ evoked release of DA is potently blocked by antagonists of the high affinity NT receptor (40). Thus, midbrain DA neurons respond to NT with effects that are similar to the effect of the DA-receptor antagonists such as haloperidol. This similarity does not extend to the frontal cortex, however, because NT effects following application to this region resemble those of DA. Interestingly, this region also responds to neuroleptic drug administration with decreases in NT concentrations rather than the increases seen in the striatum or nucleus accumbens. A recent report indicates that acute application of NT receptor antagonists into the frontal cortex or systemically injected dose-dependently increases the number of spontaneously active ventral tegmental area dopamine neurons with little effect on substantia nigra neurons while five weeks of antagonist delivered systemically decreased ventral tegmental area neuronal firing again without involving substantia nigra neurons (78).

Behavioral Effects

The behavioral effects of NT also depend upon the site of application within the CNS. Intraventricularly administered NT decreases spontaneous and psychostimulantinduced locomotor activity in rats and mice. This effect is likely mediated by actions of the peptide in the nucleus accumbens, because similar results are obtained when NT is microinjected directly into this brain region. However, NT administered into the VTA increases locomotor activity and rearing behavior in rats that can be blocked by pretreatment with antipsychotic drugs. In fact, the increase in locomotion induced by intra-VTA neurotensin can be blocked by intra-accumbens NT. Thus, NT can block the behavioral effects of DA at either end of this circuit. When DA agonists are used to produce penile erection or yawning behavior, NT administered intracerebroventricularly blocks these responses without altering the increase in sniffing stereotypies. Swerdlow and colleagues have also demonstrated that NT administration into the nucleus accumbens in low doses (0.25 or 1.0m g) blocks amphetamine induced decreases in pre-pulse startle inhibition (31) while a higher dose of NT (5m g) into the nucleus accumbens potentiates the disruption of pre-pulse startle inhibition in rats by the dopamine agonists amphetamine and apomorphine (32), indicating a neuroleptic-like effect at lower doses and a pro-dopamine effect of higher dose levels of NT in this paradigm modelling an aspect of schizophrenic symptoms.

The intra-VTA administration of NT exhibits rewarding properties as measured by place-preference and self-administration paradigms, but it decreases the response for food rewards in an operant paradigm. This latter effect may be due in part to a role in gut endocrinology, as NT is increased in hypothalamus by fatty meals and may be responding to the cascade of gut hormones that are thought to induce satiety. These responses are disregulated in cystic fibrosis with pancreatic insufficiency with the result that pre- and post-prandial NT plasma levels are increased, but are not correlated with the severity of the pancreatic insufficiency (43). One of the newest research fronts on the endocrine actions of NT indicate that hypothalmic NT gene expression is increased after three days of leptin administration that produces food intake and body weight (75) and leptin receptors have been mapped to NT-containing neurons of the arcuate nucleus of the hypothalamus (38). Thus, in addition to the well described role of peripheral NT in mediating intestinal motility (65), NT may regulate feeding through effects on CNS satiety circuits.

When rats are cannulated with stimulating electrodes in the posterior lateral hypothalamus or PAG matter, the self-stimulation rate of bar presses is decreased by NT injection into the VTA, suggesting that NT applied to this region is reinforcing. Cholecystokinin application into the VTA has the opposite effect. This effect of NT on VTA self-stimulation is apparently mediated through NTR1 high affinity receptors but self stimulation of electrodes planted into the lateral hypothalamus are not blocked by NTR1 antagonists (86). Both NT and NMN decrease self-stimulation when injected into the medial prefrontal cortex, indicating possible reinforcing properties of NT and NMN in this circuit.

NEUROCHEMICAL PATHOLOGY

Schizophrenia

The primary evidence for NT involvement in the pathogenesis of schizophrenia rests on several clinical studies. Our group has repeatedly observed decreased group mean NT concentrations in CSF in drug-free schizophrenic patients compared to normal, healthy, sex- and age-matched volunteers and we have recently confirmed the initial observations in a further experiment where 42 drug-free schizophrenic or schizoaffective patients submitted a pre-treatment CSF sample and 18 of these patients allowed collection of another CSF sample after 4 weeks of antipsychotic drug treatment. Psychopathology was highest in the patients with the lowest pre-treatment CSF concentrations of NT and degree of improvement in overall psychopathology and particularly negative symptoms correlated with increases in NT concentrations after treatment (84). These observations are consistent with the hypothesis of reduced NT synaptic availability contributing to psychotic symptom presence in schizophrenia. Another CSF study (17) demonstrated that several metabolites of NT may differ between normals, depressed patients and schizophrenics suggesting that peptidase activity may differ among these clinical populations. We have shown that NT concentrations increase in CSF after antipsychotic drug treatment, an effect similar to that observed in the brain tissue of similarly treated laboratory animals. Obviously, the presence of some period of antipsychotic drug pretreatment in almost all of the patients studied using post-mortem tissue thus far is a potential confound for assessing NT involvement in the primary pathology of schizophrenia. The four published studies in which NT concentrations have been measured in postmortem brain tissue from schizophrenic patients are encumbered by the same potential confound, although antipsychotics are usually withdrawn relatively early in the terminal phases of illness. Of the seven cortical regions examined among all of these studies, only Brodmann’s area 32 from the schizophrenic tissue exhibited any group mean change in NT concentrations, an increase relative to controls. None of the seven total subcortical regions examined thus far have shown any group mean differences in NT concentrations, including two studies using nucleus accumbens and one with caudate nucleus tissue. Thus, the schizophrenic postmortem tissue data do not confirm the expectations engendered by the CSF studies, where decreases in NT have been observed. A recent study of NT—NMN messenger RNA expression and genomic sequence in ventral midbrain neurons from schizophrenic patients compared to controls did not find any differences. Kleinman and colleagues (personal communicdtion), using autoradiography, have not found differences between schizophrenics and controls as regards NT-receptor density. However, Watson and co-workers have found evidence of polymorphisms in the NT-receptor mRNA sequences from human CNS tissue. Whether similar polymorphisms are present in schizophrenia and what the functional implications of such differences would represent await future studies.

PARKINSON’S DISEASE

In Parkinson’s disease, NT receptors residing on DA neurons in the substantia nigra are reduced in number as these neurons degenerate over the disease course. However, group mean NT concentrations is postmortem Parkinson’s disease brain regions are generally not found to be different from age- and sex-matched controls, including the caudate nucleus, nucleus accumbens, and VTA. A decrease in hippocampal NT concentrations in Parkinson’s disease patients relative to controls that was reported by our group has not been confirmed by others. This finding in Parkinson’s disease is similar to the lack of effect on NT concentrations in the rat brain in response to 6-OHDA-induced destruction of DA neurons—except in regions such as the medial prefrontal cortex, where NT and DA are extensively colocalized. Interestingly, in an animal model of Parkinson’s disease using 6-OHDA, Jolicoeur and colleagues found that intraventricular NT administration would reduce the tremor and rigidity of rats.

Alzheimer’s Disease

In Alzheimer’s disease, there have been four postmortem tissue reports where regional NT concentration changes were sought. Our group reported a decrease in amygdala NT concentrations in Alzheimer’s patients relative to controls, and this has been confirmed by Benzing and co-workers using immunohistochemical methods. No other brain regions have been reported to exhibit alterations in NT systems in Alzheimer’s disease.

Others

Patients with Huntington’s chorea have also been investigated for postmortem regional alterations in NT concentrations. Increases in NT concentrations in the caudate nucleus and in the globus pallidus have been reported.

Two groups have reported alterations in NT systems in the CNS of infants dying of sudden infant death syndrome (SIDS). Coquerel and colleagues described increased NT concentrations in CSF of SIDS patients relative to control children and adults and measured NT in brainstem regions of SIDS patients. Because no controls were used in the brainstem studies, the results are difficult to interpret, but the SIDS brainstem NT concentrations are much higher than previously reported adult concentrations. The second study measured high-affinity NT receptors in SIDS brain-stem sections and compared them to age-matched, non-SIDS controls. They observed increased densities of NT binding sites in the nucleus tractus solitarius, but no differences were found in other brainstem regions. Another study using SIDS brainstem tissue reported a transient increase in GTP-sensitive, high-affinity NT receptors over the first month of life, with levels decreasing to near adult values by 15 months of age. The transient increase of NT receptors at 1 month in SIDS tissue was interpreted as putative evidence for NT playing a role in brainstem development and points to possible disruption of this process in SIDS.

Rett’s syndrome is a progressive neurological disorder in female children characterized by decelerated brain and body growth, cortical atrophy, and severe mental deficiency. Stereotypical "hand washing" movements are often present, and this syndrome is now considered to be an X-linked dominant mutation that is lethal to males. An investigation into several neurochemical systems thought to be involved in the resulting neuropathology of Rett’ s syndrome found no changes in NT concentrations in CSF nor in NT-receptor numbers in postmortem brain tissue despite evidence that both dopaminergic and cholinergic systems are targets of Rett’s syndrome neuropathology.

DRUGS TARGETED TOWARD NT RECEPTORS

Antagonists

The goal of availability of agonists and antagonists at the NT—NMN receptor has begun to be realized. Recently described nonpeptide NT-receptor antagonists, such as the polycyclic compound SR-48692, promise to afford the long-awaited opportunity to block the actions of endogenous NT, hopefully providing information on the physiological role(s). This compound, in particular, promises to be extremely useful because it is lipophilic and penetrates into the CNS after peripheral administration (see 72 for review). It has greater affinity for the high affinity NT receptor than for the levocabastine-sensitive, low affinity NT receptor, although when expressed in Chinese hamster ovary cells (106) or Xenopus oocytes (15), SR-48692 acts as an agonist at the low affinity NT receptor (106). Peripherally, this antagonist blocks the production of nitric oxide and thus a NT receptor effect on nitric oxide synthase may be inferred (21). Another recently characterized non-peptide NT receptor antagonist, SR-142948A (37), blocks both the low and high-affinity forms of the NT receptor (11) and is more potent than SR-48692 in blocking the cardiovascular effects of NT (81).

A completely novel approach to receptor blockade has been launched with peptide nucleic acids which are able to hybridize to complementary DNA strands encoding the NT high affinity receptor. This approach was able to block the hypothermia and analgesia seen after NT administration into the PAG region (95).

Agonists

One of the obvious tests of the endogenous neuroleptic hypothesis of NT would be the administration of an NT-receptor agonist to symptomatic schizophrenic patients to determine if the NT-receptor agonist would possess antipsychotic properties. Previously, substituted peptide NT analogues required direct CNS injection to produce effects. However, this goal may soon be possible, because a C-terminal NT analogue linked to a small, CNS accessible molecule has recently been observed to potentiate pentobarbital sedation and produce hypothermia in a dose-dependent manner after intraperitoneal injection (Bissette and Richelson, personal communication). Should this compound prove nontoxic at useful doses, clinical trials will be subsequently initiated. In a related development, N-acylprolytyrosine analogs of NT have been shown to block apomorphine-induced climbing in mice and dopamine-induced exploratory behavior in rats and the low toxicity of these analogs encourages exploration of possible clinical utility (36). The Eisai hexapeptide NT agonist crosses the blood brain barrier and may also represent a NT receptor agonist that would allow direct investigation of antipsychotic ability without extrapyramidal liability in human schizophrenic patients (79).

DISCUSSION

Whether NT is truly an endogenous neuroleptic or not, the past 25 years of research have certainly shown it to be one of the most interesting peptides investigated to date. At present, NT is definitely implicated in the physiological regulation of luteinizing hormone and prolactin release and a variety of CNS dopaminergic systems. The involvement of NT systems in the clinical neuropathology of several neurologic and psychiatric diseases is adumbrated by the available data; in most cases, however, independent verification has been difficult, probably because of the relatively modest populations sampled thus far. The close relationship between NT and dopaminergic systems in the CNS, combined with the extra-opioid analgesia mediated by NT, would be enough to sustain interest in NT apart from the tantalizing possibility of NT systems mediating some of the psychopathology of schizophrenia. The future development of positron emission tomography ligands to directly image NT-receptor density in humans will be a major advance in the NT research field. Given the explosion of knowledge in the past two and a half decades, even the most grandiose visions of NT may be fully realized in the next several years.

ACKNOWLEDGMENTS

This work was supported by NIMH MH-39415.

published 2000