Additional related information may be found at:
Neuropsychopharmacology: The Fifth Generation of Progress

Back to Psychopharmacology - The Fourth Generation of Progress

The Noradrenergic Receptor Subtypes

Magdalena Wozniak, Nicole L. Schramm and Lee E. Limbird

HISTORICAL PERSPECTIVE ON ADRENERGIC RECEPTORS

The modern concept of a- and b-adrenergic receptors (ARs) was introduced in 1948 by Ahlquist (1), who studied the effects of catecholamines on a variety of physiological responses. These included contraction and relaxation of the uterus, dilation of the pupil and stimulation of myocardial contraction. He demonstrated that norepinephrine (NE), epinephrine (Epi), isoproterenol (Iso), methylnorepinephrine, and methylepinephrine could cause either contraction or relaxation of smooth muscle, depending on the amine, its dose and the site of action. For instance, Ahlquist reported that isoproterenol and norepinephrine caused smooth muscle relaxation and contraction, respectively, while epinephrine could cause both contraction as well as relaxation of smooth muscle. Based on these observations, Ahlquist proposed that the effects of the amines were mediated by two distinct receptors, designated a-AR and b-AR. b-ARs were defined by the catecholamine potency series of Iso > Epi > NE, whereas a-ARs were defined by the series of Epi = NE > Iso. This pharmacological concept of distinct catecholamine receptors was further solidified with the development of highly selective antagonists, including propranolol for the b-AR and phentolamine for the a-AR.

In 1974, it was postulated that a-ARs could be subdivided into two subtypes: a1-AR and a2-AR. This classification was based on anatomical localization, whereby a1-ARs were located on postsynaptic membranes, and a2-ARs were located on presynaptic nerve terminals (48a;72a). Subsequently, a functional classification of a-ARs was proposed, defining a1-ARs as postsynaptic, excitatory receptors found on vascular smooth muscle and a2-ARs as presynaptic receptors that attenuated NE release from the sympathetic nerve terminals. However, subsequent data showed that a2-ARs are also located postsynaptically (70) and in non-synaptic settings, such as human platelets, adipocytes, and vascular smooth muscle.

We now know that there are multiple a1 and a2 AR subtypes, members of structural subfamilies of G protein coupled receptors that are encoded by distinct genes. The current subclassification of a-ARs is based on the differential affinities of agonists and antagonists in tissue responses and in radioligand binding studies (11,70). Among a-AR agonists, phenylephrine is selective for a1-ARs, clonidine for a2-ARs; the endogenous agonists epinephrine and norepinephrine are not subpopulation selective. Historically, the a-AR antagonist prazosin was described as selective for a1-ARs and yohimbine for a2-ARs (70,79). However, recent studies have demonstrated that there are multiple a2 AR subtypes, some of which also are potently antagonized by prazosin (10).

a1-ADRENERGIC RECEPTORS

Pharmacological Definition and Molecular Cloning of a1-Adrenergic Receptor Subtypes

In the late 1980s, two subtypes of a1-ARs(a1A-AR and a1B-AR) were defined based on pharmacological criteria (10,11,59,70). The a1A-AR can be distinguished from the a1B-AR by the higher affinity of the a1A-AR for the agonists NE, Epi, phenylephrine, oxymetazoline and methoxamine, compared with the a1B-AR. In addition, a1A-ARs display higher affinity for the a-AR antagonists WB4101, phentolamine, and 5-methyl-urapidil, when compared with a1B-ARs. In contrast, chlorethylclonidine (CEC), an alkylating agent, irreversibly inactivates only a1B-AR binding sites.

Some reports suggested the existence of yet another subtype of a1-ARs, termed a1D-AR. Radioligand binding studies as well as photoaffinity labeling strategies indicate that the a1D-AR shares some of the pharmacological characteristics with a1A-AR, including its higher affinity for WB4101 and 5-methyl-urapidil, compared with the a1B-AR. However, the a1D-AR subtype, unlike a1A-AR subtype, can be inactivated by CEC. Pharmacological characterization of a1-AR subtypes has revealed tissue-specific distribution of receptor subtypes: liver and spleen express only a1B-ARs; in contrast, both brain and heart express both a1A-AR and a1B-AR subtypes.

The cloning of cDNAs for three a1-AR subtypes (tentatively designated as a1a-AR, a1b-AR and a1c-AR subtypes) has confirmed pharmacological evidence for multiple a1-AR subtypes and led to an analysis of structural differences between these three receptor subpopulations (30,34,59,70). (Lowercase subscripts are used to describe cloned receptor subtypes until their identity with pharmacologically characterized subtypes is confirmed.) Table 1 summarizes the cDNA designation, relative pharmacological specificity and tissue distribution of the a-AR subtypes. The cDNAs encoding the a1a-AR, a1b-AR and a1c-AR subtypes originally were cloned from libraries created from rat cerebral cortex, Syrian hamster smooth muscle DDT1 MF2 cells, and bovine brain, respectively. The independent chromosomal locations of the genes encoding these cDNAs are consistent with the interpretation that a1-AR subtypes are the products of different genes. Thus, the genes encoding the a1a-AR and a1b-AR subtypes are located on human chromosome 5, locus q23-32, while the gene encoding the a1c-AR subtype is located on human chromosome number 8. The a1b-AR clone contains a 1545-bp open reading frame which encodes a protein of 515 amino acids, while the a1a-AR clone has an open reading frame of 1680 bp encoding a protein of 560 amino acids. In the case of the a1c-AR gene, its open reading frame encodes a polypeptide of 466 amino acids with a molecular weight of 51 kDa. Shortly after the clone assumed to encode the a1A-AR subtype was isolated, an essentially identical clone was isolated from a rat hippocampal library. This novel clone was initially thought to encode a novel a1-AR subtype and was thus designated a1d-AR. Upon closer analysis, it became clear that the two independently isolated a1a-AR and a1d-AR clones were identical, with the exception of two codons, and thus they are now believed to encode the same a1-AR subtype—the a1A-AR.

A thorough characterization of each cloned a1-AR subtype led to the realization that the nomenclature for the cloned a1-AR subtypes does not correspond to that for the pharmacologically identified a1-AR subtypes (34). In addition, at least one of the cloned a1-AR subtypes (the a1D-AR subtype) had not been previously identified pharmacologically. This led to a confusion regarding the identity, and therefore the nomenclature, of the pharmacologically identified vs. the cloned a1-AR subtypes (Table 1). Currently, it is apparent that the pharmacologically defined a1A-AR, known for its role in mediating vasoconstriction in rat mesenteric and renal arteries, is indeed encoded by the cDNA designated, at its time of cloning, as the a1c-AR.

The a1-AR subtypes also demonstrate selectivity in interacting with a variety of agonist and antagonist agents. A number of compounds, including 5-methyl-urapidil, oxymetazoline, A-61603, a modified (+)niguldipine compound (SNAP 5089), KMD-3213 and RS17053 display higher affinity for the a1A-AR than for the other a1-AR subtypes (Table 1). On the other hand, the cloned a1b-AR was indistinguishable from the pharmacologically defined a1B-AR, which selectively recognizes compound AH11110A, spiperone and respiperone. The cloned a1d-AR, however, which was initially labeled as the a1a-AR and later as a1a/d-AR, is now known to represent a novel, not previously pharmacologically identified, a1-AR subtype. This novel subtype, expressed in vascular smooth muscle, cerebral cortex and rat lung, was recently designated pharmacologically as the a1D-AR; BMY7378 and SKF105854 are considered a1D-AR-selective antagonists.

Some researchers favor the opinion that a1A-AR, a1B-AR and a1D-ARs are not the only subtypes of a1-AR. These three a1-AR subtypes are known to bind prazosin with high affinity, thus comprising the so called a1H-AR group; atypical a1-AR binding sites that bind prazosin with low affinity and are referred to as a1L-AR sites (34). Moreover, two types of sites can be further distinguished among these a1L-ARs; a1L-ARs: defined by binding yohimbine and the a1-AR antagonist HV732 with low affinity, and a1N-ARs, which are characterized by moderate affinity for yohimbine and high affinity for HV732. These potential novel a1-AR subtypes have yet to be thoroughly characterized pharmacologically, biochemically, or physiologically and have not been cloned.

Structure of a1-Adrenergic Receptors

The a1A-AR, a1B-AR and a1D-AR subtypes belong to the G protein-coupled receptor superfamily. Although members of this superfamily are known to bind various ligands and mediate multiple physiological processes, they share a highly analogous seven transmembrane-spanning topography: seven hydrophobic transmembrane-spanning domains consisting of 20–28 amino acids are separated by hydrophilic intra- and extracellular loops (34). There is a remarkable 75% sequence similarity in the seven putative transmembrane domains of the a1-AR subtypes. The first and second extracellular loops each contain single cysteine residues that form a disulfide bond, likely critical for correct protein folding and posttranslational modification (17,34). The membrane-proximal regions of the second and third intracellular loop are involved in receptor-G protein coupling. The length of the third intracellular loop is similar for the three subtypes of a1-ARs, although the amino acid sequences are different; the length of the intracellular carboxyl terminus is almost identical in the a1D-AR and a1B-AR. Although there is very little sequence homology in this region among the a1-AR subtypes, the a1-AR carboxy termini contain consensus sites for phosphorylation by serine/threonine protein kinases. In contrast, there are marked differences in the length of the extracellular amino terminus among the three receptor subtypes, with the a1D-AR amino terminus being much longer (about 90 amino acids) than the a1A-AR (25 amino acids) or the a1B-AR amino terminus (42 amino acids). The extracellular amino terminus also contains potential sites for N-linked glycosylation in all three a1-AR subtypes. Even though glycosylation has only been demonstrated directly for the a1B-AR to date, the a1D-AR subtype is thought, on the basis of its migration pattern on SDS-PAGE, to also be glycosylated. The functional relevance of these biochemical differences among a1-ARs has yet to be clarified.

As indicated above, regions within the third intracellular loop of a1-ARs are important for receptor-G protein coupling. The construction of a variety of receptor mutants helped define a 27-amino acid region (residues 233–259) in the N-terminal portion of the third intracellular loop of the a1B-AR, which is involved in the receptor coupling to phospholipase Cb (PLC) [17]. Comparison of the amino acid sequences revealed that this region of the a1B-AR is 62% and 55% homologous to the corresponding regions of the a1A-AR and a1D-AR, respectively. Except for this short region of sequence, the amino acid sequence in the third loop of the a1-ARs is quite distinct. When selected fragments of the third intracellular loop of the a1B-AR were introduced by recombinant DNA substitution into the corresponding regions of the cDNA encoding the b2-AR, the resultant b2-AR/a1-AR chimeras were able to activate phosphoinositide hydrolysis, in distinct contrast to coupling to Gs and stimulation of adenylyl cyclase characteristic of the native b2-AR. These studies indicated that the region of the third intracellular loop of the a1-AR substituted into the b2-AR (amino acids 233–254 in the membrane proximal segment of the carboxy-terminus of the third intracellular loop) is an important determinant in the selectivity of receptor-G protein coupling (17). In addition to the 27-amino-acid stretch of the a1B-AR described above, a 23-amino acid fragment present at the carboxy terminal end of the third loop also is necessary for productive receptor-G protein coupling. The sequence involved in the interaction between the receptor and G protein resides in the N-terminal fragment of the third loop, while the C-terminal region of the loop modulates the efficacy of receptor coupling to PLC (17). Both agonist affinity for the a1B-AR and agonist potency in activating PLC are increased by 1–2 orders of magnitude after single amino acid substitutions (Ala293 to Leu, Lys290 to His) in the C-terminal portion of the third intracellular loop (17).

Site-directed mutagenesis studies have revealed amino acid residues crucial for determining a1-AR subtype selectivity. When both Ala204 in the fifth transmembrane domain and Leu314 in the sixth transmembrane region of the a1B-AR were mutated to the corresponding Val and Met residues found in the a1A-AR, the agonist binding profile was reversed from a1B-AR-like to a1A-AR-like. Conversely, the substitution of Ala204 to Val185 and Leu314 to Met293 in the a1A-AR led to the a1B-AR-like agonist binding profile. These findings implicate these two residues, and sequence regions, as critical for defining binding selectivity at a1A-AR vs. a1B-AR subtypes (34).

Tissue Distribution of a1-AR Subtypes

Radioligand binding techniques have been used to quantitate a1-ARs and to characterize their regulation in various tissues and cell culture systems (11,70,72,81). Radioligands that have been employed to study a1-ARs include the antagonists [3H]prazosin, [3H]WB4101, [125I]HEAT, and [3H]dihydroergocryptine (DHEC). The a-AR agonists [3H]epinephrine and [3H]norepinephrine also have been used to measure a1-ARs, although to a lesser degree. a1-ARs have been demonstrated, based on both radioligand and pharmacological data, in the liver, heart, vascular smooth muscle, brain, spleen and other tissues; subtypes have been distinguished based on their differential affinities for the antagonist WB4101 and their sensitivities to inactivation by CEC. Analysis of various rat brain regions showed that thalamus and cerebral cortex had the highest proportion of the low-affinity subtype for WB4101 (a1B-AR), whereas hippocampus and pons-medulla had the highest proportion of high-affinity WB4101 binding sites (a1A-ARs). The a1A-AR also was found in the rat thalamus and cerebral cortex, while the a1B-AR was detected also in the hippocampus. Studies also have characterized human a1-AR subtype expression in human tissues. Both a1A-AR and a1B-AR subtypes can be identified in human brain (70,72). Two a1A-AR selective agents, 5-methyl-urapidil and (+)-niguldipine, were utilized to resolve the relative contributions of these subtypes in membrane preparations derived from human frontal, temporal and parietal lobe cortical tissue obtained during surgery or autopsy. Approximately 40–50% of binding detected appeared to correspond to the a1A-AR subtype, generating interest in revealing the physiological role of multiple a1-AR subtypes in human brain.

Detection of mRNA encoding a1-AR subtypes by a variety of strategies demonstrated that a1D-ARs in rat tissues are expressed in the vas deferens, hippocampus, cerebral cortex, brain stem, aorta, heart and spleen, while in the human brain this subtype was found in the cerebral cortex and cerebellum (34,72). The a1B-AR subtype is found in rat liver, heart, cerebral cortex, hippocampus, brain stem, kidney, lung and spleen. In the human brain, the a1B-AR is expressed in the cerebral cortex and cerebellum. Expression of the a1A-AR was detected in human and rat heart and cerebral cortex, rat hippocampus, lung and vas deferens, and human cerebellum, dentate gyrus and pituitary gland. In general, the distribution of mRNA encoding a1A-AR and a1B-AR in different tissues correlates well with their pharmacologically determined localization.

a1-Adrenergic Receptor Signal Transduction Pathways

Early studies examining a1-AR subtypes suggested that the a1A-AR and a1B-AR were coupled to distinct second messenger systems controlling intracellular Ca2+ levels. However, as the literature has matured, it is clear that all three subtypes (a1A-AR, a1B-AR and a1D-AR) are able to mobilize Ca2+ from intracellular stores as well as increase extracellular Ca2+ entry via voltage-gated Ca2+ channels. Stimulation of all three a1-AR subtypes leads to the hydrolysis of membrane phospholipids via a G protein-mediated activation of phospholipase Cb (70). The resultant inositol triphosphate (IP3) mediates the a1-AR-elicited Ca2+ release from intracellular stores, thereby increasing cytosolic Ca2+ concentration. The simultaneous production of diacylglycerol (DAG) activates protein kinase C (PKC). PKC also is activated by a group of Ca2+ and calmodulin-sensitive protein kinases. Active PKC phosphorylates many cellular substrates, including membrane channels, pumps, and ion-exchange proteins such as the Ca2+ ATPase, leading to changes in plasma membrane conductances for various ions and presumably facilitating an influx of extracellular Ca2+, further increasing its cytoplasmic concentration. The a1-ARs also have been reported to modulate other signaling pathways, including increased cyclic AMP and cyclic GMP accumulation, potentiation of cyclic AMP responses elicited by Gs-linked receptors, activation of phospholipase A2 and phospholipase D, activation of cyclic AMP phosphodiesterase, release of adenosine, and stimulation of arachidonic acid release (34).

The recent subclassification of a1-ARs into a1A-AR, a1B-AR and a1D-AR subtypes has raised questions concerning the selectivity of different a1-AR subtypes for interaction with various G proteins. Indeed, different subtypes of a1-AR were shown to couple to different members of the Gq/G11 class of G proteins. The a1D-AR appears to couple only to Gq and G11, whereas the a1B-AR mediates phosphoinositol turnover via coupling to Gq, G11, G14 and G16 (34). Interestingly, the a1B-AR was also shown to couple to pertussis toxin-insensitive Gh protein. Gh is a novel signal mediator consisting of a 74-kDa GTP-binding a-subunit and a 50 kDa b-subunit. Unlike the heterotrimeric G proteins, Gh is a multifunctional protein that combines transglutaminase and receptor signaling functions (34).

The genes encoding the a1-AR subtypes have been characterized as protooncogenes involved in cellular growth and differentiation (17). The first indication that the a1B-AR gene may contribute to proliferation came from studies demonstrating that mutations of the intracellular loops of the a1B-AR that resulted in ligand-independent receptor activity and G protein activation also elicited neoplastic transformation of fibroblasts following transfections with the "constitutively active" a1B-AR structures. These studies suggest that mutations occurring spontaneously in this receptor gene could initiate tumorigenesis in human beings.

Regulation of a1-Adrenergic Receptors and a1-AR Function

Continued exposure to an agonist often results in a diminished response to that agonist, a process termed desensitization. Agonist-induced desensitization has been divided into two categories, referred to as agonist-specific or homologous, and agonist-nonspecific or heterologous. The term homologous designates the form of desensitization involving diminished response to the desensitizing agent only, while the efficacy of other receptor activators is unaltered. In contrast, heterologous desensitization occurs when an agonist attenuates a response mediated by a broad range of ligands acting via distinct receptors.

Stimulation of a1-AR leads to a desensitization of response and a decrease (downregulation) of cell surface receptor density (76,17). The activation of protein kinase C by phorbol diesters causes phosphorylation and functional uncoupling of a1-ARs from signal transduction pathways in cultured smooth muscle cells. It is not yet known whether the desensitization and downregulation processes differ qualitatively or quantitatively among the receptor subtypes.

In contrast to the wealth of information regarding the molecular basis for homologous as well as heterologous desensitization of adenylyl cyclase-coupled receptors (see Dopamine Receptors: Clinical Correlates), little is known about the mechanisms underlying the desensitization of PLC-coupled receptors. Recent studies of the a1B-AR provided evidence that the carboxy terminus of the receptor becomes phosphorylated in parallel with signal diminution in a PKC-independent fashion.

The density of a1-ARs is dynamically regulated in various physiological and pathological conditions, perhaps contributing to changes in cellular responsiveness to a1-AR agonists. For example, a decrease in a1-AR density was observed in the brain of the aging rat; a similar age-related change was reported in the heart and spleen (36).

Certain pathological conditions are associated with changes in expression of the a1-AR. For example, the density of cardiac a1-ARs was significantly decreased in the ventricles, but not in the atria of spontaneously hypertensive rats (86). A similar decrease in the level of a1-AR density was observed in the kidney of both the hyperactive and hypertensive rat induced by deoxycorticosterone and salt. By contrast, the density of a1-ARs was significantly increased in some brain regions, such as the midbrain and cerebral cortex, with hypertension but was unchanged in others such as the hippocampus, hypothalamus, cerebellum and brain stem (81).

Changes in thyroid status alter the functional potency of various agonists for cardiac a1-ARs; hyperthyroid rats express a lower level of a1-ARs in the ventricles than euthyroid animals and also are less sensitive to ambient agonist concentrations, whereas the density of a1-ARs is increased in the ventricles of hypothyroid rats and this is paralleled by an enhanced sensitivity to a1-AR agonists. These effects of thyroid status, however, are tissue-specific, since vas deferens, caudal artery and cerebral cortex express a similar density of a1-ARs regardless of thyroid status (29). Other endocrine perturbations also alter a1-AR responsiveness, including diabetes.

Other examples of regulation of a1-AR responsiveness exist. For example, electrophysiological techniques have revealed an enhanced a1-AR responsiveness following ischemia and reperfusion of the myocardium that is paralleled by an increased a1-AR density (16).

Physiological Effects of a1-Adrenergic Receptors

a1-ARs are a part of the sympathetic nervous system and therefore are involved in the control of various physiological functions (70). However, the most prominent a1-AR-mediated physiological responses occur in the periphery, not in the CNS. Thus, a1-ARs are crucial in the stimulation of smooth muscle contraction; in the vasculature, these receptors evoke increased arteriolar and venous constriction leading to an increase in peripheral resistance to blood flow. Activation of a1-ARs in the heart increases cardiac output and force of contraction. Since blood pressure is influenced by both cardiac output and peripheral resistance, the a1-AR-stimulated increase in both parameters leads to an elevation of blood pressure.

Contraction of smooth muscle sphincters in the stomach, intestine and bladder, as well as contraction of the uterus, also are mediated via a1-ARs. In contrast, stimulation of a1-ARs in intestinal smooth muscle results in membrane hyperpolarization and muscle relaxation.

The a1-ARs modulate a variety of metabolic responses. For example, the liver a1-ARs mediate the breakdown of glycogen (glycogenolysis) and the synthesis of glucose (gluconeogenesis). Activation of a1-ARs also results in decreased pancreatic secretions and enhanced salivary (K+ and water) and sweat gland secretions.

a1 Adrenergic Receptors in the Brain

Catecholamine-rich brain regions, including the nucleus tractus solitarius, dorsal vagal complex and hypothalamus, are involved in the maintenance of cardiovascular function. Lesions of the nucleus tractus solitarius result in a chronic increase in blood pressure. a1-AR agonists microinjected into the area of the nucleus tractus solitarius evoke a concentration-dependent reduction in blood pressure, whereas injections of NE into the hypothalamus elevate blood pressure. These observations are consistent with the hypothesis that central a1-ARs play a role in blood pressure regulation, corroborated by the observation that a1-AR density and responsiveness parallel certain pathophysiological states, including hypertension.

Central a1-ARs have been implicated in the regulation of hormone secretion. Norepinephrine and other biogenic amines modulate the hypothalamic control of growth hormone secretion. The a1-ARs in the hypothalamic ventromedial nucleus are involved in regulation of satiety and body weight; of interest are observations that a1-AR density in the ventromedial nucleus is inversely proportional to body weight gain in rats (84).

Clinical Relevance of a1-Adrenergic Receptors

The clinical applications for a1-ARs were long thought to involve only a1-AR antagonists as antihypertensives and a1-AR agonists as nasal decongestants. Particularly well known is the use of the selective a1-AR antagonists terazosin and doxazosin in the treatment of hypertension. Recently, however, several newer therapeutic applications are being evaluated, including depression and prostatic hypertrophy (71).

The activation of the a1-ARs in the CNS may be associated with an alerting or antidepressant action. The a1-ARs in the brain can be stimulated with a systemically administered compound (SDZ NVI 085) that does not concomitantly stimulate vascular a1-ARs and thus avoids untoward increases in blood pressure. This a1-AR agonist crosses the blood-brain barrier and produces alerting effects in normal rats and monkeys. SDZ NVI 085 also prevented the behavioral and learning deficits induced by noradrenergic neurotoxins. Particularly encouraging studies on human volunteers concluded that subjects reported positive behavioral effects at oral doses that did not cause an increase in blood pressure (71). Due to its lipophilicity, SDZ NVI 085 may produce central a1-AR-mediated effects without increasing blood pressure. However, this is not a sufficient explanation, since a number of lipophilic a1-AR agonists known to produce CNS effects in rodents also increased blood pressure in these animals.

Most recently, novel a1-AR antagonists are being utilized and refined for the treatment of benign prostatic hypertrophy (BPH). Activation of a1-ARs localized in the stromal smooth muscles significantly contributes to urethral obstruction resulting from an enlarged prostate gland. The a1-AR antagonists prazosin, terazosin, doxazosin, alfuzosin, bunazosin and tamulosin are being used currently to relieve the symptoms of BPH. One novel a1-antagonist that may prove useful in the treatment of BPH is SB 216 469, known to produce a selective blockade of norepinephrine-induced contraction in the prostate artery but not in the general vasculature. Other compounds of interest include SL 89.0951, which shows in vivo selectivity for urethral vs. blood pressure responses, and naftopidil. Further subclassification of a1-ARs may offer a route to achieving greater tissue and organ selectivity. For example, the identification of the a1A-AR subtype has led to the development of selective a1A-AR antagonists that may be more effective in the treatment of BPH than the non-subtype selective blockers.

Another area of current drug discovery is the differential sensitivity of various blood vessels to prazosin, a fact that exploits the potentially pharmacologically useful existence of prazosin-insensitive a1-ARs.

a2-ADRENERGIC RECEPTORS

Pharmacological as well as molecular cloning strategies have revealed three sub-subtypes of a2-ARs, now known as the a2A-AR, a2B-AR, and a2C -AR subtypes. The genes encoding these receptors (all of which appear to be intronless) have been cloned from multiple species, allowing extensive structural, functional and biochemical characterization (70). The human forms of these receptors are called a2-C10, a2-C2, and a2-C4, respectively, to indicate their chromosomal location. The a, b, c designation will be used in this chapter to refer to the genes; the A, B, C designation will be used to refer to the proteins (gene products), which are the pharmacologically defined subtypes.

Pharmacological Selectivity of a2-Adrenergic Receptor Subtypes

The endogenous agonists of a2-ARs are norepinephrine (noradrenaline) and epinephrine (adrenaline), and they are recognized with similar affinity by the three a2-AR subtypes. Table 2) provides the pharmacological specificities and structures of agents that interact with various a2-AR subtypes. The values given for ligands were obtained in competition binding studies for radiolabeled antagonists and have been corrected for radioligand concentration and affinity. All three a2-AR subtypes are stimulated by the agonists UK 14304, clonidine, and apraclonidine. Guanfacine and guanabenz preferentially stimulate the a2A-AR but have some activity at all three subtypes. Oxymetazoline is an agonist with varying efficacy, depending on the physiological preparation, that demonstrates preferential interaction with the a2A-AR. Like oxymetazoline, the efficacy of clonidine and its analog UK 14304 as full or partial agonists depends on the target cell, perhaps consistent with findings that a2-AR receptor reserve varies considerably in different physiological settings (58).

All three receptors are blocked by the plant-derived antagonists yohimbine and rauwolscine. Prazosin, on the other hand, has a 30-fold higher affinity for the a2B-AR and a2C-AR subtypes than for the a2A-AR. Similarly, ARC 239 and spiroxatrine have a higher affinity for the a2B-AR and a2C-AR subtypes than for the a2A-AR. BAM 1303 and WB 4101 are two antagonists that have a much higher affinity for the a2C-AR than for the a2B-AR. The antagonist BRL 44408 has a higher affinity for the a2A-AR than for the a2B-AR and a2C-AR. Dose ratios of these antagonists can be used to discriminate among receptor subtypes in vitro and in some isolated organ studies. However, the relatively unknown and probably differing bioavailability of these agents in vivo has precluded their use in the unequivocal clarification of the role of a2-AR subtypes in various physiological responses.

Early pharmacological studies seemed to indicate the existence of a fourth a2-AR subtype, the a2D-AR, in the bovine pineal gland, the rat submaxillary gland and a cell line derived from a rat pancreatic islet cell tumor. This receptor has a lower affinity for rauwolscine and yohimbine than the other receptors and an a2A-like affinity for prazosin, spiroxatrine and ARC 239. Extant data, however, are more consistent with the interpretation that the bovine and rodent a2D-ARs are species homologs of the a2A-AR (50,62).

Structure of the a2-Adrenergic Receptors

The a2-ARs, like all catecholamine-binding receptors cloned to date, are predicted to be seven-transmembrane-spanning molecules. They most commonly mediate their effects via interaction with the Gi/Go family of heterotrimeric G-proteins. There are subtle differences in the biochemical properties of the three receptors, as shown schematically inFigure 1a. Some structural and biochemical differences among the three subtypes of a2-Adrenergic Receptors.a). Note that the N-terminal extracellular domains of the a2A-AR and a2C-AR contain asparagine residues which serve as substrate sites for N-linked glycosylation. The a2A-AR and a2B-AR both contain cysteine residues that can be palmitoylated in their C-terminal intracellular domains. The functional relevance of these post-translational modifications is unclear, although in contrast to findings for the b2-AR, elimination of acylation of the a2A-AR does not alter the efficiency of receptor-G-protein coupling (43).

As shown in Figure 1b. Human a2-Adrenergic Receptor, see text for details.b, several structural regions of the a2-ARs have been assigned particular functions, based on studies performed with mutant a2-ARs and chimeras of sequence regions of the a2A-AR fused to complementary regions of other G-protein coupled receptors (GPCRs). The region of the receptor that lies in the lipid bilayer is responsible for binding catecholamines and the variety of agonist and antagonist ligands that bind to the a2-ARs (85). Biochemical and recombinant DNA strategies have shown that a2-AR coupling to G-proteins is mediated by stretches of sequence that lie in the second intracellular loop and very near to the bilayer in the third intracellular loop, projecting from predicted transmembrane (TM) domains 5 and 6 (18,23). The sequences responsible for G-protein activation are predicted to exist as amphipathic a-helices, based on computer-assisted analysis of secondary structure as well as insertional mutagenesis strategies performed on the analogous M2 muscarinic receptor (52).

The third intracellular loop is a region of significant diversity among the three a2-AR subtypes. This region possesses the serine/threonine substrate sites for phosphorylation by the G-protein coupled receptor kinases (GRKs), which recognize agonist-occupied receptors as substrates for phosphorylation (25). In the a2B-AR, the third intracellular loop possesses a highly acidic sequence, which is required for GRK-mediated phosphorylation, but the phosphorylation sites themselves have not been mapped (39). Removal of this entire loop from the a2A-AR accelerates its removal from the cell surface (42), suggesting either that this region interacts with proteins that stabilize its surface expression or with proteins that serve to block receptor endocytosis from the plasma membrane.

Within the bilayer, particular residues have been implicated in ligand recognition by a2-ARs (Figure 1a. Some structural and biochemical differences among the three subtypes of a2-Adrenergic Receptors.a). By analogy with the b2-AR, aspartate 113 (in the a2A-AR), predicted to lie near the top (exofacial surface) of TM 3, is predicted to interact with the protonated amine of catecholamines (73). Phenylalanine 412, which lies in TM 7, is critical for recognition of a2-AR antagonists like yohimbine; a2/ b2 receptor chimeras possessing TM 7 contributed by the a2-AR possess yohimbine-blockable isoproterenol- (a b2-AR agonist) stimulated adenylyl cyclase activity (45). The species differences between human a2A-AR and rodent a2D-AR have been localized to residue 201, which is a serine in the mouse a2D-AR and a cysteine in the human a2A-AR. Site-directed mutagenesis of this residue in the a2A-AR to the corresponding residue in the a2D-AR, and vice-versa, switches the yohimbine affinity of the receptor to that of its species homolog (50).

Tissue Distribution of a2-AR Subtypes

a2-ARs are widely distributed throughout the central nervous system (CNS). Most studies have been performed using in situ hybridization to localize the cell bodies expressing the mRNA encoding the a2-AR subtypes (61,82). Although some studies have been performed in postmortem human brains, the most detailed regional expression has been explored in the rodent brain. In some cases, studies of mRNA localization have been complemented by in situ radioligand binding analysis (78) and, although less frequently, by immunochemistry (69).

Localization studies have revealed that mRNA encoding the a2A-AR subtype is expressed in multiple nuclei of the brain stem and pons, the midbrain, the hypothalamus, the septal region, amygdala, olfactory system, hippocampus, cerebral cortex, spinal cord, and, to a lesser extent, in the cerebellum (82). The mRNA encoding the a2A-AR is particularly enriched in the locus ceruleus and in nuclei of the brainstem implicated in central control of blood pressure, such as the nucleus tractus solitarius (NTS) and the ventrolateral medulla (VLM) [61].

The mRNA encoding the a2B-AR is expressed in the thalamus, olfactory system, pyramidal cell layer of the hippocampus, some regions of the basal ganglia, and the cerebellar Purkinje layer. At least in the rodent, there is not considerable overlap in a2A-AR and a2B-AR mRNA expression (82).

The mRNA encoding the a2C-AR subtype is expressed in some regions of the brain stem and pons, midbrain, thalamus, amygdala, olfactory system, hippocampus, cerebral cortex, basal ganglia and dorsal root ganglia. The expression of the mRNA encoding the a2A-AR and a2C-AR subtypes overlaps extensively, although there are CNS loci where a2A-AR mRNA is expressed without significant co-expression of the a2C-AR mRNA. In contrast, only in the basal ganglia is the a2C-AR mRNA expressed without significant co-expression of the a2A-AR mRNA (61,82).

These regions of mRNA expression, particularly for the well-studied a2A-AR subtype, correlate well with the CNS regulation of behavioral and physiological functions delineated previously by in vivo studies, as summarized below.

a2-Adrenergic Receptor Signal Transduction Pathways

As mentioned earlier, most known effects of the a2-ARs are mediated by heterotrimeric GTP-binding proteins of the pertussis toxin-sensitive Gi/Go subfamily. The three a2-AR subtypes also share multiple signaling pathways (49). Three of these signaling pathways often exist in the same cell, particularly in neurons or neuroendocrine tissue: 1) a2-ARs inhibit adenylyl cyclase via Gi and thereby inhibit the production of cyclic adenosine monophosphate (cAMP). 2) a2-ARs, via Go proteins, suppress voltage-activated calcium channels, thus reducing the flow of extracellular Ca2+ into target cells; and 3) a2-ARs increase the conductance of K+ ions through receptor-operated, inwardly rectifying potassium channels, an effect mediated predominately if not exclusively via the bJ subunits of Gi proteins. All three of these activities (inhibition of adenylyl cyclase, suppression of voltage-sensitive calcium channels, and stimulation of potassium channels) can contribute to the reduction of neurotransmitter or hormone release (48).

In one physiological setting documented to date, a2-ARs regulate responses via pertussis-toxin-insensitive mechanisms. Thus, in renal epithelial cells, a2-ARs have been shown to increase intracellular calcium and stimulate the production of inositol-1,4,5-triphosphate (IP3) and diacylglycerol (DAG) in a pertussis toxin-insensitive manner (31).

Other signaling pathways are activated by a2-ARs in heterologous expression systems, such as stimulation of phospholipase A2 (PLA2) [41], phospholipase D (PLD) [55], ras (1) and MAP kinases (77), and adenylyl cyclase (24). However, activation of these additional signaling pathways has only been observed to date in heterologous systems where a2-ARs are expressed at relatively high density, compared with current estimates of native target cell expression. It remains to be established whether or not these heterologous systems exaggerate pathways that play no functional role in vivo, or whether they reveal pathways activated in native cells or tissues but which are not detectable with the sensitivity of current technology.

Regulation of a2-Adrenergic Receptor Function

Signal transduction via a2-ARs is regulated by phosphorylation. The a2-AR, like the b-AR, is a substrate for phosphorylation by the G protein-coupled Receptor directed Kinases, or GRKs. One consequence of this phosphorylation is a functional uncoupling of the a2-AR from its cognate G-proteins. GRKs recognize G protein-coupled receptors only in the agonist-bound, or active, state and phosphorylate serine or threonine residues embedded in regions of sequence rich in acidic aspartate or glutamate residues. To date, the GRK 2 and GRK 3 isoforms recognize and hyperphosphorylate both the a2A-AR (40) and the a2B-AR (64) in the third cytoplasmic loop. The a2C-AR is not a substrate for GRK isoforms designated GRK 2, 3, 5, or 6 (40); whether or not the a2C-AR is a substrate for other members of the GRK superfamily has not been established.

The a2-ARs also contain sequence motifs characteristic of substrates for cAMP dependent protein kinase (85); whether or not cAMP-dependent protein kinase-catalyzed phosphorylation influences a2-AR functions has not been revealed to date.

Physiological Effects of a2-Adrenergic Receptors in the Nervous System

A major role of a2-ARs in the CNS is the modulation of the autonomic nervous system. a2-ARs located on presynaptic termini of the ventrolateral medulla mediate a reduction in sympathetic outflow, which leads to a decrease in arterial blood pressure and bradycardia. Concomitantly, a2-ARs on the dorsal vagal nucleus can stimulate vagal (parasympathetic) outflow. One consequence of these CNS effects is the lowering of blood pressure via adrenergic mechanisms and attenuation of cardiac inotropy and chronotropy via cholinergic mechanisms (8). This is the basis for the use of a2-AR agents as antihypertensive agents (see below).

The locus ceruleus (LC) contains noradrenergic axons that terminate in the hippocampus. These fibers are tonically inhibited by norepinephrine release from recurrent axon collaterals. Release of NE onto the hippocampus is controlled by presynaptic a2-ARs (predominantly of the a2A-AR subtype). Excessive stimulation of the release-inhibiting a2-ARs in the LC leads to a dearth of NE in the hippocampus. The catecholamine hypothesis of depression suggests that a lack of NE in the hippocampus may lead to depression. This provides the rationale for the use of a2-AR antagonists, in combination with NE transporter reuptake inhibitors, in treating depression (56).

The LC also innervates the thalamus and cortex. NE released onto these sites stimulates alertness and attention. These cortical and thalamic release sites are inhibited by a2-ARs. This explains the sedative effects of the a2-AR agonists, an undesired side effect in some therapies and the reason for drug choice in others.

a2-AR antagonists also mediate analgesia and are synergistic with opiate effects, but the receptor subpopulation that performs this role is as yet unclear. In fact, since analgesia involves spinal and supraspinal loci, it is probable that more than one a2-AR subtype might contribute to this analgesic response, speculated to account for stress-induced analgesia.

a2-ARs play important roles outside of the nervous system as well, although these roles will not be described here at length. Cardiovascular effects of a2-AR agonists include not only CNS modulation of NE outflow by the a2A-AR but also direct effects on the peripheral vasculature. Transient vasoconstriction is effected by a2-ARs; recent findings with genetically manipulated mice indicate that the a2B-AR subtype may play a principal role in mediating this vasoconstriction (51). Some experimental models also have implicated a2-ARs in release of endothelial-derived relaxation factor (presumably nitric oxide) from endothelial cells lining some regions of the vasculature (14).

a2-ARs regulate excitation-secretion coupling events at loci other than synaptic terminals. Thus, a2-ARs maintain tonic suppression of insulin release from the b cells of the pancreas (70). In addition, a2-ARs in human platelets stimulate the secretion of ADP and production of thromboxane, which rapidly recruits other platelets to a nascent platelet plug, thereby potentiating aggregation.

Therapeutic Use of Agents Directed at a2-Adrenergic Receptors

Clonidine, a partial a2-AR agonist, is commonly used as an antihypertensive agent. Its effects are attributed to the stimulation of central a2-ARs which reduce sympathetic outflow. Although studies performed in the last decade suggest that clonidine can interact with another population of binding sites, termed imidazoline sites, recent data from genetically manipulated mice expressing a mutant a2A-AR provide strong evidence that this subtype may elicit the hypotensive effects not only of native catecholamines but also of a2-AR agonists, even those with an imidazoline structure. One drawback to clonidine therapy is the sedative side effect of the drug; another, if the drug is not removed gradually, is rebound hypertension.

The sedative effect of a2-AR agonists can be exploited to reduce the minimum anesthetic concentration of halothane required to maintain anesthesia. Dexmedetomidine (66) and clonidine (37) have been employed experimentally for this "anesthetic sparing" response.

Clonidine also is used to suppress the symptoms of opiate withdrawal. Since a2-ARs and opioid receptors utilize similar signal transduction mechanisms, a2-AR stimulation can replace opiate receptor stimulation in a non-addictive way. The therapeutic use of a2-AR agonists in managing opiate withdrawal is especially effective when combined with the opiate antagonist naloxone.

Mianserin, an antagonist at a2-ARs (as well as at 5-HT and some other neurotransmitter receptors), is marketed as an antidepressant in Europe and Japan. Mirtazapine1 an a2-AR antagonist and serotonin type 1A receptor agonist, is a promising new antidepressant (18a). Analogs of idazoxan, another a2-AR antagonist, as well as of mianserin are under development to refine the selectivity for a2-ARs and perhaps improve the therapeutic utility of a2-AR antagonists in certain depressive states (56). The anti-depressive actions of these drugs may arise from relieving the a2-AR-mediated inhibition of LC stimulation of the hippocampus (described above).

Agonists at a2-ARs also are being explored clinically to treat spasticity (resulting from stroke, cerebral trauma, or multiple sclerosis), glaucoma, short stature (in children with constitutional growth delay), and diarrhea associated with diabetes. Antagonists at a2-ARs are used clinically to treat Raynaud's phenomenon, psychogenic and organic impotence, non-insulin-dependent diabetes, and obesity (70).

b-ADRENERGIC RECEPTORS

Pharmacological Identification and Cloning of b-Adrenergic Receptor Subtypes

Three b-AR subtypes have been identified in native tissues based on selective activation and blockade of responses by agonists and antagonists. Initially, only two subtypes of b AR, b1-AR and b2-AR, were pharmacologically characterized (11,74). More recently, however, a third subtype, b3-AR, has been identified that is enriched in expression in brown adipose tissue.

The endogenous catecholamines norepinephrine and epinephrine interact differentially with the three b-AR subtypes: epinephrine and norepinephrine are equipotent at b1-AR, epinephrine has a higher affinity than norepinephrine at b2-AR, and b3-AR has a lower affinity for both catecholamines when compared with b1-AR and b2-AR. The data in Table 3 indicate that many synthetic agonists and numerous antagonists differentiate among the b-AR subtypes. Interestingly, most of the potent b1-AR and b2-AR antagonists act as b3-AR agonists (27,75).

The cloning of cDNAs encoding the three b-AR subtypes has confirmed the existence of three distinct subtypes and has allowed for characterization of structure/function relationships for each subtype (4,11,32). Generally, the cloned b-AR subtypes seem to correspond well to the subtypes previously identified pharmacologically. The cDNA for b2-AR was originally cloned from a hamster genomic library using oligonucleotides complementary to peptide sequences of the purified hamster lung b2-AR. The b2-AR gene is localized to chromosome 5 and codes for a protein of 413 amino acids. Ironically, the mammalian b1-AR subtype, a protein of 477 amino acids, was first identified as an "orphan receptor" when cDNA encoding the 5-HT-1A receptor was used to probe a human placental cDNA library (4). The b3-AR gene, encoding a protein of 402 amino acids, originally was isolated from a human genomic library using probes corresponding to the avian b1-AR and human b2-AR cDNA. The b3-AR was cloned later from rat brown adipose tissue cDNA, mouse genomic DNA and rat colon cDNA libraries. The overall sequence homology among the three cloned b-AR subtypes is approximately 50–55%: amino acid identity is much higher in the transmembrane-spanning domains (4,32).

Characterization of the Structure of b-Adrenergic Receptor Subtypes

Several structural features distinguish b3-AR from the other two subtypes. First, the rat b3-AR gene has an intron that interrupts the 3' end of the coding region, whereas the genes encoding the b1-AR and b2-AR subtypes lack introns. Second, the serine/threonine potential phosphorylation sites for the b-AR kinase present in the cytoplasmic tail of b1-AR and b2-AR are absent from all clones of b3-AR. These findings parallel the known lack of desensitization and downregulation of the b3-AR subtype (32).

Signal Transduction via the b-Adrenergic Receptors

All three b-AR subtypes are coupled to adenylyl cyclase activation via a stimulatory Gs protein (4,11,32; c.f. Table 3). However, recent data suggest that the b3-AR subtype may interact not only with Gs but also with Gi in isolated adipocytes (12).

Desensitization of b-Adrenergic Receptors

The multiple molecular mechanisms that can contribute to desensitization of receptors have been extensively studied for the b2-AR subtype (4,32,57). As described earlier in Dopamine Receptors: Clinical Correlates of this volume, continuous exposure of the b2-AR to agonist results in a rapid desensitization (within minutes) of cAMP accumulation. Heterologous (agonist-nonspecific) desensitization results from receptor phosphorylation in the third cytoplasmic loop by cyclic AMP-dependent protein kinase (PKA) and a functional uncoupling of receptor-Gs interactions. In contrast, homologous (agonist-dependent) desensitization is mediated by a receptor-directed kinase (b-adrenergic receptor kinase or bARK) that phosphorylates the agonist-occupied receptor as substrate. Although bARK-catalyzed phosphorylation attenuates the b-AR signal, it does not eliminate it. However, another molecule, dubbed b-arrestin due to its functional analogy to arrestin in the visual system, binds to the phosphorylated receptor and virtually prevents receptor-G protein interaction. b-arrestin also has been implicated in serving as an adapter protein that enriches receptors in clathrin-coated pits and mediates their internalization from the cell surface. The sequestered receptor is dephosphorylated (and hence resensitized) before recycling to the cell surface. Although the quantitative contribution of receptor sequestration to the overall desensitization process is not significant, the sequestration is essential for receptor resensitization.

The three b-AR subtypes vary extensively in their sensitivity to desensitization. Phosphorylation of the b2-AR by bARK and by PKA is readily observed, resulting in receptor uncoupling from Gs and downstream effectors. The b1-AR is overall less sensitive to both short- and long-term desensitization. This is likely a consequence of fewer potential phosphorylation sites on the endofacial surface of b1-AR, compared with b2-AR, and a longer C-terminal tail of the b1-AR that has been implicated in blocking receptor sequestration. The b3-AR, in turn, does not demonstrate any detectable desensitization consensus sequences for phosphorylation by protein kinase A and is deficient in serine/threonine phosphorylation sites for the b-AR kinase. This lack of phosphorylation sites likely explains why functional receptor uncoupling in response to sustained agonist stimulation has not been demonstrated for the b3-AR subtype in native tissues. Stably expressed b3-ARs also do not display short-term agonist-induced desensitization or sequestration (4,32).

Distribution of b-Adrenergic Receptors

In the brain, the b1-AR was detected by quantitative autoradiography in cingulate cortex, layers I and II of the cerebral cortex, the hippocampus, the islands of Calleja and gelatinosus, and the mediodorsal and ventral nuclei of the thalamus, whereas the b2-AR subtype was detected in the molecular layer of the cerebellum, pia mater, in the central, paraventricular and caudal thalamic nuclei, in neostriatum and the cerebral cortex (3,63,67; summarized in Table 3). Both b1-AR and b2-AR are present in equal amounts in substantia nigra, olfactory tubercle, layer IV of the cerebral cortex, medial preoptic nucleus and in all nuclei of the medulla ( 67). Although the b3-AR has been identified in the hypothalamus (32), this subtype is primarily expressed in brown and white adipose tissue (46). Studies also have reported the presence of b3-AR in esophagus, stomach, ileum, gallbladder, colon, skeletal muscle, liver and cardiovascular system (6,26,46). Whether these b3-ARs are on neurons and are involved in innervating these tissues or whether they exist in the end organ is not known. b1-AR and b2-ARs are found in the lungs, including airway smooth muscle, epithelium, cholinergic and sensory nerves, submucosal glands and pulmonary vessels (11,32,71). b1-ARs and b2-ARs are also found in the heart; here the b1-AR are predominantly in the myocytes, and the b2-ARs are on innervating neurons. b2-ARs are also present in saphenous vein, mast cells, macrophages, eosinophils and T lymphocytes.

Physiological Functions of b-Adrenergic Receptor Subtypes

The primary physiological function mediated by b1-AR is an increase in cardiac rate and force of contraction. Renin secretion also is stimulated via b1-AR (11,71), as is relaxation of coronary arteries and gastrointestinal smooth muscle.

b2-ARs stimulate smooth muscle relaxation in the airways, blood vessels and uterus, elicit positive inotropic and chronotropic effects (7), and enhance both glucose production and insulin release (35).

b3-ARs mediate lipolysis in white adipose tissue and thermogenesis in brown adipose tissue. The b3-AR has been implicated in the stimulation of insulin secretion from pancreatic islets, inhibition of glycogen synthesis in skeletal muscle and inhibition of contractility in gastrointestinal smooth muscle (11). Based on ex vivo studies, the b3-AR also modulates certain cardiovascular effects, including relaxation of rat carotid artery, a decrease in blood pressure due to peripheral vasodilation and an increase in heart rate via the baroreflex (6).

Therapeutic Implications for the b-Adrenergic Receptor Subtypes

Most of the therapeutic efficacy of the b-AR agonists and antagonists is directed towards amelioration of peripheral disease, not CNS dysfunction. The diseases treated with b-AR agents include bronchoconstriction associated with asthma, congestive heart failure, arrhythmias and obesity, as discussed briefly below. Although asthma is associated primarily with immunologic mechanisms, b2-AR dysfunction has been implicated in some cases of asthma (20). b2-AR agonists (i.e., salmeterol and formoterol) are used clinically to treat the bronchoconstriction characteristic of asthma, and potentially can be used for other pulmonary disorders.

b-AR agonists, such as dobutamine, are useful in the acute treatment of congestive heart failure. In addition, b3-AR agonists may prove useful for the treatment of obesity (32,71).

Interestingly, several therapeutic applications for certain b-AR antagonists are independent of their b-AR blocker activity. For instance, celiprolol (b1-AR selective antagonist/b1-AR weak agonist) has vasodilator properties via an unknown mechanism; sotalol (nonselective b-AR antagonist) has antiarrhythmic effects; and propranolol, timolol and metoprolol are effective for the prophylaxis, but not treatment, of migraine (38).

SUMMARY

This chapter briefly summarizes knowledge of the structure, function and physiological roles of adrenergic receptors relevant to CNS activities. Hopefully, the references, which cite seminal papers, comprehensive reviews and recent advances will provide the reader with useful cues for further reading.

 

published 2000