Serotonin receptors are prevalent throughout the nervous system and the periphery, and remain one of the most lucrative and promising drug discovery targets for disorders ranging from migraine headaches to neuropsychiatric disorders such as schizophrenia and depression. There are 14 distinct serotonin receptors, of which 13 are G protein coupled receptors (GPCRs), which are targets for approximately 40% of the approved medicines. Recent crystallographic and biochemical evidence has provided a converging understanding of the basic structure and functional mechanics of GPCR activation. Currently, two GPCR crystal structures exist for the serotonin family, the 5-HT1B and 5-HT2B receptor, with the antimigraine and valvulopathic drug ergotamine bound. The first serotonin crystal structures not only provide the first evidence of serotonin receptor topography but also provide mechanistic explanations into functional selectivity or biased agonism. This review will detail the findings of these crystal structures from a molecular and mutagenesis perspective for driving rational drug design for novel therapeutics incorporating biased signaling.
Serotonin or 5-hydroxytryptamine (5-HT) remains one of the most widely studied chemical messengers. Serotonin produces a myriad of physiological effects in humans, mediated through 14 distinct receptor subtypes, of which 13 are G protein coupled receptors (GPCRs), and one ligand-gated cation channel (Berger et al., 2009; Hoyer et al., 1994). 5-HT receptors have evolved over the course of 700–800 million years (Kroeze and Roth, 1998; Peroutka and Howell, 1994). In the human central nervous system (CNS) alone, all the serotonin receptor subtypes, with the exception of 5-HT5b, are expressed, and they are involved in the modulation of sleep-wake cycles, emesis, appetite, mood, memory, breathing (Ray et al., 2011), cognition, and many other functions (Berger et al., 2009; Meltzer and Roth, 2013). Much of the serotonin in the body, however, is not found in the CNS, rather in the gastrointestinal (GI) tract, where it causes peristalsis through either smooth muscle contraction or enteric nerve depolarization (Gershon et al., 1990). Lesser known, but historically important, serotonin is found in the blood, isolated from platelets in the serum, where it is involved in blood coagulation and vasoconstriction, a function which led to its name “sero” (from serum) and “tonin” (to induce contraction)(Rapport et al., 1948). Not surprisingly, more than 125,000 published articles as of 2014 (pubmed search) have been published on serotonin and its receptors.
Much research regarding serotonin has been in the area of neuropsychiatric drug discovery in treatment of affective disorders, where there continues to be an extreme interest in the design of more efficacious pharmaceuticals. Drugs either targeting serotonin receptors or serotonin itself represent a large share of the top selling pharmaceuticals in the past decade, with more being approved for future use. This is most evident with antidepressants such as SSRIs (serotonin-selective reuptake inhibitors) dominating the drug market, with over $11 billion in sales in 2008 alone, and Cymbalta (duloexetine), a dual serotonin-norepinephrine reuptake inhibitor (SNRI), being in the top 10 drugs sold in the US in 2012 (Lindsley, 2013). Additionally, newer antipsychotics, such as aripiprazole (Abilify) and quetiapine (Seroquel), which have partial agonist activity and/or antagonist activity at serotonin receptors, were also in the top 10 drugs sold in the US. Recently, newer SNRIs are being approved by the FDA with an indication to treat depression, including the approved levomilnacipran (Fetzima), and SSRIs with direct agonist activity at the 5-HT1A receptor have been approved by the FDA including vilazodone (Viibryd) and vortioxetine (Brintellix) for treatment of major depressive disorder (Celada et al., 2013). There is, however, a need for the discovery of more efficacious CNS pharmaceuticals with reduced side effects associated with off-target activity. For example, agents for Parkinson’s disease, such as pergolide and bromocriptine, were discontinued in the US because of their ability to induce cardiac valve hypertrophy, which has been linked to off-target agonist action at the 5-HT2B receptor (Horvath et al., 2004; Roth, 2007; Rothman et al., 2000; Setola et al., 2003).
In addition to drug selectivity, the concept of functional selectivity (Urban et al., 2007) in pharmacology has revolutionized the drug discovery process with recent findings highlighting that functionally selective β-arrestin biased drugs can show efficacy in preclinical models of schizophrenia (Allen et al., 2011), and β-arrestin biased angiotensin II type 1 compounds are currently in Phase 2 clinical trials for treatment of congestive heart failure (Soergel et al., 2013). Thus, rational drug design of newer serotonin receptor drugs incorporating ligand bias appears to be a fruitful area of drug development that potentially could be facilitated by insights into the structure and function of their respective targets.
Recently, there has been a GPCR structural “renaissance” with crystal structures from several representative receptor classes being solved, including rhodopsin (Palczewski et al. 2000), β2 adrenergic (Cherezov et al., 2007), β1 adrenergic (Warne et al., 2008), A2A adenosine (Jaakola et al., 2008), H1 histamine (Shimamura et al., 2011), D3 dopamine (Chien et al., 2010), smoothened (Wang et al., 2013b), CXCR4 (Wu et al., 2010), sphingosine 1-phosphate (Hanson et al., 2012), protease-activated receptor 1(Zhang et al., 2012), M2 muscarinic (Haga et al., 2012), M3 muscarinic (Kruse et al., 2012), mu opiate (Manglik et al., 2012), delta opiate (Granier et al., 2012), kappa opiate (Wu et al., 2012), nociceptin/orphanin (Thompson et al., 2012), and recently the 5-HT1B and 5-HT2B serotonin receptors (Wacker et al., 2013; Wang et al., 2013a). These receptor crystal structures provide a wealth of structural information for the design of more receptor subtype selective agents, and also provide insights into the chemo-mechanical processes involved in GPCR activatation (Mustafi and Palczewski, 2009). Thus GPCR structures likely will serve medicinal chemists and pharmacologists to design new functionally selective drugs as potential therapies or pharmacological tools (Kenakin and Christopoulos, 2013).
The recent publications of two serotonin receptor crystal structures in complex with ergotamine (ERG), a former antimigraine agent, has opened up two major discussions relevant to the serotonin field, which concern the molecular basis for 1) serotonin receptor subtype recognition and 2) functional selectivity. The 5-HT1B crystal structure illustrates the molecular basis for the selectivity of antimigraine triptan drugs that are devoid of 5-HT2B receptor off-target activity, and the 5-HT2B crystal structure provides clues into the functional mechanics of a receptor in a β-arrestin biased state. These crystal structures not only further our understanding into serotonin’s therapeutic and off-target effects in the current drug discovery process, but also serve to refine our understanding of GPCRs in general.
This review will be divided into two major sections. The first part will provide a background and relevance of serotonin as it relates to our current understanding of its actions in the body and in medicine, with a particular emphasis on the 5-HT1B and 5-HT2B receptors. The second part will focus on the structural importance of the recently published serotonin receptor crystal structures, especially as it relates to our current understanding of GPCR structure and function. The final section will suggest future areas for the development of our understanding of functional selectivity and biased agonism.
Serotonin was actually discovered independently by two laboratories, Vittorio Erspamer’s lab in Rome, Italy (Erspamer and Boretti, 1950) and Irvine Page’s lab at the Cleveland Clinic (Rapport et al., 1948). Although the action of serotonin is usually associated with CNS function, the majority of serotonin actually resides in the gastrointestinal (GI) tract, produced by enterochromaffin cells lining the lumen of the gut, where it causes increased peristaltic activity (Beleslin and Varagic, 1958). It was from the GI tract that Vittorio Erspamer first isolated serotonin in 1938, a substance that he deduced was an indole amine, which he named enteramine (Erspamer and Asero, 1952). Years later in 1948, a substance in blood serum that produced vasoconstriction was isolated by Maurice Rapport, Arda Green, and Irvine Page of the Cleveland Clinic (Rapport et al., 1948; Rapport et al., 1947), which was later named serotonin (“sero” serum, “tonin” constrict). It wasn’t until 1953, however, when Betty Twarog in Irvine Page’s lab analyzed brain extracts using a serotonin-sensitive mollusk bioassay (Twarog and Page, 1953), an assay initially discovered by Erspamer (Erspamer and Ghiretti, 1951), that the presence of serotonin in brain extract was recognized. Interestingly, Erspamer had initially posited an indole nucleus being contained in enteramine’s structure, and later Rapport determined the structure of serotonin to be 5-hydroxytryptamine (Rapport, 1949) shown in Figure 1.
Shown are the structures of 5-hydrotryptamine (5-HT or serotonin) and ergolines, lysergic acid diethylamide (LSD), ergotamine (ERG), and dihydroergotamine (DHE), which are ligands at all serotonin receptors. Also shown is an example triptan ligand, sumatriptan, which posseses 5-HT1B receptor selectivity and also norfenfluramine, the active metabolite of fenfluramine, which possesses 5-HT2B agonist activity.
Prior to the discovery of serotonin in the 1930’s, Sandoz pharmaceuticals had been marketing ergotamine, a drug that produced uterine contractions and was used to induce labor (Hofmann, 1978). Ergotamine is part of a larger class of compounds called ergolines, which includes ergopeptines like ergotamine, lysergamides like lysergic acid diethylamide (LSD), and ergoclavines, all of which retain the indole nucleus in their structures (Figure 1). By 1938, research into other ergolines at Sandoz by the Swiss chemist, Albert Hofmann, who under Arthur Stoll, synthesized several hundred analogs of ergotamine. Albert Hofmann is often famously noted for the discovery of LSD, which serendipitously had been discovered by him via synthesizing analogs of ergotamine and changing the amide substituent into simpler alkyl chains, a subclass of ergolines usually termed lysergamides. LSD, however, did not satisfy the pharmacological profile using rat uterus contraction assays, and interest in it as an obstetric drug waned. Hofmann, however, was compelled to re-synthesize LSD in 1943 based on the diethyl substitution found in a known analeptic drug coramine (nicotinic acid diethylamide, forming the D-ring of LSD, Figure 1), which was known to induce CNS stimulation and increased respiration. On April 16th 1943, Hofmann experienced what he thought was a “mystical” experience around his newly synthesized LSD. Thinking he may have accidently ingested a small amount, three days later on April 19th, he ingested what he thought was a small test dose, 0.25 mg or 250 μg, which today is known to be a substantial dose of LSD, and the unexpected effects were far beyond any respiratory or CNS stimulant as seen with coramine. Instead, they were “kaleidoscopic, fantastic images” and “alterations that I perceived in myself, in my inner being” (Hofmann, 1980). From that time forward, the unique and powerful psychoactive properties that LSD can evoke became widely known.
Later in the 1940s, LSD was promoted as a research tool in the psychiatry, especially in understanding schizophrenia. Until then, the roots of schizophrenia’s etiology had been argued to be more nurture-based rather than nature-based, and attributed to improper parenting. LSD’s ability to induce transient thought disorders, perceptual and visual hallucinations, and delusions in certain individuals, however, began to spur more research into LSD’s pharmacological actions. After the structure of serotonin was published, a hypothesis was put forth that linked serotonin to several mental disorders (Woolley and Shaw, 1954). Woolley, who was almost completely blind from diabetes, conceived the importance of serotonin in mental disorders simply by realizing that the tryptamine scaffold of 5-HT was embedded within LSD’s structure (Figure 1). It wasn’t until much later, however, that the importance of serotonin receptors in schizophrenia and depression was fully appreciated.
LSD became a powerful tool at the disposal of pharmacologists to clarify the physiological actions of serotonin in whole tissue assays. John Gaddum pioneered the serotonin receptor pharmacology field by first discovering that LSD antagonized the action of 5-HT in these tissues (Gaddum, 1953), which ultimately led to the classification of two types of serotonin receptors, then referred to as tryptamine receptors (Gaddum and Picarelli, 1957). In serotonin-induced contractions of guinea pig ileum, the D-type was designated as being sensitive to blockade by the irreversible antagonist, dibenzyline, also known as phenoxybenzamine, and the M-type was sensitive to blockade by morphine (Gaddum and Picarelli, 1957). In that study, Gaddum and Picarelli hypothesized that serotonin-induced contractions of the ileum by the D-type receptors acted in smooth muscle and the M-type acted in nerve ganglia. With only a delineation of serotonin receptors based on available pharmacological agents in animal tissues, much of serotonin receptor research waned for more than two decades.
In the 1970’s, the development of radioligands specifically to label receptors in animal tissue were developed, leading to the use of [3H]-LSD and [3H]-5-HT. Peroutka and Snyder (1979) used 3H-LSD to label two distinct receptor populations, one that had higher affinity for [3H]-5-HT, which they named 5-HT1, and another that had higher affinity for [3H]-spiperone, which they named 5-HT2. The affinity for 5-HT at these receptor types was sensitive to the presence of guanine nucleotides (Peroutka et al., 1979), indicating that these receptors were likely G protein coupled. Later, it was found that 5-HT2 receptors were correlated with the actions of the previously named D-type receptor (Bradley et al., 1986). The M-type was then found to be a new serotonin receptor type, the 5-HT3 receptor, which was revealed to be a ligand-gated ion channel permeable to cations (Derkach et al., 1989), in contrast to the 5-HT1 and 5-HT2 receptors.
On the basis of selective agonist/antagonist ligand affinities, cloned sequence homology, and intracellular transduction mechanisms, serotonin receptors were reclassified into seven types, comprised of 14 subtypes (Hoyer et al. 1994). With this new classification system, much serotonin receptor research has flourished in the past 20 years, linking receptor distribution to physiological actions (Berger et al., 2009), and the design of subtype selective agents, which will be discussed in the next section.
According to the current classification system, there are seven types of 5-HT receptors, 5-HT1-7. All 5-HT receptors are GPCRs with the exception of 5-HT3, which is a ligand-gated cation channel. Serotonin GPCRs couple to all three canonical signaling pathways through Gαi, Gαq/11, and Gαs, allowing this receptor family to modulate several biochemical signaling pathways, and leading to several physiological consequences, a discussion which is beyond the scope of this review, (but see Nichols and Nichols (2008) for a more comprehensive review).
The Gαi-coupled serotonin receptors encompass the 5-HT1 and 5-HT5 types. The 5-HT1 subfamily, which is comprised of 5-HT1A, 5-HT1B, 5-HT1D, 5-HT1E, and 5-HT1F subtypes, usually couple to Gαi(2) leading to inhibition of adenylyl cyclase and a decrease of intracellular cAMP (Bockaert et al., 1987; Lin et al., 2002). There is no 5-HT1C receptor because this subtype was re-designated the 5-HT2C receptor based on its functional Gq coupling. In this subfamily, the 5-HT1A receptor is expressed as both a pre-synaptic receptor in the raphe nuclei, where it serves to inhibit further release of CNS serotonin, and as a post-synaptic receptor in medial prefrontal cortex and other cortical areas to modulate dopamine release (Altieri et al., 2013; Bockaert et al., 1987). The 5-HT1A receptor has been implicated in anxiety, and as mentioned, novel anti-depressants and antipsychotics have begun to incorporate partial 5-HT1A agonist activity to enhance their efficacy in such diseases (Celada et al., 2013). Historically, the other 5-HT1 subtypes have been a focus on for the treatment of migraine and will be discussed in more detail in the following sections.
Little is known about the other Gαi-coupled serotonin receptor family, the 5-HT5 subfamily, which includes subtypes 5-HT5A and 5-HT5B. Only 5-HT5A has only been identified in human cortical layers and cerebellum (Pasqualetti et al., 1998). 5-HT5A has been shown to couple to Gαi in HEK cells (Hurley et al., 1998), and the selective 5-HT5A antagonist, SB-6995516, impairs memory (Gonzalez et al., 2013) and reduces acoustic startle (Curtin et al., 2013) in animal models. It has been suggested that the 5-HT5 receptor may be a therapeutic target in psychiatric disorders (Thomas, 2006), but to date, no selective agonists for this subtype exist to clarify its role.
The Gαq/11-coupled serotonin receptors include the 5-HT2 receptor subtypes, 5-HT2A, 5-HT2B, and 5-HT2C. These receptors traditionally have been linked to Gq-related pathways, with activation of phospholipase C producing inositol triphosphate (IP3) and diacylglycerol (DAG), ultimately leading to an increase in intracellular calcium (Roth et al., 1984; Roth et al., 1998). There is evidence, however, that 5-HT2 receptors may exhibit G protein promiscuity, at least in cultured cell systems. For instance, inhibition of cAMP possibly through Gαi activation has been found to occur for 5-HT2A (Garnovskaya et al., 1995), and in the case of 5-HT2C with high expression levels (Lucaites et al., 1996). The 5-HT2A receptor is, by far, one of the most studied types of serotonin receptors owing to its widespread expression in the cortex (Willins et al., 1997), its involvement in the mechanism of hallucinogen action (see Nichols, 2004 for a review), and its implications and effectiveness as a receptor target in a handful of mental illnesses including schizophrenia, depression, and Tourette’s syndrome to name a few (Meltzer and Roth, 2013). The 5-HT2C is also widely distributed in the CNS (Mengod et al., 1996; Molineaux et al., 1989), and is also found in the ventral tegmental area negatively regulating dopamine release into nucleus accumbens (Di Giovanni et al., 2000). In fact, 5-HT2C-induced control of dopamine release has been proposed as a mechanism for some antidepressants (Millan et al., 2005), and 5-HT2C antagonism may be responsible for the rewarding effects associated with these antidepressants (McCorvy et al., 2011). In addition, selective 5-HT2C agonists such as lorcaserin have shown efficacy as appetite suppressants (Thomsen et al., 2008) and are currently FDA approved (Hess and Cross, 2013), albeit with slight off-target effects at the 5-HT2B receptor. Less well-known is the ability for 5-HT2 receptors to induce cell proliferation (Banasr et al., 2004), especially during tissue development, which will be further discussed as it pertains to off-target effects stemming from chronic 5-HT2B receptor activation.
The Gαs-coupled receptors include 5-HT4, 5-HT6, and 5-HT7. The 5-HT4 receptor as been shown to couple positively to adenylyl cyclase leading to increased levels of cAMP in guinea pig hippocampus (Bockaert et al., 1990), but may also result in increased calcium current, at least in human atrial myocytes (Ouadid et al., 1992) through a possible Gαq/11-coupled mechanism, which has been demonstrated in reconstituted systems (Ponimaskin et al., 2002). 5-HT4 receptor distribution varies widely in humans (Bonaventure et al., 2000), and select 5-HT4 activation can range from peristalsis in the gut (Kadowaki et al., 2002), to long-term potentiation (LTP) in hippocampus (Kulla and Manahan-Vaughan, 2002). Selective 5-HT4 knock-out mice exhibit stress-induced feeding disorders (Compan et al., 2004). Selective 5-HT4 antagonists have been proposed to treat atrial fibrillation, irritable bowel syndrome, and urinary incontinence (Brudeli et al., 2013).
The 5-HT6 and 5-HT7 receptors, also Gαs-coupled, lead to an increase in cAMP, and are located in the CNS in the thalamus, hypothalamus, hippocampus, as well as the peripheral tissues. Many typical and atypical antipsychotics function as antagonists at cloned human 5-HT6 and 5-HT7 receptors (Roth et al., 1994). There is evidence that selective 5-HT6 receptor antagonists, such as SB 271046, may also have cognitive enhancing properties, and may potentially be used for “add-on” therapy in conjunction with antipsychotics with low affinity for the 5-HT6 receptor (Roth et al., 2004). In addition, the 5-HT7 receptor is a potential drug target for depression and 5-HT7 receptor antagonism has been posited as necessary for amisulpride’s antidepressant (Abbas et al., 2009).
In addition to canonical G protein signaling, GPCRs have also been found to be involved in non-canonical signaling via arrestin recruitment. Arrestin recruitment is a now appreciated as a mechanism in GPCR desensitization (Freedman and Lefkowitz, 1996) with clinical importance, especially regarding the tolerance and therapeutic utility of drugs (Violin et al., 2014). GPCR desensitization occurs through GPCR kinase (GRK) recruitment and subsequent GPCR phosphorylation, which recruit β-arrestin1 and β-arrestin2 (also known as arrestin 2 and arrestin3, respectively) from which the binding of β-arrestin sterically prevents further G protein coupling, ultimately leading to desensitization (Krupnick and Benovic, 1998). The GPCR-β-arrestin complex can also target the GPCR to clathrin-coated pits (Krupnick et al., 1997), ultimately leading to internalization (Ferguson et al., 1996). GPCR internalization serves to control both the duration of the response and subsequent downstream signaling of the receptor-ligand complex. In addition to β-arrestins acting to attenuate G protein-dependent signaling, they can simultaneously initiate parallel, G protein-independent signals (Lefkowitz and Shenoy, 2005). In view of the recent findings with non-canonical arrestin-dependent signaling, it was previously thought that all agonist-receptor complexes stimulated all available signaling pathways to an equal extent. However, over the last decade, mounting evidence questions such a simplistic view.
The introduction and use of recombinant cell systems to express and measure specific receptor functional responses ushered in a new concept in pharmacology. Functional selectivity, sometimes called “agonist-directed trafficking” or more recently “ligand bias” was proposed to account for differences between two or more functional responses resulting from receptor activation by an agonist (Urban et al., 2007). The earliest conceptual framework for functional selectivity was proposed by Roth and Chuang (1987) to account for the phenomenon whereby one or more distinct biochemical transduction mechanisms may be activated by a single GPCR. Roth and Chuang (1987) also predicted that this phenomenon would lead to the development of selective agonists for a particular receptor-linked effector. Although functional selectivity exhibited by a given ligand may be system-dependent in select cell lines, it has been well-established that functional selectivity reflects a ligand’s ability to select different conformations of the receptor, leading to a specific functional response over another (Kenakin, 2011), thus exhibiting conformational ligand bias. Selecting different conformations of the receptor may be key to the design of novel functionally selective agonists or antagonists with improved therapeutic potential over currently available medications; however our understanding of the molecular determinants that lead to biased agonism is not well-understood.
The first serotonin receptor to be cloned was the 5-HT1A receptor subtype (Kobilka et al., 1987; Fargin et al., 1988). Interestingly, the 5-HT1 subtype exhibits a high degree of sequence and pharmacological similarity to the β-adrenergic receptors, and initially, it was thought to be a β-adrenergic receptor. Later, the 5-HT1B subtype was designated as a distinct receptor subtype from 5-HT1A on the basis that it did have high affinity for β-adrenergic receptor antagonists, but also had low affinity for spiperone (Pedigo et al., 1981; Nelson et al., 1981), whereas the 5-HT1A receptor binds spiperone with high affinity. Confusion still arose, however, over the existence of a human ortholog of 5-HT1B, because it exhibited responses in rodents (Hoyer et al., 1992) that were not similar in humans (Peroutka, 1994). Therefore, the 5-HT1Dα and 5-HT1Dβ receptor subtypes were classified to account for the human orthologs (Hoyer and Middlemiss, 1989). These two receptor subtypes were often confused as being pharmacologically distinct based on their ability to be activated by β-adrenergic antagonists, such as propanolol and pindolol, where a single amino acid residue distinguishes this reverse pharmacology (Hamblin et al., 1992; Metcalf et al., 1992). This species difference will be discussed more in detail as it pertains to ligand selectivity in later sections.
As mentioned, the 5-HT1B/1D receptors have been shown to couple negatively to adenylyl cyclase resulting in lower intracellular cAMP, which has been demonstrated in vivo in substantia nigra (Bouhelal et al., 1988; Schoeffter et al., 1988), rabbit mesenteric artery (Hinton et al., 1999), and in transfected heterologous cell lines (Levy et al., 1992; Adham et al., 1992; Weinshank et al., 1992). Expression of 5-HT1B receptors is most prevalent in human cerebral arteries (Nilsson et al., 1999), where vasospasm has been implicated in the pathogenesis of migraine. In fact, activation of 5-HT1B/1D receptors in cerebral arteries by agonists causes vasoconstriction and has been posited as a therapeutic explanation for the antimigraine drugs.
Ergotamine, as discussed previously, was developed by Sandoz in the early 1930’s as a drug to induce pronounced uterine contraction, but ergotamine also shows vasoconstriction in the periphery as well as the cerebral arteries, suggesting therapeutic efficacy for migraine. The ergolines as a class of drugs, in general, are recognized as having several receptor targets including but not limited to 5-HT1A-1F, 5-HT2A-C, D1-5, as well as α1 and α2 adrenergic receptor types (Silberstein, 1997). The antimigraine effects, however, are likely mediated through agonist activity at the 5-HT1B, 5-HT1D and possibly 5-HT1F receptors present on trigeminal nerve terminals (Ramirez Rosas et al., 2013). However, ergotamine as well as other antimigraine ergolines, such as dihydroergotamine and methysergide, suffer from acute side-effects due to vasoconstriction in the periphery leading to hypertension and coronary vasoconstriction. Ultimately with continued use, severe chronic side-effects can occur such as retroperitoneal fibrosis, pleuropulmonary fibrosis, and cardiac valvulopathy due to off-target activity at the 5-HT2B receptor (Roth, 2007; Rothman et al., 2000), which will be discussed in the next section.
In light of the serious side-effects and off-target activity of the ergolines, newer generation antimigraines incorporate an extended substitution on the tryptamine scaffold giving rise to the triptan class of drugs (Humphrey, 2008) (Figure 1). Triptans, such as sumatriptan, are extremely effective as antimigraine drugs potentially alleviating symptoms by vasoconstriction at cerebral arteries via 5-HT1B receptors and inhibition of proinflammatory neuropeptides in trigeminal fibers via 5-HT1D receptors. This class of medications avoids the cardiac valvulopathy and coronary vasoconstriction by having lower affinity at 5-HT2 receptors, yet retaining selectivity and agonist activity at 5-HT1B/1D/1F receptors.
The rat 5-HT2B receptor, formerly known as the 5-HT2F receptor, was first cloned and characterized from the rat stomach fundus (Kursar et al., 1992; Foguet et al., 1992), which was previously discovered to cause contraction upon application of several trypamine agonists (Vane, 1959). The human 5-HT2B receptor was cloned from a placental genomic library (Kursar et al., 1994; Schmuck et al., 1994), and shares considerable homology with the previously cloned human 5-HT2A and 5-HT2C (formerly 5-HT1C) receptors. Functional activation of 5-HT2B was linked to inositol triphosphate (IP3) production in cultured cells (Wainscott et al., 1993) and calcium influx (Cox and Cohen, 1996), which leads to contraction of the stomach fundus. The ex vivo stomach fundus contraction assay was commonly used for detection of early serotoninergic agents including hallucinogenic drugs like LSD, but this assay likely reflects 5-HT2B receptor activity rather than 5-HT2A activity, which is more predictive for hallucinogenic agents (Nelson et al., 1999).
Expression patterns of 5-HT2B receptor using mRNA detection in mouse (Kursar et al., 1994) showed high levels in the liver and kidney, with lower amounts in the heart and brain (Loric et al., 1992). Other mRNA expression studies confirmed the presence of 5-HT2B receptors in brain, specifically in the septal nuclei, dorsal hypothalamus and medial amygdala, with comparable levels detected in the stomach fundus (Duxon et al., 1997). Later, it was shown that the cardiovascular (Choi et al., 1994) and cardiac physiological functions of 5-HT2B were more prominent compared to its role in the CNS compared to 5-HT2A and 5-HT2C receptor CNS functions (Bonhaus et al., 1995).
The hypothesis that 5-HT2B activation leading to cardiac hypertrophy gained prominence with the observation that the 5-HT2B knockout mouse demonstrated a lethal phenotype as a result of hypotrophy of the trabeculae cardiac muscles (Nebigil et al., 2000), which help control backflow of blood through the mitral (bicuspid) and tricuspid valves. Studies linked the involvement of the 5-HT2B receptor in cardiac hypertrophy through Gq activation and concomitant phosphatidylinositol-3 kinase/Akt and extracellular signal-regulated kinase (ERK) 1/2 signaling pathways (Nebigil et al., 2003).
Coincidentally, the fenfluramine and phenteramine weight loss combination (fen-phen) marketed by Wyeth in the early 1990s resulted in patients exhibiting abnormal echocardiograms (Connolly et al., 1997), pulmonary hypertension (Brenot et al., 1993), among other cardiac valve deficiencies (Smith et al., 2009). Indeed, the major metabolite of fenfluramine, norfenfluramine (Figure 1), was found to have potent 5-HT2B agonist activity in Gq calcium release assays, whereas fenfluramine exhibits little activation (Fitzgerald et al., 2000; Roth, 2007; Rothman et al., 2000), indicating that the off-target effects of the fenfluramine metabolite, norfenfluramine, can lead to serious side-effects. This fact coupled with the finding that ergotamine also exhibits 5-HT2B agonist activity clearly indicates that the 5-HT2B receptor is an important off-target receptor to avoid in drug design. Finally, recent evidence regarding ergotamine’s pharmacological actions may be explained by its differential signaling profile at the 5-HT2B receptor (Huang et al., 2009), for which the major biochemical signaling pathway leading to cardiac valvulopathy remains unknown.