Synthesis and In Vitro Evaluation of Novel Dopamine Receptor D2 3,4-dihydroquinolin-2(1H)-one Derivatives Related to Aripiprazole

In this pilot study, a series of new 3,4-dihydroquinolin-2(1H)-one derivatives as potential dopamine receptor D2 (D2R) modulators were synthesized and evaluated in vitro. The preliminary structure–activity relationship disclosed that compound 5e exhibited the highest D2R affinity among the newly synthesized compounds. In addition, 5e showed a very low cytotoxic profile and a high probability to cross the blood–brain barrier, which is important considering the observed affinity. However, molecular modelling simulation revealed completely different binding mode of 5e compared to USC-D301, which might be the culprit of the reduced affinity of 5e toward D2R in comparison with USC-D301.


Introduction
Dopamine primarily mediates its effect through activation of dopamine receptors (DRs) [1]. There are a total of five DR subtypes that are members of the G-protein-coupled receptors (GPCRs) [2], and are further divided into two classes according to the structure. The D 1 -like family includes D 1 Rs and D 5 Rs, whereas D 2-4 Rs belong to the D 2 -like family [1][2][3]. The main difference between both families is that D 1 -like family activate adenylate cyclase (AC), which leads to production of cyclic adenosine monophosphate (cAMP), whereas D 2 -like stimulates AC activity.
Dopamine D 2 type receptors (D 2 Rs) are integral membrane receptors coupled to G proteins with three extracellular loops, seven transmembrane domains, and three intracellular loops [4,5]. D 2 Rs are present in two isoforms, short D 2S and long D 2L , which differ by the insertion of 29 amino acids in the third intracellular loop on D 2L Rs [6]. The "short" version is exclusively expressed presynaptically as an autoreceptor, whereas, the "long" one is mainly found at the postsynaptic cells [7]. Furthermore, both isoforms can inhibit intracellular cyclic adenosine monophosphate via G i [4]. The highest levels of D 2 Rs in the human brain are expressed within striatum, the olfactory tubercle, and the nucleus

Chemistry
The chemicals were purchased from Sigma-Aldrich Co., LLC (Prague, Czech Republic) and were used without additional purification. Analytical thin-layer chromatography was carried out using plates coated with silica gel 60 with the fluorescent indicator F254 (Merck, Prague, Czech Republic). The thin-layer chromatography (TLC) plates were visualized by exposure to ultraviolet light (254 nm) or by the detection reagent ninhydrin. Column chromatography was performed using silica gel 100 at atmospheric pressure (70-230-mesh ASTM, Fluka, Prague, Czech Republic). The NMR spectra were all recorded on a Varian S500 spectrometer (500 MHz for 1 H and 126 MHz for 13 C). Chemical shifts are reported in δ ppm referenced to residual solvent signals (for 1 H NMR and 13 C NMR: chloroform-d (CDCl3; 7.26 (D) or 77.16 (C) ppm). A CEM Explorer SP 12 S was used for the MW-assisted reaction. The final compounds were analyzed by LC-MS with a Dionex Ultimate 3000 RS UHPLC system coupled with a Q Exactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) to obtain high-resolution mass spectra. Gradient LC analysis with UV detection (254 nm) confirmed >95% purity.

Chemistry
The chemicals were purchased from Sigma-Aldrich Co., LLC (Prague, Czech Republic) and were used without additional purification. Analytical thin-layer chromatography was carried out using plates coated with silica gel 60 with the fluorescent indicator F254 (Merck, Prague, Czech Republic). The thin-layer chromatography (TLC) plates were visualized by exposure to ultraviolet light (254 nm) or by the detection reagent ninhydrin. Column chromatography was performed using silica gel 100 at atmospheric pressure (70-230-mesh ASTM, Fluka, Prague, Czech Republic). The NMR spectra were all recorded on a Varian S500 spectrometer (500 MHz for 1 H and 126 MHz for 13 C). Chemical shifts are reported in δ ppm referenced to residual solvent signals (for 1 H NMR and 13 C NMR: chloroform-d (CDCl 3 ; 7.26 (D) or 77.16 (C) ppm). A CEM Explorer SP 12 S was used for the MW-assisted reaction. The final compounds were analyzed by LC-MS with a Dionex Ultimate 3000 RS UHPLC system coupled with a Q Exactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) to obtain high-resolution mass spectra. Gradient LC analysis with UV detection (254 nm) confirmed >95% purity. To a stirred solution of 3,4-dihydroquinolin-2-(1H)-one (1) (2.7 mmoL) and 60% NaH (272 mg) in DMF (20 mL), 1-bromo-3-chloropropane (2) (3.0 mmoL) was added in dropby-drop manner under ice-cooled condition. After the addition of 2, the reaction mixture was stirred for 4 h at room temperature (r.t.) [42]. After the completion of the reaction (monitored by TLC), the mixture was diluted with toluene (30 mL) and concentrated under reduced pressure. This operation was done three times. To the resulting mixture, DCM (200 mL) and distilled water (100 mL) were added, and it was vigorously stirred at r.t. for 30 min. The organic phase was then separated, dried over Na 2 SO 4 , filtered and concentrated under reduced pressure. The crude residue was purified by silica gel chromatography (DCM:EtOAc = 4:1 v/v).
The product was isolated as yellowish oil in 89% yield (540 mg); 1 H NMR of 4a agrees with the literature-reported spectra [43]. 1  To a stirred solution of 3,4-dihydroquinolin-2-(1H)-one (1) (6.1 mmoL) and 60% NaH (624 mg) in DMF (18 mL), 1-bromo-4-chlorobutane (3) (12 mmoL) was added in a dropby-drop manner under ice-cooled condition. After the addition of 3, the reaction mixture was stirred at r.t. overnight [42,44,45]. After the completion of the reaction (monitored by TLC), the mixture was diluted with toluene (30 mL) and concentrated under reduced pressure. This operation was done three times. To the resulting mixture, EtOAc (300 mL) and distilled water (100 mL) were added, and it was vigorously stirred at r.t. for 30 min. The organic phase was then separated, dried over Na 2 SO 4 , filtered and concentrated under reduced pressure. The crude residue was purified by silica gel chromatography (DCM:EtOAc = 98:2 v/v).
The product was isolated as yellowish oil in 70% yield (1.0 g); 1  To a stirred solution of appropriate analogue 4a,b (0.5 mmoL) and amine a-g (1.5 mmoL) in MeCN (5 mL), K 2 CO 3 (1.5 mmoL) was added and the reaction mixture was stirred for overnight at reflux [46]. After the completion of the reaction (monitored by TLC), the mixture was diluted with CHCl 3 (30 mL), the solid was filtered off and the residue was concentrated under reduced pressure. The crude product was purified by silica gel chromatography (DCM:MeOH = 95:5 v/v). Final compounds (5a-g, 6a-g) were prepared as hydrochlorides by mixing with small portion of hydrochloric acid (37% aq.) in MeOH at r.t.   13  . The receptor used was pdb structure 6CM4, which contains a structure of the atypical antipsychotic drug and D 2 R antagonist risperidone bound to the D 2 R [48]. An induced fit docking protocol was used, with the Triangle Matcher used for placement (10.000 placements), London dG scoring function used for initial scoring (100 conformations retained), refinement with Amber10:EHT force field, and rescoring with GBVI/WSA dG (100 conformations retained and ranked by docking score). Before docking, compounds were prepared by protonation at physiological pH and energy minimization with Amber10:EHT force field.

Thermodynamic Integration and Free Energy Calculations
Final docked conformations of risperidone, aripiprazole, USC-D301, and 5e were generated using the preceding method. For thermodynamic integration standard settings were used ( Figure S60, Supplementary Material). The used force field was Amber10:EHT. Preparation of the molecular system was done using MOE, the thermodynamic free energy calculation was performed using the PMEMD package of the AMBER molecular dynamics toolkit [47].

BBB Score Prediction
Blood-brain barrier (BBB) score of newly developed compounds was calculated using an algorithm defined by Gupta et al. [49]. A MarvinSketch software (ChemAxon Ltd., v. 20.15.0; https://www.chemaxon.com) was used to predict some of the physicochemical descriptors like number of aromatic ring, number heavy atoms, MWHBN (a descriptor comprising molecular weight, hydrogen bond donor, and hydrogen bond acceptors), topological polar surface area, and pK A . The binding reactions were terminated by filtration of the membranes through APFC filter plate (Millipore, Prague, Czech Republic) pre-soaked with 0.5% PEI and washed with ice-cold distilled water using a Brandel cell harvester (Brandel, Gaithersburg, MD, USA). Then, filters with labelled membranes were dried. After 24 h, scintillation cocktail (Rotiszint eco plus, Carl Roth) was added to each sample and radioactivity was quantified by liquid scintillation spectrometry using Wallac Microbeta scintillation counter (Wallac, Turku, Finland where y is the specific binding at free concentration x. K D values are expressed as µM and B MAX values as pmol of binding sites per mg of membrane protein.
Competition Binding

The binding of tested agonists was determined in competition experiments with 180 pM [ 3 H]-Spiperone fitting of Equation (2)
where y is the specific radioligand biding at concentration x of competitor expressed as a percent of binding in the absence of a competitor, IC 50 is the concentration causing 50% inhibition of radioligand binding. Inhibition constants K I for analyzed agonists were calculated as: where IC 50

MTT Assay
Standard MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay (Sigma Aldrich, Prague, Czech Republic) was used according to the manufacturer s protocol on the CHO-K1 (Chinese hamster ovary, ECACC, Salisbury, UK) in order to compare the effect of different compounds within the series. The cells were cultured according to ECACC recommended conditions and seeded in a density of 8000 per well as described previously [52]. Briefly, tested compounds were dissolved in DMSO (Sigma Aldrich, Prague, Czech Republic) and subsequently in the growth medium (F-12) so that the final concentration of DMSO did not exceed 1% (v/v). Cells were exposed to a tested compound for 24 h. Then the medium was replaced by a medium containing 10 µM of MTT and cells were allowed to produce formazan for another approximately 3 h under surveillance. Thereafter, medium with MTT was sucked out and crystals of formazan were dissolved in DMSO (100 µL). Cell viability was assessed spectrophotometrically by the amount of formazan produced. Absorbance was measured at 570 nm with 650 nm reference wavelength on Synergy HT (BioTek, Winooski, VT, USA). IC 50 was then calculated from the control-subtracted triplicates using non-linear regression (four parameters) of GraphPad Prism 9 software. Final IC 50 and SEM values were obtained as a mean of three independent measurements.

PAMPA Assay
PAMPA is a high-throughput screening tool applicable for prediction of the passive transport of potential drugs across the BBB [53]. In this study, it has been used as a non-cellbased in vitro assay carried out in a coated 96-well membrane filter. The filter membrane of the donor plate was coated with PBL (Polar Brain Lipid, Avanti, AL, USA) in dodecane (4 µL of 20 mg/mL PBL in dodecane) and the acceptor well was filled with 300 µL of PBS pH 7.4 buffer (V A tested compounds were dissolved first in the DMSO/phosphatebuffered saline mixture (maximum 0.5% v/v of DMSO) and subsequently diluted with phosphate-buffered saline (pH 7.4) to final concentrations of 40-100 µM in the donor wells). Concentration of DMSO did not exceed 0.5% (v/v) in the donor solution. About 300 µL of the donor solution was added to the donor wells (V D ) and the donor filter plate was carefully put on the acceptor plate so that coated membrane was "in touch" with both donor solution and acceptor buffer. Test compound diffused from the donor well through the lipid membrane (Area = 0.28 cm 2 ) to the acceptor well. The concentration of the drug in both donor and the acceptor wells were assessed after 3, 4, 5, and 6 h of incubation in quadruplicate using the UV plate reader Synergy HT (Biotek, Winooski, VT, USA) at the maximum absorption wavelength of each compound. Besides that, solution of theoretical compound concentration, simulating the equilibrium state established if the membrane were ideally permeable was prepared and assessed as well. Concentration of the compounds in the donor and acceptor well and equilibrium concentration were calculated from the standard curve and expressed as the permeability (P e ) according to the equation:

Design of Novel Compounds
Aripiprazole (Figure 1), a partial D 2 Rs agonist, belongs to the third generation of antipsychotic drugs and has been approved by the Food and Drug Administration (FDA) agency for the use as an adjunctive medication in the treatment of depressive and bipolar disorders [9,11,[54][55][56]. Aripiprazole has a unique, biased, mode of action comprising partial agonism for Gα i/o and a robust antagonism for Gβγ signaling and an antagonism or a partial agonism for β-arrestin-2 signaling [57,58]. Furthermore, if extracellular concentration of dopamine levels are high (e.g., in mesolimbic areas), aripiprazole competes with dopamine and acts as a partial antagonist. On the other hand, in the presence of low dopamine concentration (e.g., dopamine areas that are involved in working memory), aripiprazole can activate other receptors. Thus, aripiprazole can be classified as a "dopamine stabilizer" [9,59,60]. In addition, aripiprazole is a D 3 and 5-HT 1A receptors partial agonist and 5-HT 2A Rs antagonist [54,61]. In its structure, aripiprazole combines 3,4-dihydro-7hydroxyquinolin-2(1H)-one fragment attached at position 7-to 2,3-dichlorophenyl piperazine and thus is a member of large group of antipsychotics, so called 1,4-disubstituted arylpiperazines. The biological activity of this subgroup of compounds is encoded by an aromatic warhead, which controls intrinsic activity, and an amine moiety, which is responsible for the formation of a hydrogen bond to the crucial residue Asp 3.32 in the transmembrane helix 3 of D 2 R [62]. A linker controls subtype selectivity; 3-methylene linker was found suitable for D 2 R selectivity [63,64]. Aromatic/heteroaromatic appendage on the opposite site of the ligand orchestrates receptor affinity [62]. In the past decade, many compounds have been generated containing 2,3-dichlorophenylpiperazine fragment with unique pharmacological profile exhibiting high D 2 Rs affinity. From the extensive SAR, it was deduced that the central linker has only moderate impact on affinity but huge effect on functional activity at D 2 Rs [65][66][67][68]. Besides various substitutions made to the central linker, it has been shown that lipophilic appendages strongly influence functional and subtype selectivity [68][69][70]. 3,4-Dihydroquinolin-2(1H)-one scaffold-containing ligands have shown to possess an affinity to D 2 Rs as well. In this case, the nature of the central linker showed a moderate impact on D 2 Rs affinity [71]. Modifications within the amine moiety influenced D 2 Rs affinity and functional selectivity [72]. Besides, substitutions in aromatic warheads also strongly affected D 2 Rs affinity and functional and subtype selectivity [71,72]. Recently, some 3,4-dihydroquinolin-2-(1H)-one scaffold-containing compounds exhibited high D 4 Rs selectivity over other D 2 -like family receptors [73]. These findings show that small structural modifications within one region of the molecule based on aripiprazole can tune significantly the properties of the ligand.
In the study of Lopéz et al., researchers tethered arylpiperazine-like core (different phenylpiperazine moieties) with 3,4-dihydroquinolin-2(1H)-one moiety (red color, Figure 1) at position 1- [41]. USC-D301 (Figure 1) is an example of small molecule exhibiting strong D 2 R antagonism and high selectivity over D 3 Rs [41]. On the other hand, eticlopride ( Figure 1) is a substituted benzamide analog without 1,4-disubstituted arylpiperazine fragment exhibiting very high affinity for D 2 Rs [74]. Thus, we wanted to explore the effect of novel prepared analogs containing various tertiary amines with 3,4-dihydroquinolin-2(1H)-one fragment connected by aliphatic linker at position 1-of the quinolinone core towards the D 2 Rs as the main targets. We are aware of the fact that other dopamine receptors (especially D 3 R and D 4 R) are of importance for the complexity of the antipsychotic action, as observed also in case of aripiprazole [54], however, this was not the aim of this study. Molecular imaging studies have revealed that striatal D 2 Rs antagonism is essential in vivo for therapeutic doses of all neuroleptics [19,[75][76][77][78][79][80], and D 2 Rs affinity of antipsychotic drugs is the crucial for their antipsychotic efficacy. The aliphatic linkers (1,3-propane-diyl and 1,4-butane-diyl) for new ligands were chosen based on previous studies [41,44]. These new derivatives were evaluated for their D 2 R antagonistic properties with the emphasis on the structure-activity relationships regarding the type of amine and length of the linker.

Binding Affinities of Novel Compounds at D2Rs and Their Cytotoxicities
The results of the affinity of 5a-g and 6a-g for D2R are summarized in Table 1. In general, 1,3-propane-diyl derivatives 5a-g exhibited slightly better D2R antagonism than their 1,4-butane-diyl counterparts 6a-g. The aliphatic analogues 5f-g and 6g possessed slightly lower D2R antagonism than the molecules with cyclic amines 5a-e and 6a-c,e. Mor-Scheme 1. Synthesis of new quinolinone derivatives 5a-g and 6a-g.

Binding Affinities of Novel Compounds at D 2 Rs and Their Cytotoxicities
The results of the affinity of 5a-g and 6a-g for D 2 R are summarized in Table 1. In general, 1,3-propane-diyl derivatives 5a-g exhibited slightly better D 2 R antagonism than their 1,4-butane-diyl counterparts 6a-g. The aliphatic analogues 5f,g and 6g possessed slightly lower D 2 R antagonism than the molecules with cyclic amines 5a-e and 6a-c,e. Morpholinecontaining compound 6d showed the lowest D 2 R antagonism from all cyclic amine derivatives (5a-e and 6a-e). On the other hand, thiomorpholine analogue 5e had the strongest antagonistic behavior at D 2 Rs from all prepared final compounds. Interestingly, the final derivatives 5a-g, 6a-g exhibited very low cytotoxicity in MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay using Chinese hamster ovary cell lines (Table 1) so that the relatively lower affinity toward D 2 R is compensated by low toxicity allowing higher dosing. Table 1. Binding affinity of tested final compounds 5a-g and 6a-g at D 2 Rs and their cytotoxicity.

Molecular Modelling Studies
A molecular modelling study was undertaken to help explain differences in binding strength. Quantitative results are summarized in Table 2, the docking score (S) and the calculated relative binding energy (ddG) from the thermodynamic integration free energy calculation follow the same tendency as the measured affinities. Docking of USC-D301 and 5e revealed strongly differing binding poses, despite their high molecular similarity ( Figure 2). The positively charged nitrogen of all docked ligands was recognized by Asp114 via a strong ionic ammonium-carboxylate interaction. A pi-pi interaction with Phe390 was also apparent, with the isoxazole, 2,3-dichlorophenyl and 2-methoxyphenyl fragment of risperidone, aripiprazole, and USC-D301, respectively. Surprisingly, the lowest scoring docking pose of 5e did not overlap with USC-D301 (Figure 3). Owing to its smaller size and thiomorpholine cap, there is no aromatic group at that position that can undergo the stacking interaction with Phe390 or other nearby residues, likely leading to a strong penalty to the binding energy. Instead, the molecule adopts a flipped conformation that places the benzolactam fragment in this pocket, where the fused benzene ring is not positioned well to undergo stacking with Phe390, Trp386, and others. that places the benzolactam fragment in this pocket, where the fused benzene ring is not positioned well to undergo stacking with Phe390, Trp386, and others.  Surface color of pocket: green is hydrophobic, purple is polar, and red is exposed. For clarity, only residues that showed selective interactions in the docking screens are rendered, sidechains in bold.

Central Nervous System Availability Prediction and Study for Novel Compounds
Before the synthesis, we have calculated the so-called BBB score to predict the compound's CNS availability. Indeed, all the compounds displayed high values above 5.0 (5.2-5.4) which is indicative of their high potential to cross BBB. The prediction was then confirmed by the data from parallel artificial membrane permeation (PAMPA) assay pointing out their potential to cross the BBB by passive diffusion (5a-g and 6a-g Pe (× 10 −6 cm s −1 ) = 7.0-24) ( Table 3). The validation of PAMPA has been performed using standard compounds whose availability or unavailability was experimentally predicted in vitro and confirmed in vivo [53,82]. Table 3. Prediction of BBB barrier penetration of the studied compounds expressed as Pe (n = 3) and BBB score of final derivatives.

Central Nervous System Availability Prediction and Study for Novel Compounds
Before the synthesis, we have calculated the so-called BBB score to predict the compound's CNS availability. Indeed, all the compounds displayed high values above 5.0 (5.2-5.4) which is indicative of their high potential to cross BBB. The prediction was then confirmed by the data from parallel artificial membrane permeation (PAMPA) assay pointing out their potential to cross the BBB by passive diffusion (5a-g and 6a-g P e (× 10 −6 cm s −1 ) = 7.0-24) ( Table 3). The validation of PAMPA has been performed using standard compounds whose availability or unavailability was experimentally predicted in vitro and confirmed in vivo [53,82].

Conclusions
In summary, a series of 3,4-dihydroquinolin-2(1H)-one analogues, inspired by aripiprazol was designed and synthesized. The substitutions of the amine group revealed a negligible impact on D 2 R affinity. Although the binding affinities at D 2 Rs of new analogues are much weaker compared to aripiprazole, they are very close to the binding affinity, for instance, of memantine acting as N-methyl-D-aspartate receptor antagonist, a wellestablished drug for the treatment of Alzheimer's disease [83,84]. Out of these ligands, 5e possessed the highest D 2 R affinity, very low cytotoxicity profile, and the highest probability to cross the BBB. Molecular modeling simulation revealed completely different binding mode of 5e compared to USC-D301, which might be the culprit of the reduced affinity of 5e toward D 2 R. The subject of further investigation of these compounds will be to assess their affinity toward other members of the D 2 -like receptors family and other GPCRs, especially 5-HT 1A and 5-HT 2A Rs. Since aripiprazole has few of the typical adverse effects of other antipsychotics, such as extrapyramidal symptoms, hyperprolactinemia, weight gain, metabolic disorders, and sedation given to its unique biased activity and/or partial agonistic/antagonistic actions on D 2 Rs supplemented by the action on other dopamine receptor subtypes (mainly D 3 R) and 5-HT receptor subtypes (mainly 5-HT 1A , 5-HT 2A ) [54], the ongoing study must determine the functional affinity of newly developed compounds at these receptors to evaluate the real antipsychotic effect and clinical potential [85].