Structure, Function, and Pharmaceutical Ligands of 5-Hydroxytryptamine 2B Receptor

Since the first characterization of the 5-hydroxytryptamine 2B receptor (5-HT2BR) in 1992, significant progress has been made in 5-HT2BR research. Herein, we summarize the biological function, structure, and small-molecule pharmaceutical ligands of the 5-HT2BR. Emerging evidence has suggested that the 5-HT2BR is implicated in the regulation of the cardiovascular system, fibrosis disorders, cancer, the gastrointestinal (GI) tract, and the nervous system. Eight crystal complex structures of the 5-HT2BR bound with different ligands provided great insights into ligand recognition, activation mechanism, and biased signaling. Numerous 5-HT2BR antagonists have been discovered and developed, and several of them have advanced to clinical trials. It is expected that the novel 5-HT2BR antagonists with high potency and selectivity will lead to the development of first-in-class drugs in various therapeutic areas.

The 5-HT 2 receptor (5-HT 2 R) subfamily is subdivided into 5-HT 2A , 5-HT 2B and 5-HT 2C receptors. The 5-HT 2B R was the last identified 5-HT 2 R family member and was first cloned in rat stomach fundus in 1992 [4], before the cloning of human 5-HT 2B R in several tissues two years later [5,6]. In humans, the 5-HT 2B R shares nearly 50% homology with the 5-HT 2A R and 5-HT 2C R, with about 70% homology in the transmembrane region [5]. Expressions of human 5-HT 2B R mRNA have been detected in many different tissues, including the liver, kidney, intestine, pancreas, stomach, heart, lung, brain, uterus, trachea, testis, prostate, and placenta [5,6]. The 5-HT 2B R is a G q/11 protein-coupled receptor. The activation of G q/11 results in several parallel signaling pathways. One branch of the canonical G q/11 signal transduction pathway is involved in the hydrolysis of guanosine triphosphate (GTP) to guanosine diphosphate (GDP) and is mediated by the G q/11 protein.
The GTP-bound G q/11 stimulates the effector protein phospholipase Cβ (PLCβ) and leads SSc is a chronic autoimmune disease characterized by progressive vascular disease and fibrosis of the skin and internal organs. Emerging evidence has suggested that 5-HT 2B R plays an important role in SSc. In 2011, Dees et al. found that the expression of the 5-HT 2B R was significantly increased in the skin of SSc patients compared with the normal skin of healthy individuals [39]. In vitro studies on SSc dermal fibroblasts suggested that the profibrotic effects of 5-HT are mediated by the 5-HT 2B R, excluding the 5-HT 1B R and 5-HT 2A R, which are also expressed in dermal fibroblasts [39]. In vivo studies on BLM-induced dermal fibrosis and tight-skin-1 (tsk-1) mouse models showed that 5-HT 2B R antagonists ameliorated fibrosis. In addition, mice lacking the 5-HT 2B gene could be protected from BLM-induced fibrosis [39]. In 2018, Chaturvedi et al. studied human adult dermal fibroblasts (HADF) isolated from SSc patients and showed that stimulation of 5-HT/TGF-β1 in HADF significantly increased the expression of profibrotic genes. Profibrotic genes were downregulated by the 5-HT 2B R antagonist SB-204741, whose antifibrotic effect might be involved in the suppression of TGF-β1-mediated non-canonical (non-Smad dependent) signaling pathways [40]. Moreover, Wenglén et al. discovered selective antagonists of the 5-HT 2B R (AM1125 and AM1476) and suggested their antifibrotic effects for the potential treatment of SSc (see Section 4.2.1 for details) [42,43]. [49], and strong evidence has suggested that the 5-HT 2B R plays a role in hepatocellular carcinoma (HCC), neuroendocrine tumor (NET) and pancreatic tumor. Through the analysis of liver tissues from patients with HCC, Sarrouilhe et al. found that the 5-HT 1B R and the 5-HT 2B R were overexpressed in tumor tissues and that their antagonists inhibited proliferation of HCC cell lines, such as Huh7 and HepG2 [50]. Liang et al. suggested that 5-HT promoted the proliferation of serum-deprived Huh7 cells by upregulating the transcription factor FOXO3a, although this pro-proliferative effect was not observed in serum-deprived HepG2 or Hep3B cells [51]. They further found that the proproliferative effect of 5-HT could be blocked by the 5-HT 2B R antagonist SB-204741 in Huh7 cells and that 5-HT 2B R mRNA was significantly higher expressed in Huh7 cells compared to HepG2 and Hep3B cells, which may contribute to the distinct 5-HT effects in different serum-deprived HCC cells [51]. Using zebrafish HCC models, Yang et al. suggested that the 5-HT 2B R was involved in HCC carcinogenesis [52,53]. In zebrafish, the expression of the 5-HT 2B R was found to be high in HSCs, much lower in hepatocytes, and practically absent in neutrophils and macrophages [53]. The activation of the 5-HT 2B R could increase both the proliferation and the activation of HSCs, as well as the expression of TGF-β1, resulting in liver enlargement and accelerating HCC carcinogenesis. In contrast, blocking the 5-HT 2B R led to opposite effects [52,53].

Neuroendocrine Tumor (NET)
NET is a rare type of tumor that most commonly arises in the GI tract and can lead to carcinoid syndrome [54,55]. Svejda et al. studied KRJ-I cells, a small intestinal-NET (SI-NET) cell line, and found that treatment with 5-HT 2B R antagonist PRX-08066 inhibited the 5-HT secretion and KRJ-I cell proliferation, simultaneously decreasing the phosphorylation of ERK1/2 and the transcript levels and secretion of profibrotic growth factors, including TGF-β1, connective tissue growth factor (CTGF) and fibroblast growth factor (FGF2). The antiproliferative and antifibrotic effects of the 5-HT 2B R suggested that this is a promising target for intervening SI-NETs [56].

Pancreatic Tumor
In 2017, Jiang et al. reported that the 5-HT 2B R could be used as a potential therapeutic target for intervening pancreatic ductal adenocarcinomas (PDACs) [57]. 5-HT was found to be increased in human PDAC tissues. Moreover, the incubation of 5-HT with PDAC cell lines resulted in an increase in PDAC cell proliferation and a decrease of PDAC cell apoptosis. Both in vitro and in vivo studies demonstrated that the pro-survival effect of 5-HT is mediated by the 5-HT 2B R, but not other 5-HTRs. The 5-HT 2B R agonist α-Me-HTP promoted the survival of PDAC cells, whereas the 5-HT 2B R antagonist SB-204741 or genetic silencing of the 5-HT 2B R blocked the pro-survival effect of 5-HT in PDAC cells and significantly reduced the tumor burden of PDAC in mice [57]. Moreover, the tumorsuppressive effects of 5-HT 2B R antagonism were further confirmed in transgenic mice with pancreatic tumors. Notably, the mechanism behind 5-HT mediated PDAC cell survival involved the activation of PI3K/Akt/mTOR signaling and the enhancement of aerobic glycolysis (Warburg effect) [57].

Gastrointestinal (GI) Tract
Previous studies have suggested a role for the 5-HT 2B R in the GI system. 5-HT 2B R mRNA was widely expressed throughout the human GI tract [58]. The high expression of 5-HT 2B R was detected in colonic smooth muscle, and the excitatory effects of 5-HT in the human colon were demonstrated to be mediated by the 5-HT 2B R [58]. The 5-HT 2B R was also found in the interstitial cells of Cajal (ICC), the "pacemaker cells" of the GI tract, which are expressed throughout the entire GI tract and required for normal GI motility. The activation of the 5-HT 2B R in mouse models increased the proliferation of ICC in vitro and in vivo [59,60]. The 5-HT 2B R triggered ICC proliferation was found to be mediated by PLC, intracellular calcium release and PKCγ [61].
Irritable bowel syndrome (IBS) is a common functional GI disorder that is characterized by abdominal discomfort and abnormal defecation. Visceral hypersensitivity is considered a hallmark characteristic of IBS. Many animal studies have demonstrated that the 5-HT 2B R antagonism could help to modulate visceral hypersensitivity, colonic motility, and defecation [62][63][64][65], which indicates that the 5-HT 2B R is a potential therapeutic target for GI disorders, especially for IBS. Notably, a study in conscious dogs showed that 5-HT 2B R antagonism had no contractile effect on normal colonic motor activity and suggested that 5-HT 2B R antagonists may be utilized for the treatment of diarrhea-predominant IBS without resulting in a constipation side effect [66].

Nervous System
As a neurotransmitter, 5-HT plays an essential role in the nervous system [67,68]. The 5-HT 2B R has been suggested to mediate 5-HT functions in cognitive processes such as learning and memory [69][70][71], motor activities like breathing [72][73][74], as well as pain disorders, neuroglia function, and the dopaminergic pathway.

Regulation of Pain Disorders
The 5-HT 2B R has been implicated in migraine and neuropathic pain, which are two common forms of pain disorders in humans [75][76][77]. Migraine is a common primary headache disorder characterized by moderate to severe recurrent headaches. In 1989, Fozard et al. proposed that the initiation of migraine is caused by the activation of the 5-HT 2C R [78]. However, this hypothesis was challenged after the cloning of rat 5-HT 2B R in 1992 [4]. Subsequent studies demonstrated that the 5-HT 2B R activation stimulated nitric oxide (NO) synthesis, which may be involved in migraine pathogenesis [75]. In guinea pigs, selective 5-HT 2B R antagonists have been found to inhibit the 5-HT 2B R/5-HT 2C R agonist meta-chlorophenylpiperazine (mCPP) or the 5-HT 2B R agonist BW723C686-induced dural plasma protein extravasation (PPE), an indicator for migraine attacks in animal models [79]. In addition, 5-HT 2B R antagonism also prevented mCPP-induced dural PPE under hypoxia in mice [80].
Increasing evidence has revealed that the 5-HT 2B R also plays a role in neuropathic pain [77]. In mouse dorsal root ganglion (DRG) neurons, the mechanical hyperalgesia induced by 5-HT or the 5-HT 2 R agonist α-m5-HT was inhibited by the 5-HT 2B R/5-HT 2C R antagonist SB-206553 [81]. Given that the 5-HT 2B R was mainly expressed in DRGs, whereas the 5-HT 2C R was detected only in trace amounts, 5-HT-induced mechanical hyperalgesia is most likely mediated by the 5-HT 2B R [81]. Another signal transduction study suggested that the 5-HT 2B R mediates the 5-HT-induced mechanical hyperalgesia through the PLCβ-PKCε pathway to regulate the function of transient receptor potential vanilloid 1 [82]. Cervantes-Durán et al. assessed the role played by peripheral and spinal 5-HT 2 Rs in formalin-induced secondary allodynia and hyperalgesia in rats. Local peripheral ipsilateral or intrathecal injection of selective 5-HT 2B R antagonist significantly prevented formalininduced nociceptive behavior monitored by flinching frequency [83]. Ipsilateral treatment with subtype-selective antagonists of 5-HT 2A R, 5-HT 2B R or 5-HT 2C R, prevented formalininduced long-term secondary mechanical allodynia and hyperalgesia [84]. Additionally, intrathecal treatment with the same antagonists inhibited formalin-induced long-term secondary mechanical allodynia and hyperalgesia in both ipsilateral and contralateral hind paws [85]. In the spinal nerve ligation-induced neuropathic pain rat model, intrathecal injection of 5-HT 2B R antagonists not only impaired spinal nerve ligation-induced allodynia but also inhibited the spinal nerve injury-induced increased expression of the 5-HT 2B R in both DRGs and spinal cord [86]. More recently, studies in female rats revealed that blocking the spinal 5-HT 2B R diminished preoperative anxiety-induced postoperative hyperalgesia [87]. However, opposite findings were reported in other pain models. For example, in a rat model of neuropathic pain induced by chronic constriction injury (CCI) of the sciatic nerve, Urtikova et al. found that intrathecal injection of the 5-HT 2B R agonist BW723C86 evidently relieved CCI-induced allodynia [88]. Clearly, further mechanistic studies are needed to explain the opposite experimental observations.

Regulation of Neuroglia Function
The 5-HT 2B R is also expressed in neuroglia, including microglia and astroglia, playing a role in regulating neuroglia function. Microglia, as the resident macrophages in the brain and the spinal cord, are responsible for the immune defense of the central nervous system (CNS) [89]. It was reported that the 5-HT 2B R was expressed on postnatal microglia and participated in postnatal brain maturation [90]. More recently, the same group showed that the ablation of the 5-HT 2B R gene in neonatal microglia was sufficient to cause enhanced weight loss and prolonged neuroinflammation in mice caused by exposure to lipopolysaccharides in adulthood. This suggested that the 5-HT 2B R is required in neonatal microglia to prevent sickness behavior in adulthood [91].
Astrocytes are primary homeostatic cells of the CNS and account for about one-quarter of brain cortical volume. The expression of 5-HTRs, including the 5-HT 2B R, has been found in both cultured and freshly isolated astrocytes [92]. Studies have suggested that conventional serotonin-specific reuptake inhibitors (SSRIs) such as fluoxetine act as agonists of astroglial 5-HT 2B R [92]. The 5-HT 2B R agonist BW723C86 could mimic the behavioral and neurogenic SSRI effects, which could be eliminated by the genetic or pharmacological inactivation of the 5-HT 2B R [93]. In cultured mouse astrocytes, fluoxetine was found to induce EGFR transactivation and ERK1/2 phosphorylation, mediated by the stimulation of the 5-HT 2B R [94], which is consistent with the observation that the drug-induced VHD involves the activation of the 5-HT 2B R and consequent ERK phosphorylation [2,20]. Similarly, 5-HT was also found to cause ERK1/2 phosphorylation, which is mediated by the stimulation of the 5-HT 2B R with high affinity and the 5-HT 2C R with low-affinity [95]. Increasing evidence has suggested that astroglial 5-HT 2B R is involved in depression [96]. In both 1-methyl-4phenyl-1,2,3,6-tetrahydropyridine-induced and 6-hydroxydopamine-induced Parkinson's disease mouse models [97,98], the decrease of astroglial 5-HT 2B R expression paralleled the development of depressive behavior. Treatment with fluoxetine corrected both the decrease of astroglial 5-HT 2B R expression and depressive behavior. All of these indicate that the downregulation of the astroglial 5-HT 2B R may promote the development of depressive behavior in Parkinson's disease. In addition, astroglial 5-HT 2B R was also found to play a role in depressive behavior associated with sleep deprivation [99,100]. Specifically, the expression of the 5-HT 2B R in a sleep deprivation mouse model was downregulated selectively in astrocytes, which was controlled by the activation of the P2X7 receptor [99]. Interestingly, leptin was found to increase the expression of astrocytic 5-HT 2B Rs and thus enhance the action of fluoxetine on depressive-like behaviors induced by sleep deprivation [100].

Regulation of the Dopaminergic Pathway
The 5-HT 2B R has been implicated in the modulation of central dopamine (DA) activity, with potential applications in DA-dependent neuropsychiatric disorders, especially in schizophrenia and drug addiction [7,101].
Schizophrenia is a serious long-term mental disorder with multimodal symptomatology, characterized by positive, negative, and cognitive symptoms [102]. There is a classical hypothesis about schizophrenia proposed that positive symptoms are the result of a specific DA hyperfunction in the nucleus accumbens (NAc), whereas negative and cognitive symptoms are associated with a DA hypofunction in the medial prefrontal cortex (mPFC) [102]. Several microdialysis studies in rats suggested that the 5-HT 2B R blockade exerts a differential control of DA ascending pathways, with increased, decreased and unaltered effects on DA outflow in the mPFC, the NAc, and the striatum, respectively. This is in accordance with the role played by DA neurotransmission in schizophrenia symptomatology [7,[103][104][105]. An additional study indicated that the distinct effects caused by 5-HT 2B R antagonists on mPFC and NAc DA outflow resulted from a functional interplay with mPFC 5-HT 1A R [102]. Moreover, behavioral experiments in rats revealed that 5-HT 2B R antagonists reduce phencyclidine-induced hyperlocomotion and reverse the phencyclidine-induced deficit in the novel object recognition test. These observations suggested that 5-HT 2B R antagonists have the potential to alleviate the positive and cognitive symptoms of schizophrenia [105]. However, it was also reported that the ablation of the 5-HT 2B R induces schizophrenic-like phenotypes, and contradictory results were observed in the DA outflow and behavior compared with 5-HT 2B R antagonism in rats [106]. Hence, additional research studies are needed in order to explain the observed discrepancies and to confirm the role of the 5-HT 2B R in the treatment of schizophrenia.
Several studies suggested that the 5-HT 2B R may represent a potential pharmaceutical target for the treatment of drug addiction. From a behavioral point of view, blocking the 5-HT 2B R may help to prevent MDMA-, amphetamine-, and cocaine-induced hyperlocomotion [103,104,107]. However, neurochemical responses vary according to different drugs. For example, 5-HT 2B R antagonists inhibit MDMA-and amphetamine-induced DA outflow in the NAc [103,107], but no effects were observed on cocaine-induced DA outflow in the NAc [104]. Recent findings showed that the dorsal raphe nucleus 5-HT 2B R blockade suppresses cocaine-induced hyperlocomotion resulting from the facilitation of mPFC DA outflow, which would subsequently inhibit accumbal DA neurotransmission [108].

Crystal Structures
The 5-HT 2B R belongs to the class A GPCRs, the largest subfamily of GPCRs, and shares a conserved architecture: seven transmembrane helices (I-VII) followed by an 8th helix (VIII), three extracellular loops (ECL1-ECL3), three intracellular loops (ICL1-ICL3), an extracellular N-terminus and an intracellular C-terminus [109]. The 5-HT 2B R contains 481 amino acids with a molecular mass of about 54.3 kDa.
Since the 5-HT 2B R crystal structure was first determined in 2013 [110], a total of eight crystal complex structures of the 5-HT 2B R bound with small-molecule ligands were published to date (Table 1). These include five representative ergolines (ergotamine (ERG, 1), lysergic acid diethylamine (LSD, 2), lisuride (3), methylergometrine (4, also called methylergonovine), methysergide (5)), and one selective antagonist of the 5-HT 2B R (LY266097 (6)) ( Figure 1) [110][111][112][113][114]. These ergolines are promiscuous ligands for many types of aminergic GPCRs and possess distinct functions. For example, despite ERG and LSD are both β-arrestin-biased agonists of the 5-HT 2B R, ERG has antimigraine effects whereas LSD is hallucinogenic; lisuride is structurally similar to LSD, but it shows antagonistic effects towards the 5-HT 2B R; methysergide is an antagonist of the 5-HT 2B R with an antimigraine effect that in vivo it rapidly undergoes N-demethylation and transforms into methylergometrine, a potent agonist of the 5-HT 2B R. In one crystal structure (PDB ID: 5TUD), the 5-HT 2B R binds with ERG and an antibody Fab fragment on its extracellular side in an active-like conformation [113]. Taken together, the structural information provides an unprecedented opportunity to understand the ligand recognition, activation mechanism and biased signaling. Antagonist (Gq partial agonist, β-ar-restin2 antagonist)   5-HT 2B R/ERG-Fab (PDB ID: 5TUD) structure shows distinct features compared with the other seven structures. Since the determined small-molecule-bound structures share common structural features in the transmembrane regions and the key motifs, we compared the 5-HT 2B R/ERG structure (PDB ID: 4IB4) with the 5-HT 2B R/ERG-Fab structure in order to illustrate the conformational changes of the 5-HT 2B R, using the well-known activestate (PDB ID: 3SN6) [116] and inactive-state (PDB ID: 2RH1) [117] β 2 -adrenergic receptor (β 2 AR) structures as references (Figure 3a). Upon activation, an outward movement of the intracellular helix VI (Figure 3a), and an inward shift of helix VII along with a sidechain rotation of Y 7.53 (Figure 3b), are believed to represent common features in class A GPCRs [118,119]. When compared to the 5-HT 2B R/ERG structure, the 5-HT 2B R/ERG-Fab structure shows a more evident outward movement of the intracellular helix VI (Figure 3a). The backbone atoms of helix VII in both the 5-HT 2B R/ERG and the 5-HT 2B R/ERG-Fab structures overlapped well with the active-state β 2 AR structure, whereas the orientation of the Y 7.53 side chain in 5-HT 2B R/ERG is slightly different (Figure 3b). In the PIF motif, the three residues in the 5-HT 2B R/ERG-Fab structure show active-like conformations. Although the P229 5.50 and the I143 3.40 residues show active-like conformations in the 5-HT 2B R/ERG structure, the F333 6.44 residue shows an inactive-like conformation (Figure 3c). In the D(E)RY motif, the R153 3.50 in the 5-HT 2B R/ERG-Fab structure displays an extended conformation towards helix VI, similar to the active-state β 2 AR structure. Hence, the salt bridge between D152 3.49 and R153 3.50 is fully broken (Figure 3d). In summary, considering the helical movement and microswitches, the 5-HT 2B R/ERG-Fab structure shows an activelike state, whereas the 5-HT 2B R/ERG and other small-molecule bound 5-HT 2B R structures are in intermediate states.  5-HT2BR/ERG-Fab (PDB ID: 5TUD) structure shows distinct features compared with the other seven structures. Since the determined small-molecule-bound structures share common structural features in the transmembrane regions and the key motifs, we compared the 5-HT2BR/ERG structure (PDB ID: 4IB4) with the 5-HT2BR/ERG-Fab structure in order to illustrate the conformational changes of the 5-HT2BR, using the well-known ac- HT2BR/ERG structure, the F333 6.44 residue shows an inactive-like conformation (Figure 3c). In the D(E)RY motif, the R153 3.50 in the 5-HT2BR/ERG-Fab structure displays an extended conformation towards helix VI, similar to the active-state β2AR structure. Hence, the salt bridge between D152 3.49 and R153 3.50 is fully broken (Figure 3d). In summary, considering the helical movement and microswitches, the 5-HT2BR/ERG-Fab structure shows an active-like state, whereas the 5-HT2BR/ERG and other small-molecule bound 5-HT2BR structures are in intermediate states.

Orthosteric Binding Pocket (OBP)
All of the small-molecule ligands in the published 5-HT 2B R co-crystal structures occupy the OBP that is presumed as the 5-HT's binding pocket, as well as regions outside the OBP and close to the extracellular loops, termed as the extended binding pocket (EBP) [120]. Taking LSD's binding mode as an example, the ergoline ring system occupies the OBP by interacting residues in helices III, V, VI, and VII ( Figure 4a). These interactions are commonly observed in aminergic GPCRs structures, including a conserved salt bridge between the positively charged nitrogen of the ligand and the carboxylic acid of D 3.32 , and π-π stacking interactions formed by the aromatic system of the ligand with residues F 6.51 and F 6.52 in helix VI. In the EBP, the diethylamide moiety of LSD interacts with residues in the extracellular side of helices III and VII as well as ECL2.

Extended Binding Pocket (EBP)
The role played by the EBP in receptor activation has been illustrated by comparing chemically similar but functionally distinct compounds such as LSD (5-HT2BR agonist) and lisuride (5-HT2BR antagonist). LSD and lisuride bear different substitutions on the Hydrogen bond interactions are shown as dashed lines. Molecular images were generated using the UCSF Chimera software [115].
A comparison of the binding modes of ergoline ligands (ERG, LSD, lisuride, and methylergometrine) shows subtle differences ( Figure 4b). Different orientations of the ergoline scaffolds seem to be caused by the different ergoline substituents, along with the rotation of D135 3.32 side chains and the different orientations of the indole nitrogens towards helix V residues. The indole nitrogen atom of LSD points towards the backbone carbonyl oxygen of G221 5.42 , whereas the indole nitrogen atoms of ERG, lisuride, and methysergide point towards the A225 5.46 residue and form hydrogen bonds (for ERG) or electrostatic interactions (for lisuride and methysergide) with the hydroxyl oxygen of L362. In fact, both T140 3.37 and A225 5.46 have been suggested to play significant roles in the activation mechanism [114]. On one hand, the T140A 3.37 mutant substantially disrupted methylergometrine's G q agonism and β-arrestin2 recruitment. On the other hand, the A225G 5.46 mutant restored both methysergide's G q agonism and β-arrestin2 recruitment. Moreover, the crystal structure of methysergide-bound 5-HT 2B R with A225G 5.46 mutation further explained the methysergide antagonism. Methysergide adopts a similar binding mode with methylergometrine in the 5-HT 2B R/methylergometrine structure (Figure 4c), and its N-methyl group (which is the only difference between the two ligands) is oriented towards the mutated G225 5.46 residue, suggesting that the A225G 5.46 mutant results in larger space accommodate the ligands. Thus, pushing outward on helix V may be responsible for methysergide's antagonism. Hence, both A225 5.46 and T140 3.37 are key residues in the OBP for the 5-HT 2B R activation with equal contributions for G q and β-arrestin2 activation mechanisms.

Extended Binding Pocket (EBP)
The role played by the EBP in receptor activation has been illustrated by comparing chemically similar but functionally distinct compounds such as LSD (5-HT 2B R agonist) and lisuride (5-HT 2B R antagonist). LSD and lisuride bear different substitutions on the ergoline scaffold, containing an (R)-diethylamide and an (S)-diethylurea substitution, respectively ( Figure 1). Therefore, the main differences between their binding modes arise from the EBP. Lisuride's (S)-diethylurea points towards helix III, forming hydrophobic interactions with W131 3.28 and L132 3.29 , and an additional hydrogen bond with D135 3.32 (Figure 4d). While LSD's (R)-diethylamide interacts with helices III and VII, and forms additional interactions with L362 7.35 in helix VII. Consequently, contact with helix VII appears to be essential for 5-HT 2B R activation. A mutagenesis experiment of the L362 7.35 residue in the 5-HT 2B R has highlighted its crucial role in receptor activation, as mutants facilitating the contact with helix VII (L362N 7.35 , L362F 7.35 , L362Y 7.35 ) restored lisuride's G q agonism. On the contrary, L362A 7.35 mutant impaired LSD's G q agonism [114].
Moreover, the ligand recognition at helix VII in the EBP was also suggested to play an essential role in biased signaling. The L362F 7.35 mutant restored lisuride's G q agonism, but not β-arrestin2 recruitment agonism [114]. For LSD, the L362F 7.35 mutant abolished LSD's β-arrestin2 recruitment without affecting its G q agonism [114]. LY266097, a selective 5-HT 2B R antagonist with modest G q partial agonism and potent β-arrestin2 antagonism, interacts with residues in helix VII, including L362 7.35 . LY266097's agonist potency was abolished in the L362F 7.35 mutant [114]. Moreover, structure-activity relationship (SAR) studies showed that an analog of LY266097 lacking substituents on the benzyl moiety was able to fully abolish its G q agonism [114].
Notably, residues in ECL2 have also been indicated to play a role in LSD's long residence time and biased signaling [112]. Molecular dynamics (MD) simulation results suggested that the fluctuation of the "lid" in ECL2 (residues 207-214) may influence LSD's dissociation and that the L209 ECL2 A mutant may act by increasing the flexibility of the lid [112]. Further experimental verification found that the L209 ECL2 A mutant decreased LSD's residence time and significantly reduced the β-arrestin2 recruitment potency and efficacy without affecting G q agonism [112]. In addition, LSD's β-arrestin2 recruitment was found to be time-dependent for both the 5-HT 2A R and the 5-HT 2B R. The L229 ECL2 A and L209 ECL2 A mutants could selectively abolish the 5-HT 2A R's, and the 5-HT 2B R's timedependent β-arrestin2 recruitment, respectively [112].
In summary, key residues have been suggested to play important roles in receptor activation and biased signaling. Ligand recognition at helices III (T140 3.37 ) and V (G221 5.42 , A225 5.46 ) in the OBP appear to contribute equivalently to G q and β-arrestin2 potency, whereas ligand recognition at helix VII (L362 7.35 ) and ECL2 (L209 ECL2 ) contributes to either G q or β-arrestin2 activity. All of these would guide the discovery of novel 5-HT 2B R ligands, especially for biased ligands. It is expected that biased ligands may provide more possibilities for enhancing therapeutic efficiencies and reducing adverse effects [121]. For example, a β-arrestin biased antagonist of dopamine D2 receptor (BRD5814) was shown to have considerable antipsychotic efficacy and reduced motoric side effects in a mouse model [122].

Pharmaceutical Ligands
As 5-HT 2B R agonism has been considered as a side effect related to VHD, pharmaceutical ligand discovery has focused on the potential application of 5-HT 2B R antagonism. Considering that many 5-HT 2B R ligands early identified have been previously summarized [123,124], we mainly focus on reviewing the clinical-related pharmaceutical ligands and promising antagonists of the 5-HT 2B R that were reported since 2010. Note that the K i values mentioned below were determined by radiobinding assays. MT-500 (7, RS-127445, Table 2) is a 5-HT 2B R antagonist with high affinity (K i = 0.3 nM) and high selectivity over many other 5-HT receptor subtypes (especially about 1000-fold selectivity over the closely related human 5-HT 2A R and 5-HT 2C R) [125]. In vitro functional assays demonstrated that MT-500 could inhibit the 5-HT-induced increases of inositol phosphates and intracellular calcium concentration (IC 50 = 0.1 nM) and block the 5-HT-induced contraction of the rat stomach fundus [125]. In addition, in vivo pharmacokinetic studies in rats showed that MT-500 was readily absorbed by oral or intraperitoneal routes [125]. Several studies used MT-500 as a tool compound to investigate the functions of the 5-HT 2B R [62,63,83,85,108]. In 1999, POZEN acquired MT-500 from Roche and took full charge of its development for migraine prophylaxis. The phase 1 clinical trial of MT-500 was completed and showed an encouraging safety profile. However, the development of MT-500 was discontinued in 2002 for unknown reasons.

PRX-08066
PRX-08066 (8, Table 2) is a potent (K i~3 .4 nM) and selective antagonist of the 5-HT 2B R discovered by Porvasnik et al. [126]. In the study using the MCT-induced PAH rat model, PRX-08066 eased the severity of PAH by significantly reducing the elevation in PA pressure and right ventricle hypertrophy [126].
In May 2005, PRX-08066 entered a phase 1 clinical trial. The results of a dose-escalation study in healthy volunteers indicated that PRX-08066 was well-tolerated over a dose ranged from 25 mg to 800 mg, and no serious adverse events were observed. Later, PRX-08066 phase 1b safety and pharmacodynamic studies in healthy athletes started in November 2005. When volunteers took 200 mg of PRX-08066 orally, the increase in pulmonary artery blood pressure was significantly reduced (by 40%) during hypoxic exercise and did not influence systemic blood pressure. PRX-08066 thus advanced into phase 2 clinical trials in June 2006 (NCT00345774). The short-term safety and efficacy of PRX-08066 were evaluated in patients with PH associated with chronic obstructive pulmonary disease (COPD). The results were positive and indicated that PRX-08066 could significantly lower systolic pulmonary arterial pressure and was well-tolerated without any obvious effect on systemic blood pressure. In August 2008, PRX-08066 open-label phase 2 clinical trials commenced in order to assess its safety and efficacy in PH and COPD patients (NCT00677872). However, the trial was also later terminated, with no results reported.

BF-1
BF-1 (9, Table 2) is a selective and high-affinity antagonist of the 5-HT 2B R (K i = 2.7 nM) [79]. Tested in a guinea pig model for dural neurogenic PPE induced by mCPP or BW723C86, BF-1 exhibited significant reductions of dural PPE, indicating its potential as a migraine drug [79]. In December 2012, BF-1 commenced phase 1 clinical trials prophylactic for the treatment of migraine. However, no progress has since been reported, and the clinical trial was postulated to be discontinued.  Table 2) is an antagonist of the 5-HT2BR with high affinity (Ki = nM) and good selectivity (> 2000-fold selectivity against more than 60 GPCRs, ion ch nels, and enzymes) [64]. In vivo studies in the trinitrobenzene sulfonate-induced visc hypersensitivity rat model indicated that RQ-00310941 has the therapeutic potential diarrhea-predominant IBS [64]. Specifically, RQ-00310941 could attenuate the distal co sensitivity (61% inhibition at 1 mg/kg, per os (p.o.)) and suppress the restraint stress duced defecation (95% inhibition at 10 mg/kg, p.o.) without affecting normal defeca at high dose (30 mg/kg, p.o.) [64].
In July 2015, RQ-00310941 entered phase 1 clinical trials. Specifically, the safety/ erability and pharmacokinetics in healthy subjects and the preliminary efficacy in mil moderate ulcerative colitis (UC) patients with IBS-like symptoms were investigated [1 The released results showed that RQ-00310941 was safe and well-tolerated without serious adverse events in both healthy subjects and UC patients. Although there was statistically significant difference observed between RQ-00310941 and placebo with spect to primary and secondary efficacy evaluations, most efficacy measures sugges slightly favorable outcomes compared to placebo, indicating a therapeutic potential RQ-00310941 in the treatment of IBS-like symptoms in UC patients.  Table 2) is an antagonist of the 5-HT2BR with high affinity (Ki = nM) and good selectivity (> 2000-fold selectivity against more than 60 GPCRs, ion ch nels, and enzymes) [64]. In vivo studies in the trinitrobenzene sulfonate-induced visc hypersensitivity rat model indicated that RQ-00310941 has the therapeutic potentia diarrhea-predominant IBS [64]. Specifically, RQ-00310941 could attenuate the distal co sensitivity (61% inhibition at 1 mg/kg, per os (p.o.)) and suppress the restraint stress duced defecation (95% inhibition at 10 mg/kg, p.o.) without affecting normal defeca at high dose (30 mg/kg, p.o.) [64].
In July 2015, RQ-00310941 entered phase 1 clinical trials. Specifically, the safety erability and pharmacokinetics in healthy subjects and the preliminary efficacy in mil moderate ulcerative colitis (UC) patients with IBS-like symptoms were investigated [1 The released results showed that RQ-00310941 was safe and well-tolerated without serious adverse events in both healthy subjects and UC patients. Although there wa statistically significant difference observed between RQ-00310941 and placebo with spect to primary and secondary efficacy evaluations, most efficacy measures sugge slightly favorable outcomes compared to placebo, indicating a therapeutic potential RQ-00310941 in the treatment of IBS-like symptoms in UC patients.  Table 2) is an antagonist of the 5-HT2BR with high affinity (Ki = nM) and good selectivity (> 2000-fold selectivity against more than 60 GPCRs, ion ch nels, and enzymes) [64]. In vivo studies in the trinitrobenzene sulfonate-induced visc hypersensitivity rat model indicated that RQ-00310941 has the therapeutic potential diarrhea-predominant IBS [64]. Specifically, RQ-00310941 could attenuate the distal co sensitivity (61% inhibition at 1 mg/kg, per os (p.o.)) and suppress the restraint stress duced defecation (95% inhibition at 10 mg/kg, p.o.) without affecting normal defeca at high dose (30 mg/kg, p.o.) [64].
In July 2015, RQ-00310941 entered phase 1 clinical trials. Specifically, the safety/ erability and pharmacokinetics in healthy subjects and the preliminary efficacy in mil moderate ulcerative colitis (UC) patients with IBS-like symptoms were investigated [1 The released results showed that RQ-00310941 was safe and well-tolerated without serious adverse events in both healthy subjects and UC patients. Although there was statistically significant difference observed between RQ-00310941 and placebo with spect to primary and secondary efficacy evaluations, most efficacy measures sugges slightly favorable outcomes compared to placebo, indicating a therapeutic potential RQ-00310941 in the treatment of IBS-like symptoms in UC patients. AnaMar AB company is focused on the discovery and development of 5-HT2BR tagonists to prevent pathological inflammatory and fibrotic processes. Apart fr AMAP102 (10) (Section 4.1.4.) and AM1030 (11) (Section 4.1.5.), which were mentio above, they have also developed EXT5 (13), EXT9 (14), AM1125 (15), and AM1476 ( The exact structures of these compounds are undisclosed. EXT5 and EXT9 are both b zylidene aminoguanidine derivatives [36], while AM1125 and AM1476 have general mula containing a 1-amidino-3-aryl-2-pyrazoline scaffold [130]. EXT5 (13) and EXT9 (14) are both mainly antagonizing the 5-HT2BR (EXT5: Ki = nM, IC50 = 82 nM in IP1 accumulation assay; EXT9: Ki = 26 nM, IC50 = 29 nM in IP1 ac mulation assay), and also exhibit low to moderate binding affinities to the 5-HT2AR the 5-HT2CR. As mentioned above (Section 2.2.2.), EXT5 and EXT9 were utilized for in tigating the role played by the 5-HT2BR in fibrosis [36]. In vitro, the co-cultivation of T β1 and 5-HT resulted in an increased α-SMA and proteoglycan production, which co be significantly decreased after the treatment with either EXT5 or EXT9. Additionally vivo studies on BLM-treated mice showed that both EXT5 and EXT9 could attenuate fibrotic tissue remodeling, demonstrated by a decrease in tissue density, collagen-prod ing cells, myofibroblasts, and decorin expression [36]. Further gene expression stu suggested that the antiproliferative effects of EXT5 and EXT9 may be associated with pAkt/p21 signaling pathway, a cell-cycle regulation pathway [38]. More recently, E  Table 2) is an orally available antagonist of the 5-HT 2B R (structure and binding affinity are undisclosed) [127]. In vitro studies showed that AMAP102 decreased the release of proinflammatory cytokine TNF-α in human macrophages and IL-6 in rat synovial fibroblasts [127]. In vivo studies on several animal models, including collagen and glucose-6-phosphate isomerase-induced arthritis induced arthritis in mice and antigeninduced arthritis in rats, showed that AMAP102 exhibited anti-arthritic effects and reduced inflammatory pain responses [127].
In April 2009, AMAP102 entered a phase 1 clinical trial (NCT00995605) to evaluate its safety and tolerability in healthy subjects. The trial was successfully completed in August 2009, and the results showed that AMAP102 was safe and well-tolerated without any serious adverse events. In October 2014, AnaMar AB company reported the phase 2a results of AMAP102 for the treatment of inflammatory pain in osteoarthritis patients. However, compared with placebo, AMAP102 did not show a statistically significant reduction in pain over a 28-day period, and the AMAP102 clinical trial was thus discontinued.

AM1030
AM1030 (11, Table 2) [128], an aminoguanidine derivative (structure is not disclosed), is a 5-HT 2B R antagonist (K i = 330 nM). In vivo and in vitro studies on various human and rodent models suggested that AM1030 has therapeutic potential in various inflammatory diseases and is able to significantly reduce both T cell-dependent and T cell-independent inflammatory responses. Moreover, a first-in-man study in atopic dermatitis (AD) patients showed that topical administration of AM1030 was suitable for treatment. AM1030 phase 1/2 clinical trials studied in AD patients were completed in June 2015 (NCT02379910), but the results were not reported. Considering no development has since been disclosed, the development of AM1030 is assumed to be discontinued.
In July 2015, RQ-00310941 entered phase 1 clinical trials. Specifically, the safety/tolerability and pharmacokinetics in healthy subjects and the preliminary efficacy in mild to moderate ulcerative colitis (UC) patients with IBS-like symptoms were investigated [129]. The released results showed that RQ-00310941 was safe and well-tolerated without any serious adverse events in both healthy subjects and UC patients. Although there was no statistically significant difference observed between RQ-00310941 and placebo with respect to primary and secondary efficacy evaluations, most efficacy measures suggested slightly favorable outcomes compared to placebo, indicating a therapeutic potential for RQ-00310941 in the treatment of IBS-like symptoms in UC patients.
EXT5 (13) and EXT9 (14) are both mainly antagonizing the 5-HT 2B R (EXT5: K i = 45 nM, IC 50 = 82 nM in IP1 accumulation assay; EXT9: K i = 26 nM, IC 50 = 29 nM in IP1 accumulation assay), and also exhibit low to moderate binding affinities to the 5-HT 2A R and the 5-HT 2C R. As mentioned above (Section 2.2.2), EXT5 and EXT9 were utilized for investigating the role played by the 5-HT 2B R in fibrosis [36]. In vitro, the co-cultivation of TGF-β1 and 5-HT resulted in an increased α-SMA and proteoglycan production, which could be significantly decreased after the treatment with either EXT5 or EXT9. Additionally, in vivo studies on BLM-treated mice showed that both EXT5 and EXT9 could attenuate the fibrotic tissue remodeling, demonstrated by a decrease in tissue density, collagenproducing cells, myofibroblasts, and decorin expression [36]. Further gene expression studies suggested that the antiproliferative effects of EXT5 and EXT9 may be associated with the pAkt/p21 signaling pathway, a cell-cycle regulation pathway [38]. More recently, EXT5 and EXT9 were also used for investigating the role played by the 5-HT 2B R on airway function and remodeling [131]. Studies showed that EXT5 and EXT9 inhibited 5-HT-induced bronchoconstriction, TGF-β1 release and the proliferation of smooth muscle cells [131]. Notably, the 5-HT-induced bronchoconstriction could also be suppressed by the 5-HT 2A R/5-HT 2C R antagonist ketanserin, but not by the 5-HT 2B R selective antagonists RS-127445 or PRX-08066 [131]. The inhibitory effects of bronchoconstriction may involve a combination of 5-HT 2 receptors and deserves further research efforts in the future. AM1125 (15) is a highly selective 5-HT 2B R antagonist (K i = 0.9 nM). In November 2016, preclinical data of AM1125 was presented at the ACR annual meeting [132]. Treatment with AM1125 at 50 mg/kg reduced fibrosis parameters (hypodermal thickness, myofibroblast counts, and hydroxyproline content) in the tsk-1 model of SSc, which indicated a potential treatment opportunity for SSc. In May 2017, preclinical data on the antifibrotic effects of AM1125 was reported at the ATS International Conference [42]. In in vitro studies on human lung, fibroblasts showed that treatment with AM1125 significantly reduced TGF-β mRNA, plasminogen activator inhibitor 1 mRNA, and phosphorylated Smad 2/3. Further in vivo studies on BLM-induced pulmonary fibrosis mice showed that the oral administration of AM1125 ameliorated pulmonary fibrosis with a reduction of the fibrotic area, myofibroblast counts and the amount of collagen protein. These results reflected the potential of AM1125 for the treatment of pulmonary fibrosis.
AM1476 (16), an orally available 5-HT 2B R antagonist with high selectivity (activity data are undisclosed), is currently in the late preclinical phase and under development for SSc [43]. Orally administration of AM1476 in the murine sclerodermatous chronic graft-versus-host disease model showed that it could significantly reduce all measured dermal and pulmonary fibrosis readouts [43]. Dermal fibrosis studied in the tsk-1 model of SSc showed reduced hypodermal thickening and the number of myofibroblast and hydroxyproline content. In addition, the number of pSmad3 positive cells was significantly reduced in skin samples suggested the inhibitory effect on the TGF-β/Smad signaling pathway [43].

Bis-Amino-Triazine Derivatives
Using the structure-based hierarchical virtual screening approach [133], Huang et al. have identified a series of bis-amino-triazine derivatives as potent 5-HT 2B R antagonists [65]. Two compounds (compound 17 and 18, Figure 5) were highlighted with comparable potency in the binding and functional assays in vitro (17: K i = 7.2 nM, IC 50 = 27.3 nM in calcium flux assay; 18: K i = 6.2 nM, IC 50 = 33.4 nM in calcium flux assay). Without the classical tertiary amine group, these compounds performed good binding selectivity for the 5-HT 2B R over ten other tested 5-HT receptors [65]. In vivo studies further indicated that compound 18 could significantly attenuate visceral hypersensitivity in an IBS rat model [65].
Pharmaceuticals 2021, 14, x FOR PEER REVIEW

Bis-amino-triazine Derivatives
Using the structure-based hierarchical virtual screening approach [133 have identified a series of bis-amino-triazine derivatives as potent 5-HT2 [65]. Two compounds (compound 17 and 18, Figure 5) were highlighted w potency in the binding and functional assays in vitro (17: Ki = 7.2 nM, IC calcium flux assay; 18: Ki = 6.2 nM, IC50 = 33.4 nM in calcium flux assay classical tertiary amine group, these compounds performed good binding the 5-HT2BR over ten other tested 5-HT receptors [65]. In vivo studies fur that compound 18 could significantly attenuate visceral hypersensitivity model [65]. The predicted binding mode for compound 17 was verified by SAR a ries of structural analogs [65]. The key interactions contain a typical salt brid between the protonated triazine ring and the carboxyl group of the cons residue, hydrogen bonds between one amino (NH) group on the triazin D135 3.32 residue, between the other amino (NH2) group and the N344 6.55 re tween the ethyl benzoate moiety and the T140 3.37 residue [65]. Notably, furth of the binding model for compound 18 showed that its potent binding aff due to a halogen bonding interaction between the bromine atom and the ca atom of the F217 5.38 residue, which was supported by the subsequent SAR simulation [134]. Originally, high throughput screening (HTS) led to the discovery of The predicted binding mode for compound 17 was verified by SAR analysis of a series of structural analogs [65]. The key interactions contain a typical salt bridge interaction between the protonated triazine ring and the carboxyl group of the conserved D135 3.32 residue, hydrogen bonds between one amino (NH) group on the triazine ring and the D135 3.32 residue, between the other amino (NH2) group and the N344 6.55 residue, and between the ethyl benzoate moiety and the T140 3.37 residue [65]. Notably, further assessment of the binding model for compound 18 showed that its potent binding affinity might be due to a halogen bonding interaction between the bromine atom and the carbonyl oxygen atom of the F217 5.38 residue, which was supported by the subsequent SAR study and MD simulation [134].

Guanidine Derivatives
Inspired by the synergistic effect of the 5-HT 2B R selective antagonist RS-127445 and the 5-HT 7 R selective antagonist SB-269970 found in the guinea pig model evaluating the antimigraine effect, Moritomo et al. discovered a series of carbonyl guanidine derivatives as dual 5-HT 2B R and 5-HT 7 R antagonists for the treatment of migraines.
Originally, high throughput screening (HTS) led to the discovery of compound 19 ( Figure 6) [135], which showed a high affinity for the 5-HT 2B R (K i = 1.8 nM) and the 5-HT 7 R (K i = 12.4 nM), but poor aqueous solubility. Further SAR studies on a series of guanidine derivatives led to the identification of compound 20 (Figure 6), with a similar binding affinity (K i = 1.8 nM for the 5-HT 2B R, K i = 17.6 nM for the 5-HT 7 R) and better aqueous solubility when compared to compound 19. The off-target assessment showed that compound 20 was selective for the 5-HT 2B R and the 5-HT 7 R over several other monoaminergic GPCRs, and further functional assay determined its antagonistic activity towards the 5-HT 2B R and the 5-HT 7 R [135]. Furthermore, in vivo studies showed that compound 20 had an inhibitory effect on 5-HT-induced dural PPE in guinea pigs at 3 mg/kg intraperitoneal administration. However, it was not able to reduce the amount of leaked protein from the dural blood vessel to the reference value at 30 mg/kg oral administration [135], which indicates a deficient oral bioavailability.

4, x FOR PEER REVIEW 19 of 32
Ki = 4.3 nM for the 5-HT7R, Figure 6) and selectivity [136]. When orally administered at 30 mg/kg in in vivo studies, compound 21 reversed the hypothermic effect of 5-carboxamidotryptamine in mice and showed a suppressing effect to normal levels on 5-HT-induced dural PPE in guinea pigs [136]. Considering about oxidation or conjugate metabolism of 9-OH in the fluorene ring of compound 21, a spiro cycloalkane ring was thus introduced to eliminate the concern about drug metabolism. This design led to a guanidine derivative 22 (Figure 6), which was found to be a high-affinity and selective antagonist of the 5-HT2BR and the 5-HT7R (Ki = 5.1 nM for the 5-HT2BR, Ki = 1.7 nM for the 5-HT7R) [137]. In vitro studies showed that both its optically pure isomers, (R)-22 and (S)-22, exhibited similar binding affinities and antagonistic activities against the 5-HT2BR and the 5-HT7R. Moreover, they both suppressed 5-HT-induced dural PPE and the amount of leaked protein to near normal levels at 10 mg/kg, p.o. in guinea pigs [137].
In order to improve oral potency and keep high affinity, the researchers further optimized guanidine 20 by taking the balance between lipophilicity and polar surface area into consideration [136]. SAR studies based on molecular modeling results led to the identification of compound 21, which showed both high affinity (K i = 4.3 nM for the 5-HT 2B R, K i = 4.3 nM for the 5-HT 7 R, Figure 6) and selectivity [136]. When orally administered at 30 mg/kg in in vivo studies, compound 21 reversed the hypothermic effect of 5-carboxamidotryptamine in mice and showed a suppressing effect to normal levels on 5-HT-induced dural PPE in guinea pigs [136].
Considering about oxidation or conjugate metabolism of 9-OH in the fluorene ring of compound 21, a spiro cycloalkane ring was thus introduced to eliminate the concern about drug metabolism. This design led to a guanidine derivative 22 (Figure 6), which was found to be a high-affinity and selective antagonist of the 5-HT 2B R and the 5-HT 7 R (K i = 5.1 nM for the 5-HT 2B R, K i = 1.7 nM for the 5-HT 7 R) [137]. In vitro studies showed that both its optically pure isomers, (R)-22 and (S)-22, exhibited similar binding affinities and antagonistic activities against the 5-HT 2B R and the 5-HT 7 R. Moreover, they both suppressed 5-HT-induced dural PPE and the amount of leaked protein to near normal levels at 10 mg/kg, p.o. in guinea pigs [137].

Chromone Derivatives
In order to study the molecular mechanism of the neuroprotective activity of 5hydroxy-2-(2-phenylethyl)chromone (5-HPEC) (23, Figure 7) [138], a natural product isolated from Imperata cylindrical, Williams and colleagues performed a screening campaign against the CNS receptors, transporters and ion channels. The results showed that 5-HPEC is a 5-HT 2B R antagonist, as was verified in radiobinding assays (K i = 2455 nM) and calcium flux functional assays (IC 50 = 8913 nM). 5-HPEC showed selectivity for the 5-HT 2B R over other 5-HT 2 Rs. A subsequent SAR study on a series of synthesized and 5-HPEC's natural analogs was performed and showed that the most potent analog, 5-hydroxy-2-(2phenylpropyl)chromone (5-HPPC) (24, Figure 7), exhibited a 10-fold improvement in the 5-HT 2B R affinity (K i = 251 nM) and was able to maintain the 5-HT 2B R antagonism [139]. Recently, further optimization of 5-HPPC guided by molecular modeling approaches helped to identify 5-hydroxy-2-(3-(3-cyanophenyl)propyl)chromone (5-HCPC) (25, Figure 7), which exhibited an improved binding affinity (K i = 79 nM) compared with 5-HPPC and maintained inhibitory activity (IC 50 = 6310 nM in calcium flux assay) at the 5-HT 2B R, as well as selectivity over the 5-HT 2A R and the 5-HT 2C R [140]. It is worth mentioning that these chromone derivatives are non-nitrogenous, which are different from the typical nitrogencontaining ligands of the 5-HT 2B R. Although the binding modes of this type of ligands were predicted by molecular docking, the evidence was insufficient to effectively demonstrate that these non-nitrogenous ligands bind to the orthosteric site of the 5-HT 2B R. Considering the relatively weak cellular activity of 5-HCPC with much stronger binding affinity, it is not possible to exclude the possibility that they represent allosteric-site binders.

C4 Phenyl Aporphines and tris-(phenylalkyl)amines
The Harding group at Hunter college focused on the synthesis and evaluation of nantenine analogs as 5-HT2AR antagonists. Nantenine (26, Figure 8) is a natural product with binding affinities to a number of CNS receptors (α1AR: Ki = 2 nM; 5-HT2AR: Ki = 850 nM, 5-HT2BR: Ki = 534 nM) [141]. In order to increase the 5-HT2AR affinity of nantenine, a series of nantenine analogs were designed. Surprisingly, nantenine analogs bearing a phenyl ring substituent at the C4 position displayed selective affinities to the 5-HT2BR [142]. Compound 27 (Figure 8) exhibited the best binding affinity (Ki = 96 nM) and was found to be a 5-HT2BR antagonist (IC50 = 1000 nM in calcium flux assay) with good selectivity over other tested CNS receptors [142]. In addition, in order to investigate whether the molecular rigidity of the aporphine template of nantenine affects the 5-HT2AR antagonism, a series of tris-(phenylalkyl)amines with increased flexibility compared with nantenine were synthesized. Similarly, these tris-(phenylalkyl)amines were found to have a high affinity and selectivity, as well as antagonist activity to the 5-HT2BR [143]. Among them, compound 28 ( Figure 8) showed the best binding affinity to the 5-HT2BR (Ki = 4.1 nM, IC50 = 1259 nM in calcium flux assay), with a > 30-fold selectivity over the 5-HT2AR and the 5-HT2CR [143].

C4 Phenyl Aporphines and Tris-(phenylalkyl)amines
The Harding group at Hunter college focused on the synthesis and evaluation of nantenine analogs as 5-HT 2A R antagonists. Nantenine (26, Figure 8) is a natural product with binding affinities to a number of CNS receptors (α 1A R: K i = 2 nM; 5-HT 2A R: K i = 850 nM, 5-HT 2B R: K i = 534 nM) [141]. In order to increase the 5-HT 2A R affinity of nantenine, a series of nantenine analogs were designed. Surprisingly, nantenine analogs bearing a phenyl ring substituent at the C4 position displayed selective affinities to the 5-HT 2B R [142]. Compound 27 (Figure 8) exhibited the best binding affinity (K i = 96 nM) and was found to be a 5-HT 2B R antagonist (IC 50 = 1000 nM in calcium flux assay) with good selectivity over other tested CNS receptors [142]. In addition, in order to investigate whether the molecular rigidity of the aporphine template of nantenine affects the 5-HT 2A R antagonism, a series of tris-(phenylalkyl)amines with increased flexibility compared with nantenine were synthesized. Similarly, these tris-(phenylalkyl)amines were found to have a high affinity and selectivity, as well as antagonist activity to the 5-HT 2B R [143]. Among them, compound 28 (Figure 8

Biphenyl Amide Derivatives
Gabr et al. identified a series of biphenyl amide derivatives as 5-HT 2B R antagonists by rational drug design utilizing a pharmacophore-based approach [144]. The pharmacophore map was based on a previously published doxepin induced-fit model of the 5-HT 2B R [65]. The pharmacophore of the lead compound 29 (Figure 9) [145], a potent 5-HT 2B R antagonist (IC 50 = 2.4 nM in calcium flux assay) with poor potency in the presence of human serum albumin (HSA) (IC 50 = 1200 nM in 4% HSA), was initially overlaid with the receptor-based pharmacophore in order to provide directions for optimization. Finally, compound 30 (K i = 4.5 nM, IC 50 = 14.1 nM in calcium flux assay, Figure 9) was identified with high potency and selectivity for the 5-HT 2B R over six other 5-HT receptors. In vitro, pharmacokinetic profile evaluation showed that compound 30 was able to almost completely maintain its antagonistic potency in the presence of 4% HSA (IC 50 = 18.7 nM). In terms of the predicted binding mode, compound 30 could form an additional hydrogen bond with residue N344 6.55 , and hydrophobic interactions with several residues in the ECL2 of the receptor compared with compound 29 [145]. These additional interactions were regarded as contributing to both the potency and the selectivity of compound 30.

C4 Phenyl Aporphines and tris-(phenylalkyl)amines
The Harding group at Hunter college focused on the synthesis and evaluation of nantenine analogs as 5-HT2AR antagonists. Nantenine (26, Figure 8) is a natural product with binding affinities to a number of CNS receptors (α1AR: Ki = 2 nM; 5-HT2AR: Ki = 850 nM, 5-HT2BR: Ki = 534 nM) [141]. In order to increase the 5-HT2AR affinity of nantenine, a series of nantenine analogs were designed. Surprisingly, nantenine analogs bearing a phenyl ring substituent at the C4 position displayed selective affinities to the 5-HT2BR [142]. Compound 27 (Figure 8) exhibited the best binding affinity (Ki = 96 nM) and was found to be a 5-HT2BR antagonist (IC50 = 1000 nM in calcium flux assay) with good selectivity over other tested CNS receptors [142]. In addition, in order to investigate whether the molecular rigidity of the aporphine template of nantenine affects the 5-HT2AR antagonism, a series of tris-(phenylalkyl)amines with increased flexibility compared with nantenine were synthesized. Similarly, these tris-(phenylalkyl)amines were found to have a high affinity and selectivity, as well as antagonist activity to the 5-HT2BR [143]. Among them, compound 28 ( Figure 8) showed the best binding affinity to the 5-HT2BR (Ki = 4.1 nM, IC50 = 1259 nM in calcium flux assay), with a > 30-fold selectivity over the 5-HT2AR and the 5-HT2CR [143].

Biphenyl Amide Derivatives
Gabr et al. identified a series of biphenyl amide derivatives as 5-HT2BR antagonists by rational drug design utilizing a pharmacophore-based approach [144]. The pharmacophore map was based on a previously published doxepin induced-fit model of the 5-HT2BR [65]. The pharmacophore of the lead compound 29 (Figure 9) [145], a potent 5-HT2BR antagonist (IC50 = 2.4 nM in calcium flux assay) with poor potency in the presence of human serum albumin (HSA) (IC50 = 1200 nM in 4% HSA), was initially overlaid with the receptor-based pharmacophore in order to provide directions for optimization. Finally, compound 30 (Ki = 4.5 nM, IC50 = 14.1 nM in calcium flux assay, Figure 9) was identified with high potency and selectivity for the 5-HT2BR over six other 5-HT receptors. In vitro, pharmacokinetic profile evaluation showed that compound 30 was able to almost completely maintain its antagonistic potency in the presence of 4% HSA (IC50 = 18.7 nM). In terms of the predicted binding mode, compound 30 could form an additional hydrogen bond with residue N344 6.55 , and hydrophobic interactions with several residues in the ECL2 of the receptor compared with compound 29 [145]. These additional interactions were regarded as contributing to both the potency and the selectivity of compound 30.

Other pharmaceutical ligands of the 5-HT2BR
In 2010, a urea type of ligand 34 (Ki = 42 nM, Table 3) was coincidentally di as a 5-HT2BR antagonist by Kwon et al. [147]. Notably, this compound showed s lectivity for the 5-HT2BR over other serotonin receptors, as well as dopamine, hi muscarinic and opiate receptors. In 2011, C-122 (35 , Table 3) was reported by Z as a 5-HT2BR antagonist (Ki = 5.2 nM) [148]. It is worth mentioning that C-122 is n tive for the 5-HT2BR and also exhibit binding affinities against serotonin receptor the 5-HT7R (Ki = 4.4 nM), the 5-HT2AR (Ki = 61 nM), and several other monoa GPCRs. In vivo studies on the MCT-induced PAH rat model showed that C-122 p MCT-induced elevations in the pulmonary arterial circuit pressure, right ventric pertrophy and pulmonary arteriole muscularization when orally administer mg/kg daily for 3 weeks [148]. In 2015, Rodrigues et al. carried out a proof-of study of a large-scale multidimensional de novo design approach, combining c tional molecular design and quantitative activity prediction with microfluidics s and discovered new chemical entities for the 5-HT2BR [149]. Several computatio signed compounds with a good predicted affinity and selectivity were subjected imental validation. As a result, piperazine 36 (Table 3) was identified as a 5-HT2B onist (Ki = 251 nM), with high binding and functional selectivity [149]. In 2016, con that adenosine derivatives were reported with micromolar activity at the 5-HT2BR 5-HT2CR, Tosh et al. applied a structure-based drug design approach to further  Table 3) was coincidentally discovered as a 5-HT 2B R antagonist by Kwon et al. [147]. Notably, this compound showed strong selectivity for the 5-HT 2B R over other serotonin receptors, as well as dopamine, histamine, muscarinic and opiate receptors. In 2011, C-122 (35, [148]. It is worth mentioning that C-122 is not selective for the 5-HT 2B R and also exhibit binding affinities against serotonin receptors such as the 5-HT 7 R (K i = 4.4 nM), the 5-HT 2A R (K i = 61 nM), and several other monoaminergic GPCRs. In vivo studies on the MCT-induced PAH rat model showed that C-122 prevented MCTinduced elevations in the pulmonary arterial circuit pressure, right ventricular hypertrophy and pulmonary arteriole muscularization when orally administered at 10 mg/kg daily for 3 weeks [148]. In 2015, Rodrigues et al. carried out a proof-of-concept study of a large-scale multidimensional de novo design approach, combining computational molecular design and quantitative activity prediction with microfluidics synthesis, and discovered new chemical entities for the 5-HT 2B R [149]. Several computationally designed compounds with a good predicted affinity and selectivity were subjected to experimental validation. As a result, piperazine 36 (Table 3) was identified as a 5-HT 2B R antagonist (K i = 251 nM), with high binding and functional selectivity [149]. In 2016, considering that adenosine derivatives were reported with micromolar activity at the 5-HT 2B R and the 5-HT 2C R, Tosh et al. applied a structure-based drug design approach to further improve the 5-HT 2 R affinity and simultaneously reduce the affinity to adenosine receptors (ARs) [150]. A SAR study on a series of adenosine derivatives assisted them in identifying several antagonists of the 5-HT 2B /5-HT 2C receptors with selectivity over ARs. Among them, compound 37 (Table 3) showed high binding affinity (K i = 23 nM) and antagonist activity for the 5-HT 2B R, with 12-fold binding selectivity and 170-fold functional selectivity for the 5-HT 2B R over the 5-HT 2C R [150]. In 2018, Rataj et al. combined a fingerprint-based machine learning approach and molecular docking that led to the identification of compound 38 (Table 3) [151]. Notably, compound 38 showed potent binding affinity (K i = 0.3 nM) and >10,000-fold selectivity over other five tested 5-HTRs [151]. In addition, in 2018, as a proof-of-concept of halogen bonding in designing 5-HT 2B R ligands, a series of halogensubstituted analogs of doxepin (K i = 25.3 nM for the 5-HT 2B R) were synthesized. As expected, the bromine-substituted compound 39 (Table 3) showed a 10-fold increased binding affinity towards the 5-HT 2B R (K i = 2.5 nM), a 10-fold improvement when compared to doxepin, and exhibited superior potency in a mouse model of diarrhea [134].

Summary of Binding Features of 5-HT 2B R Ligands
The crystal structures of the 5-HT 2B R facilitate our detailed understanding of receptorligand binding interactions. Therefore, we generated a receptor-based pharmacophore model (Figure 11a) to reconcile the critical binding interactions of the pharmaceutical ligands mentioned above. Notably, co-crystal ligands fit well in this pharmacophore model ( Figure 11b). Most of the 5-HT 2B R ligands bear positively charged nitrogens or polar NH groups to form favorable salt bridge or hydrogen bond interaction with the carboxylic side chain of residue D135 3.32 (Figure 11a). Moreover, aromatic rings or hydrophobic fragments of 5-HT 2B R ligands shall form π-π stacking or hydrophobic interactions with residues in the OBP in helices III, V, VI and VII; and residues in the EBP in helices III and VII, as well as in the ECL2 (Figure 11a). Such a pharmacophore model can be expected to guide the discovery and development of new 5-HT 2B R ligands. Table 3. Other representative pharmaceutical ligands of the 5-HT 2B R.

Compound Name
Structure Activity Data 34 [147] the 5-HT2R affinity and simultaneously reduce the affinity to adenosine receptors (ARs) [150]. A SAR study on a series of adenosine derivatives assisted them in identifying several antagonists of the 5-HT2B/5-HT2C receptors with selectivity over ARs. Among them, compound 37 (Table 3) showed high binding affinity (Ki = 23 nM) and antagonist activity for the 5-HT2BR, with 12-fold binding selectivity and 170-fold functional selectivity for the 5-HT2BR over the 5-HT2CR [150]. In 2018, Rataj et al. combined a fingerprint-based machine learning approach and molecular docking that led to the identification of compound 38 (Table 3) [151]. Notably, compound 38 showed potent binding affinity (Ki = 0.3 nM) and > 10,000-fold selectivity over other five tested 5-HTRs [151]. In addition, in 2018, as a proofof-concept of halogen bonding in designing 5-HT2BR ligands, a series of halogen-substituted analogs of doxepin (Ki = 25.3 nM for the 5-HT2BR) were synthesized. As expected, the bromine-substituted compound 39 (Table 3) showed a 10-fold increased binding affinity towards the 5-HT2BR (Ki = 2.5 nM), a 10-fold improvement when compared to doxepin, and exhibited superior potency in a mouse model of diarrhea [134].

Summary of Binding Features of 5-HT2BR Ligands
The crystal structures of the 5-HT2BR facilitate our detailed understanding of receptor-ligand binding interactions. Therefore, we generated a receptor-based pharmacophore model (Figure 11a) to reconcile the critical binding interactions of the pharmaceutical ligands mentioned above. Notably, co-crystal ligands fit well in this pharmacophore model (Figure 11b). Most of the 5-HT2BR ligands bear positively charged nitrogens or polar NH K i = 42 nM 35 [148] (C-122) the 5-HT2R affinity and simultaneously reduce the affinity to adenosine receptors (ARs) [150]. A SAR study on a series of adenosine derivatives assisted them in identifying several antagonists of the 5-HT2B/5-HT2C receptors with selectivity over ARs. Among them, compound 37 (Table 3) showed high binding affinity (Ki = 23 nM) and antagonist activity for the 5-HT2BR, with 12-fold binding selectivity and 170-fold functional selectivity for the 5-HT2BR over the 5-HT2CR [150]. In 2018, Rataj et al. combined a fingerprint-based machine learning approach and molecular docking that led to the identification of compound 38 (Table 3) [151]. Notably, compound 38 showed potent binding affinity (Ki = 0.3 nM) and > 10,000-fold selectivity over other five tested 5-HTRs [151]. In addition, in 2018, as a proofof-concept of halogen bonding in designing 5-HT2BR ligands, a series of halogen-substituted analogs of doxepin (Ki = 25.3 nM for the 5-HT2BR) were synthesized. As expected, the bromine-substituted compound 39 (Table 3) showed a 10-fold increased binding affinity towards the 5-HT2BR (Ki = 2.5 nM), a 10-fold improvement when compared to doxepin, and exhibited superior potency in a mouse model of diarrhea [134].

Summary of Binding Features of 5-HT2BR Ligands
The crystal structures of the 5-HT2BR facilitate our detailed understanding of receptor-ligand binding interactions. Therefore, we generated a receptor-based pharmacophore model (Figure 11a) to reconcile the critical binding interactions of the pharmaceutical ligands mentioned above. Notably, co-crystal ligands fit well in this pharmacophore model (Figure 11b). Most of the 5-HT2BR ligands bear positively charged nitrogens or polar NH [149] the 5-HT2R affinity and simultaneously reduce the affinity to adenosine receptors (ARs) [150]. A SAR study on a series of adenosine derivatives assisted them in identifying several antagonists of the 5-HT2B/5-HT2C receptors with selectivity over ARs. Among them, compound 37 (Table 3) showed high binding affinity (Ki = 23 nM) and antagonist activity for the 5-HT2BR, with 12-fold binding selectivity and 170-fold functional selectivity for the 5-HT2BR over the 5-HT2CR [150]. In 2018, Rataj et al. combined a fingerprint-based machine learning approach and molecular docking that led to the identification of compound 38 (Table 3) [151]. Notably, compound 38 showed potent binding affinity (Ki = 0.3 nM) and > 10,000-fold selectivity over other five tested 5-HTRs [151]. In addition, in 2018, as a proofof-concept of halogen bonding in designing 5-HT2BR ligands, a series of halogen-substituted analogs of doxepin (Ki = 25.3 nM for the 5-HT2BR) were synthesized. As expected, the bromine-substituted compound 39 (Table 3) showed a 10-fold increased binding affinity towards the 5-HT2BR (Ki = 2.5 nM), a 10-fold improvement when compared to doxepin, and exhibited superior potency in a mouse model of diarrhea [134].

Summary of Binding Features of 5-HT2BR Ligands
The crystal structures of the 5-HT2BR facilitate our detailed understanding of receptor-ligand binding interactions. Therefore, we generated a receptor-based pharmacophore model (Figure 11a) to reconcile the critical binding interactions of the pharmaceutical ligands mentioned above. Notably, co-crystal ligands fit well in this pharmacophore model (Figure 11b). Most of the 5-HT2BR ligands bear positively charged nitrogens or polar NH K i = 251 nM 37 [150] the 5-HT2R affinity and simultaneously reduce the affinity to adenosine receptors (ARs) [150]. A SAR study on a series of adenosine derivatives assisted them in identifying several antagonists of the 5-HT2B/5-HT2C receptors with selectivity over ARs. Among them, compound 37 (Table 3) showed high binding affinity (Ki = 23 nM) and antagonist activity for the 5-HT2BR, with 12-fold binding selectivity and 170-fold functional selectivity for the 5-HT2BR over the 5-HT2CR [150]. In 2018, Rataj et al. combined a fingerprint-based machine learning approach and molecular docking that led to the identification of compound 38 (Table 3) [151]. Notably, compound 38 showed potent binding affinity (Ki = 0.3 nM) and > 10,000-fold selectivity over other five tested 5-HTRs [151]. In addition, in 2018, as a proofof-concept of halogen bonding in designing 5-HT2BR ligands, a series of halogen-substituted analogs of doxepin (Ki = 25.3 nM for the 5-HT2BR) were synthesized. As expected, the bromine-substituted compound 39 (Table 3) showed a 10-fold increased binding affinity towards the 5-HT2BR (Ki = 2.5 nM), a 10-fold improvement when compared to doxepin, and exhibited superior potency in a mouse model of diarrhea [134].

Summary of Binding Features of 5-HT2BR Ligands
The crystal structures of the 5-HT2BR facilitate our detailed understanding of receptor-ligand binding interactions. Therefore, we generated a receptor-based pharmacophore model (Figure 11a) to reconcile the critical binding interactions of the pharmaceutical ligands mentioned above. Notably, co-crystal ligands fit well in this pharmacophore model (Figure 11b). Most of the 5-HT2BR ligands bear positively charged nitrogens or polar NH K i = 23 nM 38 [151] Pharmaceuticals 2021, 14, x FOR PEER REVIEW 22 of 32 the 5-HT2R affinity and simultaneously reduce the affinity to adenosine receptors (ARs) [150]. A SAR study on a series of adenosine derivatives assisted them in identifying several antagonists of the 5-HT2B/5-HT2C receptors with selectivity over ARs. Among them, compound 37 (Table 3) showed high binding affinity (Ki = 23 nM) and antagonist activity for the 5-HT2BR, with 12-fold binding selectivity and 170-fold functional selectivity for the 5-HT2BR over the 5-HT2CR [150]. In 2018, Rataj et al. combined a fingerprint-based machine learning approach and molecular docking that led to the identification of compound 38 (Table 3) [151]. Notably, compound 38 showed potent binding affinity (Ki = 0.3 nM) and > 10,000-fold selectivity over other five tested 5-HTRs [151]. In addition, in 2018, as a proofof-concept of halogen bonding in designing 5-HT2BR ligands, a series of halogen-substituted analogs of doxepin (Ki = 25.3 nM for the 5-HT2BR) were synthesized. As expected, the bromine-substituted compound 39 (Table 3) showed a 10-fold increased binding affinity towards the 5-HT2BR (Ki = 2.5 nM), a 10-fold improvement when compared to doxepin, and exhibited superior potency in a mouse model of diarrhea [134].
groups to form favorable salt bridge or hydrogen bond interaction with the carboxylic side chain of residue D135 3.32 (Figure 11a). Moreover, aromatic rings or hydrophobic fragments of 5-HT2BR ligands shall form π-π stacking or hydrophobic interactions with residues in the OBP in helices III, V, VI and VII; and residues in the EBP in helices III and VII, as well as in the ECL2 (Figure 11a). Such a pharmacophore model can be expected to guide the discovery and development of new 5-HT2BR ligands. Figure 11. Key pharmacophore features of the 5-HT2BR. Positive electrostatic, hydrogen bond donor and hydrophobic pharmacophore features are colored in blue, green, and cyan, respectively. Pharmacophore features were generated based on the 5-HT2BR crystal structure (PDB ID: 5TVN) using the CavityPlus web server [152]. (a) Association of 5-HT2BR residues and pharmacophore features. Residues close the pharmacophore and contributing to the protein-ligand interactions are shown in hot pink sticks. (b) Association of co-crystal ligands and pharmacophore features. LSD (PDB ID: 5TVN), lisuride (PDB ID: 6DRX), methylergometrine (PDB ID: 6DRY), and LY266097 (PDB ID: 6DS0) are shown as hot pink, light sea green, salmon, and purple stick, respectively. These ligands are aligned based on residues in the binding pocket. Molecular images were generated using the UCSF Chimera software [115].

Conclusions
The 5-HT2BR has been implicated in multiple diseases, including cardiovascular diseases, fibrosis disorders, cancer, IBS, migraine and neuropsychiatric disorders. The increasing number of crystal complex structures of the 5-HT2BR has provided great insights into ligand recognition, activation mechanism, and biased signaling. Nonetheless, inactive-state 5-HT2BR structural information and G protein and β-arrestin bound 5-HT2BR structures are still desirable to be obtained to further promote the development of novel 5-HT2BR ligands. Although many 5-HT2BR antagonists have been identified, with several candidates having advanced into clinical trials, no approved drug currently exists for the 5-HT2BR. Based on the recently solved crystal structures of GPCRs, particularly the closest homologs 5-HT2AR and 5-HT2CR [153][154][155], more research efforts should be employed to develop subtype-selective 5-HT2BR antagonists. Moreover, most existing 5-HT2BR ligands were discovered without the consideration of biased signaling, while it is desirable to study the biased signaling of the identified 5-HT2BR ligands, which may not only facilitate to understand the therapeutic effect-related signaling but also develop therapeutic drugs to avoid side effects.  . Key pharmacophore features of the 5-HT 2B R. Positive electrostatic, hydrogen bond donor and hydrophobic pharmacophore features are colored in blue, green, and cyan, respectively. Pharmacophore features were generated based on the 5-HT 2B R crystal structure (PDB ID: 5TVN) using the CavityPlus web server [152]. (a) Association of 5-HT 2B R residues and pharmacophore features. Residues close the pharmacophore and contributing to the protein-ligand interactions are shown in hot pink sticks. (b) Association of co-crystal ligands and pharmacophore features. LSD (PDB ID: 5TVN), lisuride (PDB ID: 6DRX), methylergometrine (PDB ID: 6DRY), and LY266097 (PDB ID: 6DS0) are shown as hot pink, light sea green, salmon, and purple stick, respectively. These ligands are aligned based on residues in the binding pocket. Molecular images were generated using the UCSF Chimera software [115].

Conclusions
The 5-HT 2B R has been implicated in multiple diseases, including cardiovascular diseases, fibrosis disorders, cancer, IBS, migraine and neuropsychiatric disorders. The increasing number of crystal complex structures of the 5-HT 2B R has provided great insights into ligand recognition, activation mechanism, and biased signaling. Nonetheless, inactive-state 5-HT 2B R structural information and G protein and β-arrestin bound 5-HT 2B R structures are still desirable to be obtained to further promote the development of novel 5-HT 2B R ligands. Although many 5-HT 2B R antagonists have been identified, with several candidates having advanced into clinical trials, no approved drug currently exists for the 5-HT 2B R. Based on the recently solved crystal structures of GPCRs, particularly the closest homologs 5-HT 2A R and 5-HT 2C R [153][154][155], more research efforts should be employed to develop subtype-selective 5-HT 2B R antagonists. Moreover, most existing 5-HT 2B R ligands were discovered without the consideration of biased signaling, while it is desirable to study the biased signaling of the identified 5-HT 2B R ligands, which may not only facilitate to understand the therapeutic effect-related signaling but also develop therapeutic drugs to avoid side effects.

Conflicts of Interest:
The authors declare no conflict of interest.