This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ben-Jonathan, N. Articles by Hnasko, R. Search for Related Content PubMed PubMed Citation Articles by Ben-Jonathan, N. Articles by Hnasko, R.

Dopamine as a Prolactin (PRL) Inhibitor

Department of Cell Biology, Neurobiology, and Anatomy, University of Cincinnati Medical Center, Cincinnati, Ohio 45267

Correspondence: Address all correspondence and requests for reprints to: Dr. Nira Ben-Jonathan, Department of Cell Biology, 3125 Eden Avenue, University of Cincinnati, Cincinnati, Ohio 45267-0521. E-mail: Nira.Ben-Jonathan{at}uc.edu

	Abstract

Top Abstract I. Introduction II. Characteristics of... III. The Hypothalamo-Pituitary... IV. Dopamine and the... V. Lessons Learned from... VI. Clinical Aspects VII. Summary and Perspectives References

 
Dopamine is a small and relatively simple molecule that fulfills diverse functions. Within the brain, it acts as a classical neurotransmitter whose attenuation or overactivity can result in disorders such as Parkinson’s disease and schizophrenia. Major advances in the cloning and characterization of biosynthetic enzymes, transporters, and receptors have increased our knowledge regarding the metabolism, release, reuptake, and mechanism of action of dopamine. Dopamine reaches the pituitary via hypophysial portal blood from several hypothalamic nerve tracts that are regulated by PRL itself, estrogens, and several neuropeptides and neurotransmitters. Dopamine binds to type-2 dopamine receptors that are functionally linked to membrane channels and G proteins and suppresses the high intrinsic secretory activity of the pituitary lactotrophs. In addition to inhibiting PRL release by controlling calcium fluxes, dopamine activates several interacting intracellular signaling pathways and suppresses PRL gene expression and lactotroph proliferation. Thus, PRL homeostasis should be viewed in the context of a fine balance between the action of dopamine as an inhibitor and the many hypothalamic, systemic, and local factors acting as stimulators, none of which has yet emerged as a primary PRL releasing factor. The generation of transgenic animals with overexpressed or mutated genes expanded our understanding of dopamine-PRL interactions and the physiological consequences of their perturbations. PRL release in humans, which differs in many respects from that in laboratory animals, is affected by several drugs used in clinical practice. Hyperprolactinemia is a major neuroendocrine-related cause of reproductive disturbances in both men and women. The treatment of hyperprolactinemia has greatly benefited from the generation of progressively more effective and selective dopaminergic drugs.

I. Introduction

II. Characteristics of Dopaminergic Neurons

A. Synthesis and metabolism

B. Storage and exocytosis

C. Transporters

D. Receptors

III. The Hypothalamo-Pituitary Dopaminergic Systems

A. Anatomy and ontogeny

B. Physiology and pharmacology

C. Regulation by PRL

D. Effects of ovarian steroids

E. Interactions with neuropeptides and neurotransmitters

IV. Dopamine and the Pituitary Lactotrophs

A. Dopamine and its receptor

B. Actions and signal transduction

C. Antiproliferative activity

V. Lessons Learned from Transgenic Mice

A. Dopamine D₂ receptor (D₂R) and transporter

B. PRL and PRL receptor (PRL-R)

C. ERs, galanin, and nerve growth factor (NGF)

VI. Clinical Aspects

A. PRL physiology in humans

B. Drug-induced PRL release

C. Hyperprolactinemia and pituitary prolactinomas

D. Treatment of hyperprolactinemia

VII. Summary and Perspectives

	I. Introduction

Top Abstract I. Introduction II. Characteristics of... III. The Hypothalamo-Pituitary... IV. Dopamine and the... V. Lessons Learned from... VI. Clinical Aspects VII. Summary and Perspectives References

 
DOPAMINE, NOREPINEPHRINE, AND epinephrine belong to a class of neurotransmitters known as catecholamines, which are structurally defined by a catechol ring and an amine side chain. Catecholamines and indolamines (i.e., serotonin) are referred to as monoamines. Monoamines are small, water-soluble molecules that are the decarboxylated derivatives of amino acids. Production from their respective amino acids is catalyzed by several enzymes that act in sequence, the first of which serves as the rate-limiting step. Monoamines are stored at high concentrations in secretory granules. These granules provide protection against degradation by metabolic enzymes and enable a regulated release via exocytosis. Like other neurotransmitters, monoamines act very rapidly and their action can be terminated by both metabolic conversion to inactive compounds as well as by reuptake into the producing cell.

Dopamine is synthesized primarily in the central nervous system (CNS), but limited production also occurs in the adrenal medulla. Dopamine is also detectable in a few non-neuronal tissues, e.g., the pancreas and the anterior pituitary. Dysfunction of dopaminergic systems is associated with a number of diseases. For example, deficiency of dopamine in midbrain nigrostriatal neurons has long been recognized in the pathogenesis of Parkinson’s disease, while overactivity of the limbic and cortical dopaminergic neurons has been implicated in schizophrenia and psychoses. These dopaminergic neurons are also affected by neurotoxins, psychostimulants, and drugs of abuse. In the neuroendocrine axis, dysfunction of hypothalamic dopamine or its pituitary receptors leads to hyperprolactinemia and reproductive disturbances. It is not surprising, therefore, that this relatively simple molecule has been at the center of interest of basic scientists and clinicians alike for many years.

Within the brain, catecholamines function as classical neurotransmitters, i.e., they communicate between neurons and act within the anatomically confined space of the synapse. However, by virtue of their presence in the circulation and action on distant target organs, catecholamines from the adrenal medulla were among the first compounds classified as hormones in the early 1900s. Not until the 1970s, however, did the role of dopamine as an inhibitor of the pituitary lactotrophs become recognized. Since then, dopamine has been clearly established as the primary regulator of PRL gene expression and release. On the other hand, among the many factors capable of stimulating PRL, none has emerged as a leading candidate for a PRL releasing factor (PRF). Therefore, PRL homeostasis should be viewed in the context of a fine balance between the action of dopamine as an inhibitor and the many hypothalamic, systemic, and local factors acting as stimulators.

In 1985, we published a review in this journal entitled "Dopamine: A Prolactin Inhibiting Hormone" (1). The present update covers pertinent information that has been gathered since the publication of this report. During the last 15 yr, this field has witnessed unparalleled progress, including the cloning of dopamine and PRL receptors (PRL-Rs), the characterization of the dopamine transporter, the recognition of the role that estrogen and its receptors play in PRL homeostasis, and the generation of transgenic animals deficient in all these genes. In terms of therapeutic applications, dopaminergic agonists have become the mainstay treatment for suppressing PRL in hyperprolactinemic patients and for shrinking prolactinomas.

The review is organized in five chapters. The first chapter covers advances in the understanding of dopamine synthesis, storage, release, reuptake, and receptor binding. The second and third chapters focus on the hypothalamo-pituitary dopaminergic systems, their regulation by various factors, and the mechanisms by which dopamine affects the lactotrophs. Lessons learned from transgenic animals with altered genes that are relevant to PRL regulation constitute the fourth chapter. Finally, the profile of PRL release in humans, clinical aspects of dopaminergic drugs, and the pathophysiology of hyperprolactinemia are presented in the fifth chapter. Since 1985, over 4000 articles have been published on various aspects of dopamine-PRL interactions. By necessity, this review is selective rather than inclusive. Consequently, the reader is referred whenever possible to other reviews for more in-depth coverage of the different topics.

	II. Characteristics of Dopaminergic Neurons

Top Abstract I. Introduction II. Characteristics of... III. The Hypothalamo-Pituitary... IV. Dopamine and the... V. Lessons Learned from... VI. Clinical Aspects VII. Summary and Perspectives References

 
A. Synthesis and metabolism
Dopamine biosynthesis begins with the amino acid tyrosine (see Fig. 1). The majority of circulating tyrosine originates from dietary sources, but small amounts are derived from hydroxylation of phenylalanine by the liver enzyme phenylalanine hydroxylase (for review, see Ref. 2). Tyrosine enters neurons by an energy-dependent uptake process and is converted to dopamine by two enzymes that act in sequence, tyrosine hydroxylase (TH) and L-aromatic amino acid decarboxylase, also called dihydroxyphenylalanine (DOPA) decarboxylase (DDC). Neurons that contain active dopamine ß-hydroxylase (DBH) convert dopamine to norepinephrine, and those that also contain phenylethanolamine N-methyl transferase convert norepinephrine to epinephrine. The latter are classified as noradrenergic and adrenergic neurons, respectively, and their distribution in the brain differs considerably from that of the dopaminergic neurons. Regardless of the catecholamine being produced, TH is the rate-limiting step in their biosynthetic pathway.

View larger version (22K): [in this window] [in a new window]   Figure 1. Diagram of dopamine biosynthesis, release, and metabolism. 1) Tyrosine is taken into the neuron by a sodium-dependent mechanism; 2) conversion of tyrosine to L-DOPA by TH is the rate-limiting step in the biosynthetic pathway; 3) L-DOPA is converted to dopamine by DDC; 4) dopamine is translocated into secretory vesicles for storage, protection, and secretion; 5) fusion of secretory vesicles with the plasma membrane results in dopamine release into the synaptic cleft or the extracellular space ((s is the case with the TIDA neurons); 6) dopamine binds to its membrane receptors and initiates multiple effects in target cells; 7) unbound dopamine is taken up by the DAT, located in the plasma membrane of the presynaptic neuron; 8) both newly synthesized dopamine and that taken up into the cell are translocated into secretory vesicles by the VMAT; 9) MAO, located in the outer mitochondrial membrane, converts dopamine to a deaminated metabolite; 10) COMT converts dopamine or its deaminated metabolite to biologically inactive products.

The TH gene is localized to chromosome 11p in humans and encodes a single form of TH that can be alternatively spliced (5). Targeted disruption of the TH gene results in perinatal lethality, which can be rescued by L-DOPA administration (6). The mature enzyme is a soluble cytosolic protein composed of four subunits of approximately 60 kDa each (reviewed in Ref. 7). Each monomer is comprised of an inhibitory regulatory domain at the N terminus and a catalytic domain at the C terminus. The regulatory domain contains four phosphorylation sites located within the first 40 amino acids: Ser⁸, Ser¹⁹, Ser³¹, and Ser⁴⁰. The catalytic domain contains the pterin binding region and a putative leucine zipper at the C terminus that participates in intersubunit binding.

TH activity is the most critical factor that controls dopamine synthesis, and considerable efforts have been devoted to understanding activation/inactivation of this enzyme. TH activity is regulated by two mechanisms: short-term activation and long-term induction (for review, see Ref. 8). Activation (seconds to minutes) occurs in response to increased nerve impulses or pharmacological agents and involves removal of feedback inhibition by dopamine, allosteric regulation by polyanions, and phosphorylation. All phosphorylation sites, but especially Ser⁴⁰, are important for TH activation (9). Phosphorylation at specific sites is accomplished by several Ser/Thr kinases, e.g., PKA, PKC, ERK1/2, and calcium calmodulin-dependent protein kinase II, and results in conformational changes that alter enzyme affinity either to the pterin cofactor or to dopamine acting as an inhibitor (4). Because TH phosphorylation can be reversed by phosphatases, the activated state of TH at any given time reflects a dynamic balance between these antagonizing forces. As discussed in Section II.B, TH appears to be constitutively activated within the hypothalamo-pituitary unit, in sharp contrast with its normally quiescent state in both the striatum and adrenal medulla.

Long-term induction of TH involves transcriptional regulation, alternative RNA splicing, RNA stabilization, and translational regulation (for review, see Ref. 8). Nucleotide sequences up to 9 kb upstream from the transcriptional start site are necessary for developmental and tissue-specific control of TH expression. The promoter region contains several positive and negative transcriptional elements that are not conserved across species and differ among tissues (10). Changes in TH mRNA levels occur in response to alterations in physiological conditions, e.g., cold exposure and chronic stress, and are often mediated by glucocorticoids (11). Three major second messenger systems, cAMP, diacylglycerol, and calcium, which use a variety of effector molecules, have been implicated in this response. A combination of enzyme activation and long-term induction maintains dopamine synthesis by diverse and seemingly redundant tissue-specific mechanisms. These overlapping actions guarantee uninterrupted supply of the neurotransmitter and permit rapid neuronal responsiveness to many physiological stimuli.

As shown schematically in Fig. 1, DDC is the second and terminal enzyme in dopamine biosynthesis (reviewed in Ref. 12). The enzyme uses pyridoxal phosphate as a cofactor and can convert both DOPA to dopamine and 5-hydroxytryptophan to serotonin [5-hydroxytryptamine (5-HT)]. Although a single gene codes for the enzyme, there are several isoforms that may be responsible for preferred decarboxylation of either dopamine or serotonin. The mature enzyme is a dimer made of 50-kDa subunits and is regulated by de novo synthesis rather than by changes in its activity. Under basal conditions, enzyme activity is so high that L-DOPA is virtually undetectable.

Unlike dopamine, DOPA can cross the blood brain barrier, and this property has been exploited in the treatment of Parkinson’s disease, especially during the early stages when a sufficient number of midbrain dopaminergic neurons are still functional. To prevent rapid decarboxylation, DOPA has to be administered together with peripheral DDC inhibitors such as carbidopa or benserazide (13). Other DDC inhibitors, e.g., NSD 1015, have been widely used in laboratory animals for measuring DOPA accumulation as an index of dopamine biosynthesis (14). This approach is based on the findings that DOPA levels without drug application are virtually undetectable. Measurement of DOPA accumulation is well suited for evaluating dopaminergic neuronal activity in the median eminence and posterior pituitary, which contain only minimal levels of norepinephrine.

Catabolism is one of the effective mechanisms for dopamine inactivation (reviewed in Ref. 15). This involves multiple pathways that include oxidative deamination by monoamine oxidase (MAO), O-methylation by catechol-O-methyl transferase (COMT), and conjugation by sulfotransferases or glucoronidases. The preferred metabolic pathway at a given site depends on the compartmentalization of the metabolic enzymes. For example, MAO is located in the external membrane of the mitochondria and acts intracellularly, whereas COMT is associated with the external cell membrane and acts only extracellularly.

MAO exists as two isoenzymes, A and B, with an apparent molecular mass of 60–63 kDa each. The two MAO genes, each comprised of 15 exons, are located on the X-chromosome and appear to have been derived from the same ancestral gene (reviewed in Ref. 16). They differ in substrate specificity as well as selectivity for inhibitors. MAO-A is more highly expressed in catecholaminergic neurons, whereas MAO-B is more abundant in serotonergic and histaminergic neurons and in glial cells (17). Enzyme inactivation in humans or its deletion in transgenic mice are compatible with life but result in neurochemical and behavioral abnormalities (16). Deamination of dopamine by MAO produces dihydroxyphenylacetic acid (DOPAC). Determination of the ratio of DOPAC/dopamine concentrations serves as a good method for estimating rapid changes in neuronal activity, with a major advantage being that it does not require drug pretreatment. O-Methylation by COMT is primarily responsible for inactivation of circulating catecholamines. Consecutive conversion of dopamine by MAO and COMT yields homovanillic acid.

B. Storage and exocytosis
Because most endocrinologists are more familiar with peptide/protein hormones than with neurotransmitters, it is appropriate to compare the different characteristics of storage and exocytosis of these two classes of compounds. Dopamine is stored in secretory vesicles at a 100- to 1000-fold higher concentration than neuropeptides. This is attributed to several distinct features of monoaminergic neurons. First, unlike neuropeptides whose synthesis occurs within the endoplasmic reticulum and Golgi apparatus, dopamine biosynthesis can take place within the terminals themselves. Second, synthesis that occurs in a close proximity to the site of release permits a much faster turnover rate than the slow axoplasmic transport that brings proteins from cell bodies to the nerve terminals. Third, a unique reuptake process replenishes most of the released dopamine back into the secretory vesicles and maintains high intragranular concentration, whereas a released neuropeptide cannot be restocked.

After synthesis, dopamine is stored in synaptic vesicles at extremely high concentrations, 0.5–0.6 M, which is near its limit of solubility. Dopamine is translocated from the cytoplasm into the vesicles by the vesicular monoamine transporter (VMAT), shown schematically in Fig. 1 and discussed in detail in Section III.C. The function of the vesicles is 4-fold: 1) to protect dopamine from enzymatic degradation by MAO, 2) to minimize constitutive secretion by diffusion from the cells, 3) to facilitate regulated release, and 4) to enable rapid replenishment of depleted stores. The life cycle of the vesicles includes: 1) targeting to the active zone of the presynaptic membrane, 2) docking, 3) fusion, 4) release of the vesicular content, 5) retrieval by endocytosis, and 6) refilling with the neurotransmitter. Selected aspects of these events are discussed below. For a comprehensive coverage, please refer to several outstanding reviews (18, 19, 20, 21).

Monoamines are stored primarily in small translucent ("clear") vesicles (50–100 nm in diameter) but are also present in large dense core vesicles (up to 500 nm in diameter), often cosequestered with neuropeptides (22). The relationship between large and small vesicles, their predominance in catecholaminergic vs. neuroendocrine cells, their membrane composition, and their precise role in quantal neurotransmitter release are not clear. Also, most information on synaptic vesicles has been obtained from chromaffin cells, which contain primarily norepinephrine and epinephrine and differ from dopaminergic neurons by the presence of intravesicular DBH, chromogranins, and other constituents (23). It remains to be determined whether the content of the vesicles and the process of exocytosis are identical in dopaminergic and noradrenergic neurons.

Storage vesicles are formed in the neuronal perikarya and are transported to the terminals by slow axoplasmic flow. Although early studies suggested formation of vesicles from the outer membrane of the terminals by pinocytosis, it was later realized that this represented retrieval by endocytosis of previously fused vesicles. In fact, to maintain adequate transmitter storage and permit a sustained response to stimuli, endocytosis must occur at a rate that parallels exocytosis. The synaptic vesicle is a highly specialized structure whose membrane is composed of a lipid bilayer with embedded integral proteins that participate in vesicular trafficking, docking, and fusion. The vesicle membrane also contains an H⁺- ATPase, which maintains the proton gradient that energizes VMAT and preserves an acidic intravesicular environment. Each vesicle is filled with several thousand molecules of dopamine as well as other soluble constituents (24).

Much information has been gathered in recent years on the docking mechanism (reviewed in Refs. 18 and 25, 26, 27). It involves a family of proteins termed soluble N-ethylmaleimide-sensitive factor (NSF) attachment proteins (SNAP) receptors (SNARE) complexes: v-SNAREs, designating vesicular-associated proteins, and t-SNAREs, designating target (plasma membrane) cognate complexes. At least seven to eight proteins are essential for docking: vesicular synaptobrevin and synaptotagmin; SNAP-25 and syntaxin, which are located in both the vesicles and plasma membrane; and two soluble proteins, NSF and SNAP, which catalyze the disassembly of the SNAP-25-syntaxin-synaptobrevin complex during docking and fusion (28). SNAP-25, in association with syntaxin, binds to and modulates voltage-gated calcium channels, thus bringing the vesicle into close proximity with a source of calcium. Both N-type and P/Q-type calcium channels have been implicated in neuronal exocytosis, with synaptotagmin I acting as a low-affinity calcium sensor (29).

The critical role of calcium in exocytosis, termed the "stimulus-secretion coupling" hypothesis, has been long recognized. Calcium is central to all aspects of exocytosis, including rapid fusion and unloading of the vesicles as well as recruitment and translocation of loaded vesicles. Resting levels of cytoplasmic calcium within the neuron are approximately 0.1 µM and can rise to 5–10 µM upon arrival of action potentials (19). Calcium influx occurs through voltage-gated calcium channels and leads to fusion of the synaptic vesicles with the plasma membrane and release of their content to the extracellular space. This is a much faster process than the relatively slow release of peptide or protein hormones from endocrine cells.

In most neurons, dopamine is released into the synaptic cleft and binds to postsynaptic receptors. In contrast, the dopaminergic neurons of the hypothalamo-pituitary unit (with the exception of the dopaminergic neurons innervating the intermediate lobe of the pituitary) lack true synaptic contacts and are classified as secretory neurons (30). In this case, dopamine diffuses away from the terminals through the perivascular space and is transported by portal blood to distal pituitary target cells. The rate of dopamine release from secretory neurons appears to be slower than that from classical neurons. It has been argued that the speed of neurotransmitter release is reciprocally related to the distance of its site of action, but the mechanism responsible for this feature is unclear (19).

Calcium influx in chromaffin cells induces an initial fast release, termed the "exocytotic burst," which occurs in milliseconds and is followed by a slower and sustained release phase that lasts several seconds (27). It is assumed that only a small fraction of docked vesicles can instantaneously release their cargo in response to calcium influx. These vesicles comprise the "fusion-ready" pool that undergoes a very rapid ATP-dependent fusion. The slower release phase is carried out by docked vesicles that exist in a different biochemical state and require priming to promote fusion. These vesicles constitute a precursor pool that replenishes the rapid release pool. Priming is ATP dependent, involves the SNARE proteins, and is associated with production of phosphoinositides and protein phosphorylation. An even slower pool is composed of vesicles that are anchored to the cytoskeleton via actin-binding synapsins but are not docked to the membrane (25). When synapsins become phosphorylated in response to an influx of calcium, the vesicles detach from the cytoskeletal elements and can translocate to the active zone of the presynaptic membrane. However, vesicular translocation is too slow to account for the immediate calcium-dependent exocytosis.

Norepinephrine is released from chromaffin granules together with ATP and chromogranins but it is unclear whether this also occurs in dopaminergic neurons (23). The dynamics of release has been studied by electrophysiological approaches capable of resolving single exocytotic events. Such techniques can detect small changes in membrane capacity, reflecting an increase in plasma membrane surface due to vesicular fusion, and can also measure the oxidation/reduction potential of minute amounts of the released transmitter (31). Two pathways have been proposed to explain formation of fusion pores that connect the vesicle lumen with the extracellular space. One is termed the "kiss-and-run pathway," in which a pore is formed to allow partial or full emptying of the vesicle content. The other is termed the "complete fusion pathway," in which the pore dilates and the vesicle membrane collapses into the plasma membrane (18, 32). At least two membrane proteins, synaptophysin and synaptoporin, have been associated with pore formation. Questions that remain to be resolved include the three- dimensional structure of the putative pores, the precise mechanism of vesicular retrieval, and the dynamic forces that drive intracellular trafficking of the internalized vesicles.

C. Transporters
Reuptake is the process by which the released transmitter is brought back into presynaptic nerve terminals or is internalized by surrounding glial cells. It is unique to monoamines and amino acid neurotransmitters and is the main mechanism by which the action of the released transmitter is rapidly terminated (see Fig. 1). As an added benefit for monoamines, reuptake permits recycling of the same molecules while saving in energy costs of their biosynthesis (33). In contrast, the action of the released neuropeptide is terminated either by diffusion or by proteolysis.

Reuptake of dopamine is mediated by two classes of transporters: dopamine transporter (DAT), which transports dopamine from the extracellular to the intracellular space, and VMAT, which reloads dopamine into the vesicles (reviewed in Ref. 34). The two transporters differ in structure, cellular localization, substrate specificity, antagonist selectivity, and energy requirements (for a schematic presentation of the two transporters, see Fig. 2). Because dopamine in the hypothalamo-pituitary axis is not released directly into synapses, it has been argued that reuptake is not physiologically important in these neurons (35). As discussed in more detail in Section III.B, compelling evidence now indicates that a dopamine reuptake mechanism, though not robust, is an essential component in the overall regulation of PRL homeostasis.

View larger version (41K): [in this window] [in a new window]   Figure 2. Comparison of the structures of dopamine receptors (upper panel) and transporters (lower panel). Receptors: Both D1Rs and D2Rs belong to the superfamily of 7 TMDs, G protein-coupled receptors. D2Rs have a larger third intracellular loop, whereas D1Rs have a longer cytoplasmic carboxyl terminus. The locations of N-glycosylation (open circles), putative phosphorylation (circles with embedded P) sites, the s-s bridge in the extracellular face, and site of palmitoylation in the cytoplasmic face are also shown. Transporters: Both DAT and VMAT have 12 TMDs but otherwise have little primary sequence homology. Unlike the receptors, both the N- and C-termini of the transporters are located at the cytoplasmic side of the membrane. DAT has a large, N-glycosylated extracellular loop between TMDs 3 and 4, whereas the largest intraluminal loop of VMAT is located between TMDs 1 and 2. Open circles designate glycosylation sites.

The -aminobutyric acid (GABA) transporter (GAT) was the first neuronal transporter isolated by classical protein purification methods, followed by cloning of the norepinephrine transporter, NET (reviewed in Refs. 36 , 37 , and 40). After screening cDNA libraries from the rat midbrain with oligonucleotide probes complementary to conserved regions of these transporters, DAT was cloned by three groups in 1991 (41, 42, 43). Many neuronal transporters have since been cloned and are now grouped into a large family characterized by multiple transmembrane domains (TMDs) and sodium dependence. DAT belongs to a subfamily that includes transporters for GABA, norepinephrine, serotonin, glycine, and proline and is distinguished by 12 TMDs and dependence on both Na⁺ and Cl^-. Another subfamily is chloride independent, has 6–9 TMDs, and includes transporters for the excitatory amino acids glutamate and aspartate (40).

The DAT gene in humans is mapped to chromosome 5p (44). It spans 64 kb and is made of 15 exons, with the coding region beginning in exon 2 and extending partially into exon 15. There is a close correspondence between the exons and the putative TMDs, with no evidence for multiple start or polyadenylation sites or alternative splicing. The proximal 5'-flanking sequences lack canonical TATA or CAAT boxes and contain only a few known response elements for transcription factors (45). More distal sequences have multiple binding sites for nurr1, an orphan nuclear receptor transcription factor that is critical for the development of midbrain dopaminergic neurons (46), although it is not expressed in the hypothalamic dopaminergic neurons (47). A combination of positive and silencing elements within the promoter region accounts for the selective cellular localization of DAT within the brain. Although a number of potential transcription factor response elements (Erg-1, E-box, AP-2) have been identified in the proximal DAT promoter, their precise role in the regulation of DAT gene expression has not been well defined (45).

DAT encodes a 69-kDa protein of 620 residues with both the N and C termini located intracellularly (reviewed in Ref. 48 ; see Fig. 2). The protein lacks a consensus signal sequence and has 3–4 potential N-linked glycosylation sites in the second large extracellular loop. The nature and extent of glycosylation are tissue specific and may be involved in transporter targeting, stability, or ligand binding. Residues within TMD 1–3 influence binding affinity for dopamine and cocaine, whereas those in TMD 11–12 affect the affinity for the 1-methyl-4-phenylpyridinium ion (MPP⁺) neurotoxin (49, 50). As is typical for all neuronal transporters, DAT has lower affinity and reduced ligand specificity than the dopamine receptor. Several potential sites can be phosphorylated by PKC and may determine the rate of uptake or serve as a signal for transporter internalization (51).

As revealed by combined in situ hybridization and immunocytochemistry, DAT has restricted localization within the brain and is not expressed outside the CNS (34). The transporter colocalizes with TH, and because it is limited to dopaminergic neurons, it serves as a unique marker for these neurons. The highest expression of DAT is in the substantia nigra, followed by the ventral tegmental area. DAT in these neurons is detected in perikarya, dendrites, and axonal processes. A significant presence of DAT is also seen in mesolimbic and mesocortical dopaminergic pathways (52), whereas the hypothalamic dopaminergic neurons exhibit moderate and restricted expression of DAT. Unexpectedly, electron microscopy reveals that the DAT protein is found primarily in the extrasynaptic area rather than in the active zone of the synapse (53). This suggests that DAT may play a role in limiting diffusion of dopamine after being released. An unresolved issue is the mechanism by which dopamine ((nd other monoamines) is taken up by glia, because DAT is undetectable in these cells.

DAT is targeted by psychostimulants such as cocaine and amphetamine. By binding to the transporter and preventing dopamine reuptake, these drugs cause a prolonged increase in extracellular dopamine, resulting in augmentation of its effects. Because DAT is an excellent marker for functional extrahypothalamic dopaminergic neurons, in vivo imaging of cocaine analogs is used to evaluate the state of dopaminergic neurons in patients with Parkinson’s disease and other neurological disorders (54). The generation of transgenic mice with DAT inactivation added significant information on the physiological role of this transporter. These mice are hyperactive, do not respond to cocaine or amphetamine, are resistant to the neurotoxic effects of MPP⁺, and their dopamine receptor expression is down-regulated (55). The state of PRL and other hormones of the hypothalamo-pituitary axis in such animals is covered in detail in Section V.A.

Two VMAT isoforms, VMAT1 and VMAT2, were identified by expression cloning. They arise from distinct but related genes that encode proteins of 520 residues (reviewed in Refs. 33 and 56). Although they also have 12 putative TMDs, there is no sequence homology with the membrane transporters. The vesicular transporters are characterized by a large hydrophilic N-glycosylated intraluminal loop between TMDs 1 and 2, with both C and N termini located on the cytoplasmic side of the vesicular membrane (see Fig. 2). VMAT1 is present in developing neurons, peripheral tissues, and in some endocrine cells. VMAT2 is expressed in all major monoaminergic neurons throughout the brain, with a broad specificity for monoamine transport rated as serotonin> dopamine>norepinephrine>epinephrine>histamine (34). The weak substrate specificity indicates that uptake and storage of a secreted transmitter by the appropriate neuron is determined by the cell-selective plasma membrane transporter rather than by the vesicular transporter. Unlike DAT-deficient mice, which survive into adulthood, deletion of VMAT2 results in early postnatal mortality, underscoring the obligatory role of vesicular storage of monoamines for survival (57, 58, 59).

VMAT recharges the vesicles with the neurotransmitter by using an electrochemical gradient generated by vacuolar ATP-dependent H⁺ pump (V-ATPase). This pump maintains an intravesicular acidic environment (pH of 5.5 in chromaffin granules), which is necessary for uptake of the transmitter against its concentration gradient (60). Hence, VMAT differs from DAT and other membrane transporters that are driven by a Na⁺ gradient. Uptake into the vesicle involves extrusion of two protons for each transmitter molecule that is taken in. The low pH also facilitates packaging/storage of the vesicular content, but it is unclear whether vesicular alkalinization plays a role in exocytosis. In contrast to cocaine and amphetamine, which inhibit dopamine uptake by binding to DAT, the antihypertensive drugs reserpine and tetrabenazine inhibit uptake by interacting with VMAT (61).

D. Receptors
Studies in the late 1970s revealed binding of labeled dopamine to two receptors that were distinguished by pharmacological, physiological, and biochemical criteria and became known as D₁ and D₂ (62). It was then recognized that the D₁ receptor (D₁R) was coupled to G_s proteins and increased intracellular cAMP levels, whereas the D₂ receptor (D₂R) interacted with G_i proteins and inhibited cAMP accumulation. The D₂R was cloned 10 yr later by adopting a cloning strategy based on sequence homology to known G protein-coupled receptors (63). Cloning of the D₁R by several groups followed (reviewed in Ref. 64). Expression of the D₁R and D₂R in host cells confirmed their specificity for the various pharmacological agents and contrasting effects on adenylyl cyclase activity. Since then, three additional dopamine receptors were cloned and characterized. At present, there are 5 distinct receptors that are grouped into two subfamilies: the D1-like family, which includes D₁ and D₅, and the D2-like family, which includes D₂, D₃, and D₄. Because the regulation of PRL by dopamine is mediated by the D₂R, its properties will be emphasized (for structural comparison of D₁R and D₂R, see Fig. 2).

The dopamine receptors are members of the superfamily of G protein-coupled receptors. They are made of single polypeptide chains that range in size from 387 to 475 residues. The receptors have 7 transmembrane (TM)-spanning helices that form a ring-like hydrophobic pocket surrounded by 3 intracellular and 3 extracellular loops. The extracellular amino terminus in all dopamine receptors contains a similar number of residues but has a variable number of N-glycosylation sites (reviewed in Refs. 64 and 65). The cytoplasmic carboxyl terminus is much longer in D₁-like receptors and is the site of receptor anchorage to the plasma membrane via palmitoylation of a conserved cysteine residue (66). The third intracellular loop (between TMs 5 and 6) is significantly larger in the D₂ subfamily, as is the case for most receptors that interact with G_i proteins (Fig. 2). Several phosphorylation sites on the third intracellular loop and the cytoplasmic tail participate in receptor desensitization, although the physiological importance of desensitization is best established for adrenergic receptors (67). Two cysteine residues on the second and third extracellular loops form a disulfide bond that stabilizes receptor conformation. A combination of site-directed mutagenesis and protein modeling suggests that conserved amino acids in TM 2 ((spartate), TM 3 ((spartate), TM 5 (two serines), and TM 6 (phenylalanine) define a narrow pocket for agonist binding (64, 68).

The five dopamine receptors have different chromosomal localization, i.e., 5q, 11q, 3q, 11p, and 4p for D₁, D₂, D₃, D₄, and D₅, respectively (69). It has been proposed that most of the genes encoding G protein-coupled receptors originated from a primordial gene, probably one of the opsin genes (70). The genomic organization of the dopamine receptors is consistent with their divergence from two gene subfamilies that differ in the presence or absence of introns. Like most G protein-coupled receptors, the D₁-like receptor genes have no introns, whereas the D₂, D₃, and D₄ receptors have 6, 5, and 3 introns, respectively (71). The presence of introns permits generation of receptor variants by alternative splicing. D₂R has two functional variants, a short isoform, and a long isoform having a 29-residue insertion in the third cytoplasmic loop (72). Expression, regulation, and signaling of these variants within the anterior pituitary are covered in Section IV.A. Splice variants of other dopamine receptors may generate nonfunctional proteins.

In general, the D₁ and D₂ receptors are expressed at higher levels and have more selective agonists and antagonists than the D₃, D₄, and D₅ receptors (reviewed in Ref. 64). The five receptors have distinct, though often overlapping, localization within the brain and are expressed in a tissue-specific manner in the periphery. D₂R mRNA is highly expressed in the substantia nigra, ventral tegmental area, and hippocampus, whereas the amygdala contains primarily D₁ with only little D₂ mRNA. Both receptors are expressed at high levels in the caudate putamen, nucleus accumbens, and olfactory tubercle (73). The hypothalamus has moderate levels of both D₁ and D₂ mRNAs and low levels of D₄ and D₅. D₂R mRNA is expressed at high levels in both the anterior and intermediate lobes of the pituitary and at lower levels in the adrenal and retina. The D₃ receptor mRNA is not detected in peripheral tissues, whereas a low expression of D₁ and D₄ in the kidney and D₅ in the heart has been reported (74).

The association of dopamine receptors with many neurological disorders has led to the development of many agonists and antagonists (reviewed in Ref. 75). While the availability of stereoselective drugs was instrumental in the initial characterization of the receptors, their cloning subsequently helped in the discovery of more selective drugs. Dopamine receptor antagonists are known as neuroleptics and are widely prescribed for the treatment of schizophrenia and other psychoses (76). Because many neuroleptics elicit Parkinsonian side effects, i.e., rigidity and akinesia, it led to the development of newer drugs, known as atypical neuroleptics, with little or no adverse effects on motor functions (77). The most common antagonists of D₂R are (+)butaclamol, chloropromazine, haloperidol, spiperone, sulpiride, and raclopride whereas apomorphine, bromocriptine, pergolide, and cabergolide are potent D₂R agonists (reviewed in Ref. 64). As discussed in Section VI.B, the latter compounds are very effective in the treatment of prolactinomas. Neither of the above-mentioned drugs is absolutely specific for any dopamine receptor subtype, and their selectivity is based on differences in binding affinity and dissociation constants to the various receptors. Thus far, there are no drugs that discriminate between the two D₂R variants.

Signal transduction by the dopamine receptors is an active area of research (reviewed in Refs. 78 and 79). As mentioned before, early studies using brain and pituitary tissues established that activation of D₁-type receptors increased adenylyl cyclase activity, whereas activation of the D₂-type receptors resulted in its inhibition. Because of receptor heterogeneity in brain neurons and lack of truly selective agonists and antagonists, transfection of non-neuronal cells with the various receptors has provided the bulk of information on receptor signaling. The major caveat is that many transfected host cells do not express the G proteins or downstream effectors that are physiologically relevant to receptors on neurons, often leading to conflicting results.

The ability of dopamine receptors to couple to appropriate G proteins is at the heart of their action. It is now recognized that a given receptor can be associated with more than one G protein, thus increasing its diversity of action. The D₁-like and D₂-like receptors are primarily associated with the G_s and G_i subunits, respectively. However, the G₀ and G_q proteins, which are associated with ion channels and phosphoinositide metabolism, are also involved (80). Many of the actions of dopamine receptors are extremely fast, involving rapid changes in ion fluxes across the cell membrane. In some target cells, including the pituitary lactotrophs, receptor activation also leads to changes in gene expression and hormone secretion as well as alterations in cell growth and differentiation. The signal transduction pathways of the pituitary D₂R and their multiple effects on the lactotrophs are covered in Sections IV.A and IV.B.

In contrast to the aforementioned receptors that are localized post synaptically, dopamine autoreceptors are found on most parts of the neuron, i.e., soma, dendrites, and terminals (81, 82). Autoreceptors are divided into three subcategories that are classified by their ability to modulate dopamine synthesis, release, or neuronal firing rates. After being released from the neuron, dopamine can interact with autoreceptors and inhibits further release of the neurotransmitter. The elevated intracellular dopamine also suppresses TH activity by binding to its pterin cofactor and decreasing the rate of synthesis. Some prefrontal neurons, as well as the hypothalamic dopaminergic neurons, lack synthesis-modulating autoreceptors (83). Although both D₂ and D₃ receptors have been proposed to function as autoreceptors, the issue remains controversial. It is also unclear whether distinct receptor proteins modulate each of these functions or the same receptor protein is coupled to each function through distinct transduction mechanisms.

	III. The Hypothalamo-Pituitary Dopaminergic Systems

Top Abstract I. Introduction II. Characteristics of... III. The Hypothalamo-Pituitary... IV. Dopamine and the... V. Lessons Learned from... VI. Clinical Aspects VII. Summary and Perspectives References

 
A. Anatomy and ontogeny
The brain contains several well-defined dopaminergic systems. The most extensively studied are the nigrostriatal, mesolimbic, and mesocortical neuronal systems that originate in the midbrain and project to the striatum, limbic system, and cortex, respectively. These are involved in the control of locomotion, emotion, and cognition and have no direct role in the regulation of pituitary function. Within the hypothalamus, dopamine perikarya are located in several sites that are classified by the alphanumeric system of Dahlstrom and Fuxe (84). These include the posterior hypothalamus ((11), arcuate nucleus ((12), zona incerta ((13), periventricular nucleus ((14), and lateral and ventral hypothalamus ((15). Neurons arising from the A11 locus send axons to the spinal cord and constitute a diencephalospinal system whose function is not well understood. The incertohypothalamic neurons from the A13 group project diffusely to different areas of the hypothalamus and participate in the control of GnRH release. They also send some projections to more remote areas such as the amygdala. Although not well characterized, the neurons originating from A11 and A13 regions constitute the majority of the hypothalamic dopaminergic neurons (for review, see Ref. 14). Only neurons originating from the A12 and A14 groups are relevant to our discussion because they are directly involved in the control of PRL secretion.

Based on anatomical and functional studies, two dopaminergic systems that regulate PRL were initially identified: the tuberoinfundibular dopaminergic (TIDA) and tuberohypophysial dopaminergic (THDA). Further refinements in neuronal tracing techniques (85) revealed that most of the THDA neurons projecting to the neurointermediate lobe actually originate from the A14 cells in the periventricular nucleus and therefore were termed the periventricular- hypophysial dopaminergic (PHDA) system (86). Because of their lower abundance and heterogeneous distribution, the hypothalamic dopaminergic neurons are not as well characterized as their nigrostriatal counterparts.

The TIDA neurons provide the major dopaminergic input to the anterior pituitary (see Fig. 3). Most of their perikarya are located in the dorsomedial part of the arcuate nucleus, with a smaller population arising from the periventricular nucleus. Their relatively short axons terminate in the external zone of the median eminence near the primary capillary loops of the hypophysial portal vessels (reviewed in Ref. 14). The median eminence contains a negligible number of cell bodies and a dearth of classical synapses. Instead, the TIDA system represents neurosecretory neurons whose product is released into perivascular spaces surrounding the capillary loops and is carried by the portal blood to the anterior pituitary. Notably, some TH-positive perikarya in the ventrolateral portion of the arcuate nucleus that also project to the median eminence do not express DDC (87) or DAT (88) and are believed to release DOPA rather than dopamine. Because DOPA has not been detected in hypophysial portal blood, the DOPA that may be released from these neurons must be decarboxylated before reaching the pituitary. Whereas the arcuate nucleus receives multiple afferent connections from VIPergic, opioidergic, serotonergic, and NPYergic neurons among others, their precise connections to TH-positive neurons needs further clarification.

View larger version (28K): [in this window] [in a new window]   Figure 3. Diagram of the hypothalamic dopaminergic systems that regulate PRL. The TIDA neurons originate in the A12 region of the dorsomedial arcuate nucleus. They have short axons that terminate in the external zone of the median eminence near the primary capillary plexus of the portal vessels. The released dopamine is carried out by the long portal vessels to the anterior lobe ((L) of the pituitary. The THDA/PHDA neurons have perikarya primarily in the A14 region of the periventricular nucleus and send projections to the neural (NL) and intermediate (IL) lobes of the pituitary. The short portal vessels connect between the NL and AL but bypass the IL which is avascular. [Derived from M. E. Freeman et al.: Physiol Rev 80:1523:1631, 2000 (105 )].

The anatomy of dopaminergic neurons in the human hypothalamus has not been well characterized (reviewed in Ref. 14). Undoubtedly, dopamine is important for the regulation of PRL release in humans because drugs that interfere with its release or action affect circulating PRL levels to the same extent as they do in experimental animals. However, there is no clear demarcation of the A11–A14 dopaminergic groups in humans, and most TH-positive, DBH-negative neurons (designating dopaminergic rather than noradrenergic neurons) reside in the magnocellular neurons (91). However, it is clear that the hypothalamic dopaminergic neurons in humans have distinct properties from those in the nigrostriatal region, and their normal function is preserved in Parkinson’s disease. Indeed, Parkinson’s patients do not have elevated plasma PRL levels, and their basal PRL release is suppressed by dopaminergic agonists (92). Another point of departure from rodents is the absence of a distinct intermediate lobe in adult humans ((s opposed to its discrete presence in fetuses) and no comparable PHDA system. Although melanotrophs are sparsely distributed throughout the human pituitary, they are neither innervated by dopaminergic neurons nor respond to dopamine agonists and antagonists (93).

Embryonic development of the hypothalamic dopaminergic neurons in the rat progresses along four chronological stages: 1) generation from neuroepithelial precursors, 2) expression of biosynthetic enzymes, production of dopamine, and establishment of mechanisms for its reuptake and release, 3) development of efferent connections, and 4) formation of afferent innervation and synaptogenesis (for review, see Ref. 94). Stage 1 neurons are first detected in the zona incerta and periventricular zone on embryonic days (ED) 12–13, followed by those in the arcuate nucleus on ED 14–15. On ED 17–18, TH-immunoreactive neurons (95) and TH mRNA (96) are detectable throughout the hypothalamus, with TH-positive axons appearing in the median eminence 2 d later. Expression of TH, however, does not imply terminal neuronal maturity. In fact, the high levels of DOPA in the fetal hypothalamus (97) and in conditioned media from cultured fetal hypothalamic neurons (98) suggest low activity of DDC and diminished ability for dopamine biosynthesis.

Stage 2 neurons expressing both TH and DDC and capable of release and reuptake of dopamine first become apparent on ED 15–16 (99). Sexual dimorphism of arcuate neurons is already evident at this time, with number of TH-positive neurons higher in males but their size and content higher in females (99). The gender difference in dopamine content is maintained during culture of fetal diencephalic neurons, which respond to chronic exposure to gonadal steroids by decreased rate of DOPA synthesis in cultures from both sexes (100).

The development of the THDA/PHDA system is delayed. TH-positive fibers are first seen in the neural lobe on ED 20 and in the intermediate lobe 3–4 d after birth (95, 101). Intermediate lobe-derived factors such as brain-derived neurotropic factor and neurotropin-3 may be involved in directing outgrowth of the dopaminergic neurons from the hypothalamus toward the pituitary (102, 103). The density of TH immunoreactivity in the intermediate lobe increases during the 1st wk of life, with dopamine reaching peak concentrations by the end of the 2nd wk and decreasing thereafter (104). The increased density of dopaminergic innervation in the intermediate lobe during early postnatal life correlates with the ontogeny of dopamine binding sites and coincides with a marked reduction in the number of melanotrophs (104), reflecting the inhibitory action of dopamine on melanotroph proliferation.

B. Physiology and pharmacology
Of the three dopaminergic systems, the TIDA neurons play the predominant role in the control of PRL release. Substantial evidence, based on the measurement of dopamine in portal blood and determination of its concentration and turnover rates in both the arcuate nucleus and median eminence, has established that the activity of the TIDA neurons is altered under many physiological conditions known to affect PRL release (reviewed in Refs. 1 and 105). However, the early notion that the TIDA neuronal activity must be negatively correlated with PRL release is too simplistic because the expected reciprocal relationship between dopamine and PRL is often masked by the action of other factors, both positive and negative, that control PRL release.

The lactotroph is unique among endocrine cells in having a high basal secretory activity. Tonic inhibition by dopamine, which maintains low circulating PRL levels, requires a continuous high input of dopamine. The high output, in turn, depends upon a sustainable high rate of synthesis. To enable rapid PRL surges, the dopaminergic input to the lactotrophs must be concomitantly decreased. This process is accomplished by a unique mechanism governing the regulation of hypothalamic TH activity. TH in most tissues exists in a quiescent, nonphosphorylated state. In response to stimuli, the enzyme is rapidly phosphorylated, resulting in increased hydroxylation of tyrosine to DOPA and its instant conversion to dopamine that is immediately available for release (9). One notable exception is hypothalamic TH. In keeping with the constant demand for high dopamine output, hypothalamic TH is constitutively active, as judged by its lower Michaelis-Menten constant (K_m) for the pterin cofactor than striatal TH (106). In response to estrogen, hypothalamic TH is transiently inactivated, presumably by dephosphorylation because this can be reversed by inhibitors of protein phosphatases such as okadaic acid (107). This is supported by the report on rapid decline in TH activity in hypothalamic slices within 1 h of E2 treatment (108). The absence of dopamine autoreceptors on the TIDA neurons may assist in maintaining high dopamine output by reducing, or eliminating, the negative feedback by dopamine on TH activity.

As discussed before, reuptake plays a fundamental role in neuronal function by conserving the released neurotransmitter and terminating its synaptic action. Because the hypothalamic neurons release dopamine into portal blood rather than into synapses, early reports suggested that that the hypothalamic dopaminergic neurons lack a functional reuptake mechanism (35, 109). This was refuted by a later study demonstrating dopamine reuptake by incubated stalk median eminence and posterior pituitary, with reuptake inhibitors such as nomifensin and diclofensin increasing media dopamine levels after neuronal depolarization (110). The molecular basis for this reuptake process was later confirmed by the detection of DAT mRNA in the dorsomedial arcuate nucleus (34) and the demonstration of immunoreactive DAT in the median eminence, pituitary stalk, and intermediate and neuronal lobes (111).

The physiological relevance of this reuptake was supported by the acute suppression of serum PRL levels in ovariectomized (OVEX) rats treated with competitive DAT inhibitors such as cocaine or mazindol (111). Moreover, DAT-knockout mice, presumably because of increased dopamine outflow, have a marked reduction in pituitary PRL content and do not lactate (112). Since the ratio of DOPAC/dopamine in the median eminence is less than half that in the striatum (113), most of the released dopamine must be carried away by the portal blood and lesser amounts are taken up by the terminals and converted to DOPAC.

Although males and females have the same density of TIDA nerve terminals, there are marked sexual differences in their activity and responsiveness to physiological and pharmacological stimuli. Basal TIDA neuronal activity is higher in females, is decreased by ovariectomy, and is restored by estrogen (for review, see Refs. 105 and 113). An opposite trend is seen in males, whereby TIDA activity increases by orchidectomy and decreases by T. The lower basal activity of these neurons in males may be due to tonic inhibition by endogenous opioids (114). The TIDA neurons in females are more sensitive to stress and to feedback stimulation by PRL but less sensitive to bombesin and -opioid antagonists (115, 116).

The TIDA neurons also exhibit an endogenous daily rhythm of activity (for review, see Ref. 105). This is controlled by the suprachiasmatic nucleus (SCN), which coordinates photoperiodicity in the neuroendocrine axis. A proestrous-like mid-afternoon PRL surge can be induced daily in OVEX rats by estrogen, implicating some form of coupling to an intrinsic diurnal rhythm. TIDA neuronal activity in the estrogen-treated OVEX rats is high in the morning, decreases before the PRL surge, but remains suppressed throughout late afternoon (117, 118). Although these changes correlate well with the initiation of the surge, they do not correspond to its termination, suggesting involvement of factors other than dopamine. Progesterone also participates in the control of the TIDA rhythm by advancing the afternoon decline in TIDA activity (119). Such rhythmic activity does not occur in males and is abolished in females by lesions of the SCN (113). These data suggest that activation/inactivation of the TIDA neurons is driven by an endogenous rhythm that is independent of the reproductive state, but its amplitude and timing are modulated by ovarian steroids. Several factors, including opioid peptides, bombesin, and acetylcholine (113), whose actions may be mediated by nitric oxide (120), have been implicated in the control of the endogenous dopaminergic rhythm. The relative importance of all these factors as well as the hierarchy of their action remain to be defined.

The role of the THDA/PHDA neurons in PRL regulation has been controversial. Early studies failed to show a good correlation between THDA neuronal activity and PRL release, suggesting that they do not contribute to the control of PRL secretion (116). Other investigators, however, found that the concentrations of dopamine and DOPAC in the intermediate lobe exhibit a daily rhythm with a significant decline that coincides with the initiation of the proestrous PRL surge (121). They also found alterations in the DOPAC/dopamine ratio in both the intermediate and neural lobes in response to ovarian steroids (122) and PRL (123), indicating that all three populations of the dopaminergic neurons are involved with PRL homeostasis.

Unlike the long portal vessels that connect the median eminence to the anterior pituitary and can be cannulated, the short portal vessels linking the neural and anterior lobes are inaccessible for sampling. Therefore, the relative contribution of the THDA/PHDA neurons to the total dopamine reaching the anterior pituitary cannot be effectively determined and requires indirect approaches. Surgical removal of the posterior pituitary (posterior pituitary lobectomy or LOBEX) provided the first evidence that dopamine from the THDA/PHDA neurons suppresses PRL release. LOBEX in either male or female rats caused an increase in serum PRL levels, but not LH or GH, which was reversed by intracarotid injections of dopamine (124, 125). Another indirect method is compression of the pituitary stalk, which disrupts the neural input to the neurointermediate lobe but does not impede the blood supply to the pituitary (126). Within 1 wk of denervation, circulating PRL and -MSH levels increased 3- to 4-fold, whereas those of LH remained unchanged. The mode of transport of dopamine from the avascular intermediate lobe to the anterior lobe remains enigmatic. One possibility is that it diffuses into the vicinity of lactotrophs that line the pituitary cleft. Another possibility is that intermediate lobe dopamine indirectly regulates PRL by suppressing the release of a local PRL releasing/regulating factor (reviewed in Ref. 127).

The pharmacology of the hypothalamic dopaminergic neurons has been the subject of several studies. These revealed that the responsiveness of the TIDA neurons to dopaminergic agents differs in several respects from their striatal counterparts. Because the TIDA neurons lack autoreceptors, they are unresponsive to acute administration of nonselective dopamine agonists, such as apomorphine, which do not discriminate between D₁-like and D₂-like receptors (128). Such drugs, therefore, act indirectly by altering the secretion of PRL, which in turns affects the TIDA neurons via a short loop feedback mechanism.

Classical antipsychotic drugs with D₂R antagonistic properties, such as haloperidol, have no direct effect on the TIDA neurons but induce their activation within several hours, secondary to the rise in circulating PRL levels (129). On the other hand, atypical neuroleptics such as clozapine acutely increase TIDA neuronal activity, possibly by activating D₁Rs. Acute administration of D₁ agonists (e.g., SKF 38393) inhibits, whereas D₂ agonists (e.g., quinpirole) stimulate, the TIDA neurons. The opposing effects of stimulatory D₂Rs and inhibitory D₁Rs likely account for the lack of net effect of mixed D₁R/D₂R agonists on the TIDA neurons (128). In males, some of the D₂R-mediated activation of the TIDA neurons occurs via disinhibition of afferent dynorphinergic neurons that provide tonic inhibition over the TIDA neurons (130).

C. Regulation by PRL
In the absence of target gland hormones to provide feedback control over the lactotrophs, PRL regulates its own release by acting on the hypothalamic dopaminergic systems. This type of interaction, termed "short loop feedback," is mostly responsible for the maintenance of PRL homeostasis. Many studies have established that an increase in either endogenous or exogenous PRL results in higher activity of the TIDA neurons, whereas a decrease in circulating PRL levels, resulting from hypophysectomy, immunoneutralization, or dopamine agonists, lowers their activity (reviewed in Refs. 1 , 105 , and 131). The TIDA neurons respond to both acute and chronic changes in PRL with few exceptions. The latter include pregnancy, lactation, and prolactinomas, when the dopaminergic neurons become refractory to the elevated PRL levels, thereby upholding physiological or pathological hyperprolactinemia.

The existence of a short loop feedback arrangement between PRL and dopamine raises several questions: Are the effects of PRL direct or indirect, and if indirect, what are the mediators? How does PRL gain access to the TIDA neurons? What are the nature and cellular distribution of hypothalamic PRL-Rs, and how are they regulated? Are both the TIDA and THDA/PHDA systems regulated by PRL? And finally, which functions of the dopaminergic neurons are affected by PRL? Presently, there are only partial answers to these questions.

The issue of direct vs. indirect effects of PRL on the TIDA neurons is difficult to resolve, given the scarcity of suitable in vitro systems. A direct effect is suggested by the increase in TH activity in cultured fetal hypothalamic neurons incubated with PRL (98). Although most of the TH-expressing neurons were immunopositive for the PRL-R, they constituted less than 10% of the total cell population in these cultures, and the receptors were also expressed by 20–25% of other neurons. In another study, dopamine concentration in TIDA neurons increased within 1 h after injecting ovine PRL to OVEX rats, indicating very rapid activation (123); a second delayed increase in dopamine turnover was seen in both the intermediate and neural lobes. Rapid activation by PRL of immediate early genes such as Fos-related antigens (132) and nerve growth factor (NGF)1-A (133) in TH-positive neurons in the arcuate nucleus also demonstrates PRL autofeedback. Neither study, however, conclusively established a direct effect of PRL, and mediation by substances such as neurotensin, NPY, and opioids cannot be ruled out.

Circulating proteins are excluded from the brain proper except for the circumventricular organs, e.g., the median eminence and neural lobe, which are outside the blood brain barrier. Given the location of the arcuate nucleus within the medial basal hypothalamus, it raises the question of how PRL reaches the TIDA neurons. PRL, which is detectable in many hypothalamic and extrahypothalamic sites (134, 135), can be derived from two sources: transport from the circulation, and local synthesis. Transport to the cerebrospinal fluid (CSF) occurs by receptor-mediated PRL uptake at the choroid plexus (136), which has the highest density of PRL-R in the brain (137, 138). The mechanism by which this receptor acts as a transporter is unknown. Upon gaining access to the CSF, PRL can be distributed to various sites, including the arcuate nuclei, which is adjacent to the third ventricle. The hypothalamus is also capable of de novo synthesis of PRL, a process that is regulated by an estrogen-sensitive mechanism (135). Although locally produced PRL may act as a mitogen for astrocytes (139), its participation in the short loop feedback on the TIDA neurons is doubtful.

The PRL-R belongs to the hematopoietic receptor family that includes GH, many cytokines, and some growth factors (for review, see Refs. 140, 141, 142). These are characterized by a single hydrophobic transmembrane domain that divides the receptor into an extracellular ligand binding domain and an intracellular domain. Features common to the extracellular domain include four paired cysteine residues and a wsxws (or WS) motif (tryptophan-serine-any amino acid-tryptophan-serine) that are involved in the formation of a ligand binding pocket. The cytoplasmic domains of the receptors differ in size and structure among the various family members. A hydrophobic proline-rich motif (homology box 1), located near the transmembrane region, is essential for signal transduction of all ligands studied. The PRL-R and several other hematopoietic receptors also contain a less-conserved cytoplasmic region, denoted box 2, whose function is not as well defined (143).

In the rat, alternative splicing generates two PRL-R isoforms, a short isoform of 291 amino acids and a long isoform of 591 amino acids. They have identical extracellular domains but differ in the length and sequence of the intracellular domain (Fig. 4). The promoter of the PRL-R gene belongs to a TATA-less/noninitiator class and has at least three regions that direct transcription from alternative sites in a tissue-specific manner (142). Both long and short receptor isoforms are expressed in most tissues, and their ratio is altered under many conditions (144). Although the exact function of each isoform remains to be fully defined, their coupling to different signal transduction pathways accounts for many of the pleiotropic actions of PRL (143). An "intermediate" form, lacking 198 amino acids due to a deletion, is uniquely expressed by the Nb2 rat T lymphocyte cell line (145) and confers growth dependence on PRL by these cells. A soluble form of the receptor, named PRL binding protein, resulting either from alternative splicing of the transcript or proteolytic cleavage of a membrane-bound receptor protein, has also been detected (146). The human PRL-R is encoded by a single gene that contains at least 10 exons and is located on chromosome 5 in close proximity to the GH receptor gene (147). In addition to the long form, a truncated isoform was identified in human breast cancer cell lines (148).

View larger version (32K): [in this window] [in a new window]   Figure 4. Diagram of the different types of PRL-Rs in the rat. The PRL-R belongs to the hematopoietic receptor family characterized by a single short TMD. The extracellular ligand binding domain has several common features, including two disulfide bonds and WS (Trp-Ser-X-Trp-Ser) motif. A proline-rich homology box 1, located near the transmembrane region, is essential for signal transduction. The long and short isoforms are the products of differential splicing, with the short isoform possessing a distinct C terminus. An intermediate isoform, unique to Nb2 lymphocytes, has a deletion of 198 amino acids. A soluble form of the receptor, termed "PRL binding protein," has also been detected. [Derived from C. Bole-Feysot et al.: Endocr Rev 19:225–268, 1998 (140 ). © The Endocrine Society.]

Unlike many growth factor receptors, the PRL-R has no endogenous kinase activity, using instead the Janus kinase (JAK)-signal transducer and activator of transcription (Stat) pathway as its major signal transduction mechanism (for reviews, see Refs.140 and 155). JAK2, which is constitutively associated with the PRL-R, is phosphorylated upon receptor activation by the ligand. In turn, the activated JAK2 phosphorylates other proteins, including the receptor itself and Stat proteins, e.g., Stat 1, Stat 3, and Stat 5 (( and b isoforms). The phosphorylated Stat proteins dimerize and translocate to the nucleus where they bind to specific DNA motifs, called -interferon-activated sites, in the promoter regions of PRL target genes. Other transduction pathways, including the MAPK cascade and Src kinases, have been implicated in PRL-R signal transduction (reviewed in Ref. 140).

The mRNA of the PRL-R is expressed in practically all tissues with the highest density seen in the liver, choroid plexus, ovary, and mammary gland (144). Next to the choroid plexus, the hypothalamus has the highest density of PRL-R in the brain, with a dissociation constant (K_d) of 0.2–0.3 nM and maximal binding (B_max) of 5–10 fmol/mg proteins (156). Many hypothalamic nuclei, including the rostral arcuate and periventricular nuclei, in which perikarya of the TIDA and PHDA neurons are located, express mRNA and contain immunoreactive PRL-R (157). However, the receptor is expressed only by some, but not all, TH-positive neurons (98, 157). The long isoform is predominant, but both isoforms appear to be regulated under some conditions (158).

The expression of hypothalamic PRL-R mRNA is altered under several physiological conditions without a clear delineation of the specific hormones involved. Expression of the long isoform is increased during proestrus, the second half of pregnancy, and in aged female rats (159, 160, 161). Alterations in circulating PRL levels during lactation (158) and after hypophysectomy (156) increase and decrease, respectively, the hypothalamic PRL-R mRNA levels, suggesting that PRL regulates its own receptors. Similarly, chronic treatment of OVEX rats with estrogens resulted in increased expression of PRL-R mRNA in the TIDA/THDA neurons, paralleling the rise in plasma PRL levels (157). This effect was most pronounced in the dorsomedial and rostral arcuate nuclei and was attenuated by cotreatment with progesterone.

Changes in hypothalamic PRL-R mRNA expression are accompanied by an increased level of receptor protein. For example, immunoreactive PRL-R is higher in the medial preoptic area and periventricular and arcuate nuclei during lactation than in diestrus (162). During lactation, the PRL-R protein becomes detectable in several sites, such as the lateral hypothalamic, supraoptic, paraventricular, suprachiasmatic, and ventromedial nuclei that are undetectable during diestrus. The long isoform is induced in both female and male rats sensitized to pups and exhibiting maternal behaviors (159, 160). In females, receptor induction is dependent on intact ovaries and pituitary because no changes are observed in OVEX or hypophysectomized females when maternal sensitization occurs. In males, this induction is facilitated by PRL administration and is suppressed by cohabitation with females.

TH is one of the targets of PRL action within the hypothalamus, and both short-term enzyme activation (107, 163) and long-term enzyme induction (164, 165) have been observed. PRL immunoneutralization reverses the estrogen-induced increase in dopamine turnover in the median eminence and the intermediate lobe, albeit at different times during the PRL surge (166). In contrast, the observed effect of PRL immunoneutralization on the THDA system is minimal. Another laboratory found that PRL antiserum suppressed basal and haloperidol-induced increases in TIDA activity within 1–2 h, establishing a rapid action of PRL (167). This feedback mechanism is operational in both males and cycling females but is attenuated during late lactation, commensurate with the requirement for prolonged maintenance of elevated PRL levels during this time (168). Chronic hypoprolactinemia and hyperprolactinemia reduce and augment, respectively, TH mRNA levels in the arcuate nucleus but not in the substantia nigra, establishing the site specificity of enzyme induction (107).

PRL also exerts tropic effects on dopaminergic neuronal differentiation, best illustrated in animals with inherited PRL deficiency. Spontaneous mutations of the Snell and Ames dwarf mice result in the absence of lactotrophs, somatotrophs, and thyrotrophs and their respective hormones. The Snell mouse has mutations in the Pit-1 gene (169), a pituitary-specific transcription factor essential for differentiation and maintenance of the three pituitary cell types. The Ames mouse has a mutation in the Prop-1 gene, which normally activates Pit-1, leaving the animals with residual numbers of lactotrophs and somatotrophs (170). Homozygous dwarfs are growth retarded, sterile, and unable to lactate. Dopamine levels are severely depressed in both the TIDA and THDA systems, but not in other brain regions, with no change in norepinephrine (171).

Notably, the TIDA neurons do not differ between dwarfs and normal siblings until d 21 of age. In wild-type mice, dopamine levels and TH-positive neurons continue to increase, whereas those in the dwarfs are unchanged or decline (reviewed in Ref. 172). The dopamine deficiency can be reversed by PRL therapy, but only if initiated at an early postnatal period; adult dwarfs are refractory to PRL. This indicates that PRL acts as a neurotropic factor for the TIDA neurons during a specific developmental window of time. An unresolved issue is whether early postnatal TIDA development is totally independent of PRL or is supported by PRL that is provided by the maternal milk (173).

PRL functions as both a mitogen and survival factor in mammary and immune cells (for review, see Refs. 174, 175, 176), and may have similar functions in the brain. Astrocytes are the most numerous cells in the brain. Unlike neurons, they are capable of proliferation in adulthood. Astrocytes often function as immunocompetent cells and react to brain injury by increased proliferation, cytokine release, and antigen presentation. There is evidence that PRL, but not GH, induces proliferation of growth-arrested astrocytes. The mitogenic effects of PRL require serum-derived factors, suggesting that it functions as a coactivator (139). PRL temporally increases expression of tumor necrosis factor-, IL-1, and TGF- in astrocytes, which in turn may synergize with PRL in promoting astrocyte proliferation and secretory activity (177). The mitogenic/secretory effects of PRL appear to be mediated by the JAK2/Stat pathway, providing the first evidence that PRL signaling within the brain is the same as in peripheral tissues (178, 179). It remains to be determined whether the action of PRL on astrocytes directly or indirectly impinges on the TIDA neurons.

D. Effects of ovarian steroids
Estrogens affect PRL homeostasis at several anatomic sites that include the hypothalamus, posterior pituitary, and anterior pituitary (for review, see Ref. 105). The importance of estrogens in the control of the TIDA neurons is underscored by the higher basa