5-HT_2CR has been extensively targeted for the design of allosteric modulators, with the end goal of reaching selective agonists as potential anxiolytics and antiobesity agents. These allosteric modulators are synthetic and were obtained by following various synthetic routes. The original papers provide complete details for the interested reader.

A screening of the Pharmacia (now Pfizer) chemical library led to the identification of PNU-69176E as a PAM of 5-HT_2CR. The refinement of this compound´s structural features has led to the development of several other allosteric modulators of 5-HT_2CR, as depicted in Figure 5. PNU-69176E showed a capability to increase [³H]-5-HT binding in 5-HT_2CR-transfected CHO cells at an optimum concentration of 25 μM, while higher concentrations disrupted membrane structures and inhibited radioligand binding. From a functional point of view, PNU-69176E at a concentration of 10 µM can decrease the EC₅₀ of 5-HT from 27 to 10 nm while increasing its E_max by 20%. The binding of the antagonist [³H]-mesulergine in 5-HT_2CR-transfected CHO cells was not affected by PNU-69176E. In addition, this compound was also capable of slowly inducing conformational changes in the whole receptor population that led to constitutive activation. From a structural point of view, the activity of PNU-69176E was linked to its amphipathic nature, which it has in common with amphipathic lipids that act as allosteric 5-HTR modulators. This is indeed an important observation, as the bulky alkyl moiety may anchor the compound to the cell membrane, facilitating the interaction of the polar groups in the α-D-galactopyranoside moiety with allosteric sites close to the membrane itself [71]. Research interest in PNU-69176E languished for several years until a follow-up study demonstrated an optimized synthetic route for its obtainment starting from commercial picolinic acid. This study also confirmed its efficacy and potency as a positive allosteric modulator of 5-HT_2CR and determined its selectivity over 5-HT_2AR. While a diastereomer of PNU-69176E was additionally obtained through this synthetic route, it was inactive at 5-HT_2CR and 5-HT_2AR [72].

The complex amphipathic scaffold of PNU-69176E compelled researchers to look for simplified derivatives. As the binding site for this compound was yet unknown, the authors first observed that its structure contained three distinct regions that could be important for binding, namely the lipophilic undecyl tail, the piperidine carboxamide core, and the sugar moiety. This latter part was the first point of structural exploration, being replaced with different amino alcohols that could emulate its capability to form hydrogen bonds, with hydroxy groups situated at a distance of two carbon atoms from the amide linker. Additionally, the undecyl tail was replaced in some cases with more rigid cyclohexyl or cyclohexylethyl chains.

These compounds were tested in a Ca²⁺ release assay in CHO cells stably transfected with human 5-HT_2CR in the presence of 5-HT. From a first screening at 1 nM, none of the tested compounds mediated Ca²⁺ release independently of 5-HT; however, most of them also had no significant activity as allosteric modulators. Compounds 1 and CYD-1-79 (Figure 6) were exceptions: the former was active as a NAM, while the latter showed activity as a PAM, increasing 5-HT E_max by about 20%, similarly to PNU-69176E but having no effect on its EC₅₀. Interestingly, the difference between the two compounds is one additional carbon in the spacer. Changing the lipophilic tail of 1 with a cyclohexylethyl moiety resulted in 2, which was active as a weak NAM, while all other modifications regarding this portion of the scaffold were ineffective at affording allosteric modulators. The stereochemistry of the 1,2-diol fragment of CYD-1-79 was also essential for PAM activity, with the S-configuration being preferred. The authors also showed that CYD-1-79 was selective for 5-HT_2CR, demonstrating no intrinsic activity at either 5-HT_2AR or 5-HT_2BR and showing weak or no displacement of the specific radioligand from other 5-HT receptors, dopamine receptors, and monoamine transporters. The recent disclosure of the 5-HT_2CR structure also enabled the authors to carry out the very first allosteric 5-HT_2CR docking study, which highlighted a possible binding site close to the orthosteric one. In vivo studies were carried out in rat models, proving that CYD-1-79 had a good pharmacokinetic profile (half-life after iv administration of 5 mg/kg was 6.59 h) while being effective in suppressing spontaneous ambulation as well as in attenuating cocaine cue reactivity in rats [73].

The optimization of the undecyl lipophilic tail was the focus of subsequent work by the same group, whose ultimate goal was to achieve smaller and more rigid molecules. These compounds were assayed similarly to the original series of PNU-69176E analogs. The optimization of the structure of CYD-1-79 led to several compounds endowed with similar activity at a concentration of 1 nM. Compound CTW0415 (Figure 6) emerged as the most significant PAM and exhibited optimal pharmacokinetics and physicochemical properties, combined with a very selective off-target profile. Compared to CYD-1-79, it had slightly enhanced efficacy, raising 5-HT E_max to 127%. Molecular docking studies were also carried out, corroborating the existence of the unique allosteric pocket of 5-HT_2CR [74].

The most recent developments in this line of research have seen the hybridization of the polar moiety of these compounds with a lipophilic tail derived from oleamide, thus combining two pharmacophoric groups of two known 5-HTR allosteric modulators. This approach also had the significant advantage of reducing the chemical complexity of the target compounds, as well as improving their drug-likeness. Several compounds of this series showed significant efficacy at 1 nM in improving 5-HT-mediated calcium efflux. Interestingly, while some of them were selective PAMs of 5-HT_2CR, others were described as dual 5-HT_2AR/5-HT_2CR PAMs. None of these compounds were reported as PAMs of 5-HT_2BR. A full characterization was conducted for dual PAM JPC0323 (Figure 6), which evoked a 44% increase in maximum 5-HT-induced Ca²⁺ intake and also showed negligible displacement at orthosteric binding sites of a number of GPCRs and transporters and exhibited favorable pharmacokinetic parameters. In rats, JPC0323 suppressed spontaneous ambulation in a 5-HT_2CR-dependent manner, suggesting that the compound has a preference for 5-HT_2CR over 5-HT_2AR [75].

The pace of research in the medicinal chemistry of 5-HT_2CR allosteric modulators has significantly increased over the past few years. Similarly to PNU-69176E, compound VA240 was identified as a moderate PAM of 5-HT_2CR (potentiating the effect of 5-HT with E_max = 20%) after high-throughput screening of a chemical library from Vivia Biotech. After its identification, the authors synthesized two series of analogs by modifying either the pyrimidine or the phenyl ring attached to the central indole scaffold of VA240 (Figure 7). While most of these compounds were completely inactive, modification of the pyrimidine afforded the best results, with 3-pyridyl analog VA012 exhibiting the most significant effect (35% at 10 μM). The potency of VA012 was also quantified (EC₅₀ = 16 nM), and its lack of intrinsic activity at 5-HT_2CR was confirmed. VA012 is selective over the other 5-HT₂R subtypes and a set of off-target GPCRs. The allosteric nature of VA012 binding to 5-HT_2CR was corroborated by the low displacement of the orthosteric ligands 5-HT, mesulergine, and clozapine from 5-HT_2CR. After administration i.p. in the rat (10 mg/kg), VA012 exhibited a brain-to-plasma ratio of 3.8 after 120 min; therefore, it was suitable for evaluation in animal models of obesity. VA012 was highly active in rodent feeding models, reducing both food intake and body weight gain. However, the compound induced hyperlocomotion in two behavioral tests, suggesting that such side effects might limit its therapeutic utility, at least at certain doses [76].

In another recent work, the structures of meta-chlorophenylpiperazine and a 2-phenylcyclopropylmethylamine derivative, two known 5-HT_2CR agonists with favorable pharmacodynamic and pharmacokinetic properties, were used as inspiration for the synthesis of a new series of compounds featuring a phenylcyclopropylmethane or phenylcyclopropylmethanone scaffold. This led to the discovery of two imidazole derivatives of the phenylcyclopropylmethanone series (compounds 3 and 4, Figure 8), which were active as PAMs of 5-HT_2CR but were non-selective and displayed a similar activity profile at 5-HT_2BR. Interestingly, their respective methanol derivatives lost their PAM 5-HT_2CR activity. Compounds 3 and 4 were the basis for the synthesis of further derivatives of the phenylcyclopropyl methanone class, substituting the imidazole with other heterobicyclic scaffolds or with a piperazine unit. One of the piperazine derivatives, (compound 5, Figure 8) showed activity as a 5-HT_2CR PAM, increasing 5-HT E_max to 150% at a concentration of 10 µM, as well as a 5-HT_2BR NAM, decreasing 5-HT EC₅₀ tenfold. All modifications of the scaffold of compound 5, which involved the introduction of bulky substituents on the basic piperazine nitrogen, resulted in a complete loss of PAM activity at 5-HT_2CR. Similarly, the introduction of a spacer between the phenylcyclopropyl methanone scaffold of 3 and the imidazole resulted in a complete loss of PAM activity. The 5-HT_2BR NAM activity of compound 5 was described as promising for the enhancement of the therapeutic window of non-selective orthosteric 5-HT₂R agonists.

Docking studies were carried out, further supporting the binding of these compounds to an extracellular allosteric site of 5-HT_2CR. Compounds 3 and 4 formed very similar, mostly hydrophobic interactions between the imidazole core and residues in the extracellular loop EL2 and between the terminal cyclopropyl and other hydrophobic residues. The bridging carbonyl group formed a hydrogen bond with a serine residue on transmembrane helix 2 (Figure 9A). Compound 4 formed additional hydrogen bonds with an asparagine residue on transmembrane helix 4 by virtue of its nitro group in position 4 of the imidazole ring (Figure 9B). Compound 5 behaved similarly to compounds 3 and 4, but the additional interactions caused by the benzylpiperazine moiety increased its binding energy to the receptor (Figure 9C).

Finally, the in vivo effects of compound 5 were evaluated via intracerebroventricular administration in the rat. The compound suppressed food intake after 3 and 6 h from administration, in a similar manner as the reference antiobesity drug lorcaserin. The authors speculated that since compound 5 did not exhibit any effect on cumulative food intake after 24 h, it was most likely metabolized. However, the pharmacokinetic profile of this compound was not disclosed [77].

To date, only one research paper has been published regarding the development of 5-HT_2CR PAAMs. This strategy was prompted by the withdrawal of the antiobesity drug lorcaserin from the market, which led the authors to switch away from orthosteric agonists. A virtual screening campaign afforded a series of compounds with an isoquinolin-1-one scaffold, which was capable of docking into the ergotamine (orthosteric) and the tripeptide (allosteric) binding sites of 5-HT_2CR. The synthesis of these compounds, as well as of their isoindolin-1-one analogs, allowed the authors to sketch some structure–activity relationships after their evaluation via a luciferase assay in HEK293T cells. The original isoquinolin-1-one scaffold afforded the best results, especially when linked to an aryl on the nitrogen atom, a 1-hydroxycyclopentyl moiety on C-3, and a fluorine atom on C-7 (Figure 10). The hydroxy group of the hydroxycyclopentyl moiety proved to be essential for the activity of the compounds. The most promising of these compounds, 6 (Figure 10), enhanced 5-HT potency by 3 times, with an EC₅₀¬ in the low nanomolar range. Thus, compound 6 was selected for further in vivo studies based on its relative selectivity towards the other 5-HT₂R subtypes. Compound 6 showed decent pharmacokinetic properties, quickly reaching the brain, maintaining a 1:1 brain/plasma ratio, and showing good stability in Balb/c mice brain homogenate. Moreover, studies conducted on Sprague Dawley rats showed that compound 6 was capable of significantly reducing food intake after 3 days from treatment with a 50 mg/kg dose, with comparable results to lorcaserin (5 days from treatment at 10 mg/kg) [78].

An overview of the 5-HT_2CR allosteric modulators discussed in this section is provided in Table 2.