N-Substituted 3-Aminooxindoles and N-Propargyl Derivatives: Potential Biological Activities against Alzheimer's Disease

Results in Chemistry 6 (2023) 101032 Available online 8 July 2023 2211-7156/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). N-Substituted 3-Aminooxindoles and N-Propargyl Derivatives: Potential Biological Activities against Alzheimer’s Disease Tereza Hofmanova a , b , c , 1 , Carolina Marques a , 1 , Alfonso T. García-Sosa d , ´ Oscar L´ opez e , Luisa Leitzbach f , Elisabete P. Carreiro a , Aday Gonz´ alez-Bakker f , Adri´ an Puerta f , Holger Stark g , Jos´ e M. Padr´ on f , Jos´ e G. Fern´ andez-Bola˜ nos e , Anthony J. Burke a , b , h , i , j , * a LAQV-REQUIMTE, University of ´ Evora, Institute for Research and Advanced Studies, Rua Rom˜ ao Ramalho, 59, 7000-671 ´ Evora, Portugal b Chemistry and Biochemistry Department, School of Science and Technology, University of ´ Evora, Rua Rom˜ ao Ramalho 59, 7000-671 ´ Evora, Portugal c Department of Chemistry, Faculty of Science, University of Hradec Kr´ alov´ e, Rokitansk´ eho 62, 50003 Hradec Kr´ alov´ e, Czech Republic d Institute of Chemistry, University of Tartu, Ravila 14 A, Tartu 50411, Estonia e Departamento de Química Org´ anica, Facultad de Química, Universidad de Sevilla, Apartado 1203, E-41071 Seville, Spain f BioLab, Instituto Universitario de Bio-Org´ anica “Antonio Gonz´ alez” (IUBO-AG), Universidad de La Laguna, c/ Astrofísico Francisco S´ anchez 2, E-38206 La Laguna, Spain g Heinrich Heine University Düsseldorf, Institute of Pharmaceutical and Medicinal Chemistry. Universitaetsstr. 1, 40225 Duesseldorf, Germany h Faculty of Pharmacy, University of Coimbra, P´ olo das Ciˆ encias da Saúde, Azinhaga de Santa Comba, 3000-548 Coimbra, Portugal i Centro de Química de Coimbra, Institute of Molecular Sciences, Departamento de Química, Faculdade de Ciˆ encias e Tecnologia, Universidade de Coimbra, 3004-535 COIMBRA, Portugal j Center for Neurosciences and Cellular Biology (CNC), Polo I, Universidade de Coimbra Rua Larga Faculdade de Medicina, Polo I, 1◦andar, 3004–504, Coimbra, Portugal ARTICLE INFO Keywords: Oxindole Alzheimer’s disease Cholinesterase Mono-amine oxidase Molecular docking STD-NMR Continuous live cell imaging ABSTRACT The oxindole core is an important structural motif in many natural and synthetic substances with various biological activities including anticancer, antineurodegenerative, and antimicrobial properties. This report focuses on the synthesis and biological activity of a series of novel N-substituted 3-aminooxindoles and their assessment in cholinesterase (ChE) and monoamine oxidase (MAO) inhibition. With regard to MAO inhibition, a series of N- propargyl containing derivatives was synthesized and screened. Despite being weak inhibitors of MAO-A and MAO-B, the compounds were selective for butyrylcholinesterase (BuChE) over acetylcholinesterase (AChE). Most of them were strong inhibitors of BuChE with IC 50 ’ s of less than 1 µM, and one compound showed an IC 50 =27 nM. The mechanism of action of the inhibition was pin-pointed through molecular modeling, and was validated using saturation-transfer-difference (STD) NMR. Some of the compounds were screened for anti-oxidant properties, but showed no activity. The same compounds were screened in the neurodegenerative disease model cellline SH-SY5Y and although some were found to be non-cytotoxic, others were moderately cytotoxic. Continuous live cell imaging experiments showed that the compounds do not induce relevant cell damage and thus, the compounds might be interesting drug candidates for Alzheimer’s disease. Furthermore, the most active compounds showed excellent drug-likeness and pharmacological properties predicted using Swiss-ADME, and the pharmacokinetic simulations indicated that all these compounds cross the blood-brain-barrier. Introduction Alzheimer’s disease (AD) is a neurodegenerative disorder that impairs mental ability and interrupts neurocognitive function. This neuropathological condition is highlighted by neurodegeneration, neural loss, and the development of both neurofibrillary tangles and amyloid beta (Aβ) plaques [1]. Effective treatments for AD are inexistent, with no known disease- * Corresponding author at: LAQV-REQUIMTE, University of ´ Evora, Institute for Research and Advanced Studies, Rua Rom˜ ao Ramalho, 59, 7000-671 ´ Evora, Portugal. E-mail address: [email protected] (A.J. Burke). 1 Both authors have contributed equally to this work. Contents lists available at ScienceDirect Results in Chemistry journal homepage: www.sciencedirect.com/journal/results-in-chemistry https://doi.org/10.1016/j.rechem.2023.101032 Received 15 June 2023; Accepted 3 July 2023 Results in Chemistry 6 (2023) 101032 2 modifying solutions despite the best efforts by the scientific and medical communities. The few drugs available on the market show only symptomatic relief (Fig. 1), except for the new β-amyloid level lowering antibody known as lecanemab, developed by Eisai and Biogen, which was granted accelerated approval by the FDA in January 2023 [2,3]. The causes of AD are not fully understood. In the β-amyloid hypothesis, the presence of Aβ 40 aggregates and tau protein fibrils are considered to be the primary pathological source of AD [1]. However, there are strong indications that AD is a multifactorial ailment, that involves other factors like neuroinflammation, acetylcholine degradation (the cholinergic hypothesis), glial cell dysfunction, increased tau phosphorylation, deregulation of calcium metabolism, enhanced glycogen-synthase kinase (GSK)-3β activity, stimulation of cell death or neuronal apoptosis and oxidative stress [4]. Moreover, various drug candidates targeting these pathological processes were unsuccessful in preclinical or clinical trials, perhaps because of their low bioavailability, blood–brain-barrier penetration, and rapid drug excretion in such treatments. Therefore, the development of new small molecule disease-modifying therapies is essential. With respect to the cholinergic hypothesis (the most successful approach to date), the three prescribed drugs currently on the market include: donepezil, rivastigmine, and galantamine (the first two are selective acetylcholinesterase (AChE) inhibitors and the latter is a dual AChE/butyrylcholinesterase (BuChE) inhibitor). The oxindole core is a unit present in many natural products and bioactive compounds [5]; some examples include: nintedanib [6] (interstitial lung ailments like idiopathic pulmonary fibrosis (IPF) treatment), sunitinib (tyrosine kinase inhibitor, used to treat drug gastrointestinal stromal tumors and renal cell carcinoma) [7], SSR- 149414 (an orally active nonpeptide vasopressin-receptor antagonist) [8], and NITD609 (an antimalarial drug candidate) [9] (Fig. 2A–D). Moreover, oxindole-containing molecules have also been shown to be very good ChE inhibitors. One specific example includes the drug tenidap (Fig. 2E) which through a combination of cyclooxygenase inhibition and cytokine modulation, leads to a slowing of plaque formation and neuronal damage [10]. With regards to BuChE inhibition we have shown that 3-hydroxyoxindoles can be potent BuChE inhibitors (in the range 1–10 μ M) [12–14]. Also a number of potent 3-aminospiroox- indole BuChE inhibitors, including tenidap, have demonstrated very good BuChE inhibitions [11,15,16]. Moreover, we also developed selective BuChE inhibiting 2-oxindole-3-carboxamides (in all cases they showed low potency for AChE) (Fig. 3) [17]. Monoamine oxidase A and B (MAO-A and B) inhibition has also been a strategy for treating both AD and Parkinson’s disease (PD) [18]. These MAOs are FAD associated enzymes found on the mitochondrial outer membrane. Whilst MAO-A catalyzes the oxidative deamination of serotonin, adrenaline and noradrenaline, MAO-B is known to preferentially deaminate β-phenethylamine and benzylamine. Some of the most potent MAO inhibitors, like; rasagiline, selegiline and clorgyline, contain an N-propargyl unit, which has been considered to be the key pharmacophore that inhibits MAO via a covalent interaction with the FAD unit [19]. The introduction of a propargyl group on the 3-amino unit (and other oxindole scaffold positions) would provide interesting rasagiline analogues (see Scheme 3) that could be active against MAO-A and/or B. Rasagline (Scheme 3) is a well-known commercial irreversible MAO-B inhibitor. In this article we discuss our efforts in developing free and Boc (tertbutoxycarbonyl) protected 3-aminooxindole derivatives, transforming them into N-propargyl derivatives and studying their potency for both BuChE and MAO inhibition. Results and discussion Chemistry In 2016 we reported a novel synthetic methodology for the synthesis of a library of 3-amino-aryl-oxindole derivatives containing an N-methyl group in the isatin core (2a-k), (3a-k) and (4a-i), via a rhodiumcatalyzed arylation of N-Boc-protected ketimines (Scheme 1) [20] [the synthesized compounds described in this report were not biologically assayed]. It also became of interest to us, for comparative purposes, to prepare a series of N-benzyloxindole analogues (5a-h) and (6a-c) using the same reported reaction conditions and with the isatin-derived N- Boc-protected ketimines (1d) and (1e) (Scheme 1). This was partly due to the fact that the N-benzyl group is an important pharmacophoric unit present in donepezil (Fig. 1A) and is also known to be an important pharmacophore for MAO inhibition [18,21]. The overall yields were moderate to very good and no significant differences were observed in the isolated yields for N-Me and N-benzyl oxindole products (Scheme 1). Fig. 1. Approved drugs for AD treatment: donepezil (A), rivastigmine (B), galantamine (C), and tacrine (D) (due to chronic hepatotoxicity, this drug has been removed from the market). T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 3 Compounds (2c), (3c), (4a-b), (5c), and (6b-c) gave the best reaction yields (>80% isolated yield, Scheme 1). Attempts to synthesize the N- benzyl-5-bromo-N-Boc-protected ketimine were unsuccessful as only the N-Boc-protected ketimine (1d) was isolated, in 57% yield, proving that the C5-Br bond was cleaved during the reaction. The deprotection of the N-Boc group was fully demonstrated using the previously optimized reaction conditions [20], using trifluoroacetic acid (TFA) in CH 2 Cl 2 , at room temperature (Scheme 2). A new family of free 3-amino-aryl-oxindole derivatives (7) (Scheme 2) was obtained in generally moderate yields (10 to >99% yield) and also biologically assayed. The best yields were recorded for compounds (7l) and (7n) with >99% and 70% yield, respectively (Scheme 2). Oxalyl chloride was also studied, but was found to be too harsh, although the isolation of the free amine was easier than using TFA and the product obtained in acceptable purity. These deprotected compounds (Scheme 2) were then transformed into a variety of N-propargylated derivatives, using the method recently developed in our laboratory [21]. Seven new 3-amino-aryl-oxindole hybrids with propargyl units attached at position 3 (11a-d) and position 5 (10a-c) of the isatinoxindole framework were successfully obtained (Scheme 3). The copper-catalyzed route to primary aromatic amines from aromatic halides using sodium azide was one of the reactions used to afford compounds (9a) and (9b). Biological evaluation In vitro cholinesterase inhibitory activity Tables 1 and 2 show the screening results of our non-propargylated compounds against eeAChE and eqBuChE. Ketimine (1d) was found to be a potent eeAChE inhibitor and even more so, for eqBuChE (IC 50 =11 and 2.30 µM, respectively; Table 1, entry 1). The N-methyl substituted 3- aminooxindoles can be split into four subgroups - the subgroup with a non-substituted oxindole skeleton (2a)-(2k) (Group MetA), a 5-Br substituent (3a)-(3k) (Group MetB), a 5-Me substituent (4a)-(4i) (Group MetC) and the free amine analogues (7 l)-(7o) (Group MetD). In the case of the N-benzyl analogues, they were divided into three groups; unsubstituted (5a)-(5h) (Group BnA), with the 5-Me substituent (6a)-(6c) (Group BnB) and the free-amines (7a), (7i), (7j), (7m), (7n) and (7o) (Group BnC). All the compounds, except (1d) showed very low inhibitory potencies against eeAChE with IC 50 values >50 µM. Interestingly, they Fig. 2. Examples of biologically active oxindoles: nintedanib (A) [6], sunitinib (B) [7], SSR-149415 (C) [8], NITD609 (D) [9] and tenidap [10]. Fig. 3. Selected 2-oxindole-3-carboxamides as potential inhibitors of ChEs [17]. T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 4 were selective for eqBuChE. In the subgroup (Group MetA), (2a-2j), only (2c) (Table 1, entry 4) could reach submicromolar inhibition concentrations (0.29 µM) for BuChE. The others were either devoid of relevant activity at the highest tested concentration (IC 50 >50 µM) or showed moderate inhibition (IC 50 =22–50 µM). In the case of (Group MetB) (3a-3 k) both compounds (3e and 3f) gave inhibition levels lower than 8.0 µM and (3i) gave an IC 50 of 11.0 µM against BuChE. In the case of (Group MetC), the observed low IC 50 values indicated the importance of the 5-Me group over the 5-Br, for example, compounds (4c), (4e), (4f), and (4i) gave values of less than 9.0 µM. In the case of the N-benzyl Scheme 1. Rh-Catalyzed synthesis of N-Boc-3-amino-3-aryl-oxindole derivatives (2)-(6) from isatin-derived N-Boc-protected ketimines (1). T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 5 substituted molecules, (Group BnA) and (Group BnB) (5a)-(5 h) and (6a)-(6c), even more potent inhibitions were observed, with all values below 4.0 µM and, and remarkably, compound (5c) was potent down to 27.0 nM, roughly 204-fold more potent than the drug galantamine, used herein as a positive control. This latter result was most probably due to the size and the aromatic nature of the compound allowing it to interact favorably with aromatic units within the BuChE active site like: Trp82, Trp231 and Phe329. (Group BnC) also furnished some very good levels of inhibition of not>18 µM and most of them had values lower than 10 µM. The free amine analogues (7a), (7i), (7j), (7m), (7n) and (7o) showed potencies of between 1.0 and 18.0 µM (Table 1). With the exception of (7i), the N-methyl analogues were the most potent ones in this class. Moreover, the absence of the Boc group did not make a substantial difference to the inhibition. It is noteworthy to say that for BuChE inhibition, most of these compounds were more potent than the benchmark, galantamine [22]. With regard to the mechanism of inhibition, we looked at two of our Scheme 2. Synthesis of the free 3-amino-aryl-oxindole derivatives (7). L (R) Scheme 3. Synthesis of the propargylated amino-oxindole hybrids (10a-c) and (11a-d). NBS: N-bromosuccinimide. T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 6 best BuChE inhibitors, namely, (2c) and (5c), and we found that they were both efficient mixed inhibitors (Table 2) (also see the Supporting information Figs. s1 and s2; Cornish Bowden Plots and s3-s41 for the IC 50 plots for the inhibition of BuChE). The most essential features for good BuChE inhibition are highlighted in Fig. 4 above. MAO inhibition studies As described above, besides being a target for Parkinson’s disease, MAO is also a target for AD, and for this reason, considering the importance of the N-propargylamine pharmacophore, we developed a series of N-propargyl-3-amino-3-aryl-oxindole analogues, (10a-c) and (11a-d) including, some Boc protected non-propargylated ((5c)), and free amine derivatives (7c), (7l) and (7m) (Table 3). The best inhibition values were obtained against MAO-B using (5c) and (11a) (19.7% and 20.8% inhibition). Unfortunately, for the other derivatives, no relevant MAO inhibition was observed. In silico studies The possible binding of compounds to ChEs was investigated using in silico docking protocols using the X ray-determined protein crystal structures of human AChE and BuChE [23] ((Table 4). Nearly all of the compounds studied showed very strong predicted ligand binding efficiencies, within values determined previously [24,25]. Some of the results shown in Table 1, corroborated the results from the screening studies, for example; compound (2c) which furnished an IC 50 of 0.29 μ M (Table 1) gave a docking score of −9.82 kcal/mol and a glide ligand efficiency of −0.339. Compound (5c) which exhibited an IC 50 of 27 nM (Table 1) gave a docking score of −8.02 kcal/mol and a glide ligand efficiency of −0.229 for the (R)-enantiomer. Both (S)- and (R)-(4i) gave scores of −9.72 and −10.54 kcal/mol (with ligand efficiencies of −0.324 and −0.351, respectively) when the IC 50 of the racemate was determined to be 3.4 μ M (Table 1). Also (3i) gave a value Table 1 IC 50 values (µM) for eeAChE and eqBuChE of selected compounds. Compound eeAChE (µM) a eqBuChE (µM) a Compound eeAChE (µM) a eqBuChE (µM) a (1d) 11 ±2 2.3 ±0.0 (4d) >50 23.0 ± 1.0 (2a) >50 >50 (4e) 8.9 ±0.5 (2b) >50 (4f) 5.9 ±0.8 (2c) 0.29 ± 0.04 (4 g) 27.0 ± 1.0 (2d) 22.0 ± 6.0 (4 h) >50 (2e) >50 (4i) 3.4 ±0.9 (2f) >50 (5a) 1.1 ±0.0 (2 g) >50 (5b) 1.3 ±0.0 (2 h) 38.0 ± 4.0 (5c) 0.027 ± 0.000 (2i) 50.0 ± 0.0 (5d) 3.9 ±0.6 (2j) >50 (5e) 0.89 ± 0.16 (3a) >50 (5f) 3.5 ±0.3 (3b) 18.0 ± 1.0 (5 g) 0.91 ± 0.14 (3c) 44.0 ± 1.0 (5 h) 0.62 ± 0.04 (3d) >50 (6a) 2.2 ±0.1 (3e) 7.3 ±1.1 (6b) 0.86 ± 0.03 (3f) 7.7 ±0.6 (6c) 1.4 ±0.1 (3 g) 39.0 ± 2.0 (7a) 12.0 ± 1.0 (3 h) >50 (7i) 18.0 ± 1.0 (3i) 11.0 ± 2.0 (7j) 1.0 ±0.1 (3j) 23.0 ± 1.0 (7 m) 7.4 ±0.7 (3 k) 20.0 ± 0.0 (7n) 2.4 ±0.0 (4a) 24.0 ± 3.0 (7o) 2.7 ±0.4 (4b) 24.0 ± 1.0 Galantamine b 1.3 ± 0.1 5.5 ±0.4 (4c) 1.9 ±0.1 a Average of 5–9 different inhibitor concentrations and data obtained in duplicate, expressed as the mean ±SD. Substrate concentrations were fixed at 121 μ M for AChE and 112 μ M for BuChE. Data are expressed as mean ±standard deviation. b Positive control. Table 2 Kinetic parameters a for eqBuChE inhibition for the best compounds. Compound K ia (µM) K ib (µM) Inhibition mode (2c) 0.82 ±0.10 3.5 ±0.5 Mixed (5c) 0.019 ±0.003 0.065 ±0.005 Mixed a 5 different substrate concentrations were used (¼ Km −4 Km), and 4 different inhibitor concentrations. Cornish-Bowen plots were used for calculating the kinetic parameters. Fig. 4. Summarizes the most desirable features for good BuChE inhibition (both 1- and 2-naphthyl units provide excellent potencies). Table 3 MAO screening data for selected 3-amino-aryl-oxindole analogues. Compound a Inhibition [%] at 1 μ M ±SD b MAO-A MAO-B 5c 6.7 ±4.3 19.7 ±6.5 7c 10.3 ±4.7 12.9 ±8.3 7 l 6.8 ±5.2 13.1 ±11.4 7 m 6.4 ±7.3 11.9 ±11.2 10a 8.9 ±3.6 9.8 ±4.5 10b 11.1 ±2.8 18.0 ±11.6 10c 9.4 ±3.4 18.3 ±6.6 11a 9.1 ±5.6 20.8 ±12.3 11b 12.1 ±11.6 17.6 ±4.4 11c 9.4 ±11.1 8.1 ±4.8 11d 6.8 ±1.8 11.9 ±4.6 No inhibitor 0.0 ±2.5 0.0 ±3.8 Clorgyline 101.6 ±1.3 31.4 ±1.8 Safinamide 7.2 ±3.4 96.9 ±1.0 a Compounds dissolved in DMSO (10 mM, 1 mM). b MAO-A and MAO-B inhibition were calculated as percentages related to control at a test concentration of 1 μ M and given as mean ±standard deviation of two independent experiments in triplicate, using clorgyline and safinamide as a positive control. T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 7 of 11.0 μ M and a docking score of −9.78 kcal/mol for the (R)- enantiomer. Since both enantiomers were screened, it was of interest to determine if one enantiomer was favored over the other. In the case of the (2) series, it was generally the (S) enantiomer that was preferred, however, in the case of the other series, there was no common trend. It should also be noted, that gratifyingly there was a correlation between the IC 50 values reported in the bioassay studies and these in silico docking results. In the case of compounds (5c) and (7c), the simulations for both enantiopodes predicted good binding to the hBChE binding site through favourable π stacking of the oxindole ring with Trp82 (Fig. 5), in addition to π stacking with Trp231 and Phe329 in the case of (R)-(5c), and His438 for (S)-(7c). These are in addition to general van der Waals and polar interactions between the ligands and the protein, such as the hydrogen bond donated from Gly117 to the 3-carbonyl of the oxindole ring of (S)-(7c). Saturation transfer difference (STD)-NMR studies Over the last years, the STD-NMR method has become an indispensable technique to validate docking studies (in silico) [13,14,22]. To gain better insight into the interactions of our 3-aminooxindoles with eqBuChE, we executed STD-NMR experiments with our most potent oxindole ligands, namely (2c) and (5c). These STD-NMR studies showed why (5c) is a stronger inhibitor of eqBuChE than (2c). STD-NMR of ligand (2c). Upon analyzing the results obtained in the STD-NMR experiment of (2c) (Fig. 6) with eqBuChE, (the graph of the STD amplification factor as a function of saturation time is shown in the Supporting Information (Fig. s42)), it was observed that all the hydrogens interact with the eqBuChE binding pocket. Nevertheless, the H1 and H13 hydrogens from the methyl and t-butyl groups showed weaker interactions with the amino acid residues (with attenuations of 55% and 29%, respectively). In fact, these results are in agreement with the docking studies, in which the role of the t-butoxide group interacts weakly with the enzyme. As can be observed from the epitope intensity table shown in Fig. 6, the aromatic hydrogens are the most affected by the active site residues, having attenuations in the range 81–100%. This was indeed corroborated by the docking study (Fig. 5), showing very good placement of the aromatic units within the binding pocket. STD-NMR of ligand (5c). The STD-NMR results for (5c) with eqBuChE are shown in Fig. 7 (the graph of the STD amplification factor as a function of saturation time is shown in Fig. s43; SI). The signals for the non-aromatic protons H2 and H7 showed a weaker interaction with the enzyme, which are completely in agreement with docking studies (Fig. 5). As can be seen from the epitope intensity table shown in Fig. 7, the aromatic protons are the most affected, with attenuations in the range 87–100%. This was in fact supported by the docking study (Fig. 5), showing good proximity of the Trp231 and Phe329 units to the phenyl ring of the benzyl group (97% attenuation). The oxindole hydrogens, H3-H6 showed high STD attenuation values (95.7 – 99.6%) indicating that they should be very close to the active site residues of the protein, and most probably to Trp82 as suggested by our docking studies (Fig. 5). The naphthyl protons also showed high proximity to the amino acid residues of the enzyme, showing high STD attenuation values, more specifically H10 and H11 with values of 100%. The docking studies indicated the proximity of these hydrogens to the polar amino acid residues of the enzyme. On comparing the STD-NMR results for both inhibitors, we confirm that (5c) is a stronger binder than (2c), which is principally due to the benzyl group which establishes a stronger interaction with aminoacid residues of the binding site, which is corroborated by our docking studies. Calculated physicochemical properties For some of the chosen compounds we determined their physicochemical properties, using SwissADME [26]. All of the calculated results were within the limits defined for Lipinski’s rules, except in the case of (5c), (5d), and (11b), which were slightly outside the defined limits. Also, encouragingly, none of these compounds were identified as Pan- Table 4 Calculated docking scores and efficiency indices for compounds with hBuChE. Compound Docking Score (kcal/ mol) Glide Ligand Efficiency Glide Ligand Efficiency SA Glide Ligand Efficiency Ln S-(2a) −9.49 −0.365 −1.082 −2.229 R-(2a) −6.89 −0.265 −0.785 −1.618 S-(2b) −6.19 −0.238 −0.706 −1.454 R-(2b) −5.80 −0.223 −0.660 −1.361 S-(2c) −6.73 −0.232 −0.713 −1.540 R-(2c) −9.82 −0.339 −1.040 −2.248 S-(2d) −8.47 −0.292 −0.897 −1.938 R-(2d) −8.34 −0.288 −0.884 −1.910 S-(2e) −6.75 −0.241 −0.732 −1.558 R-(2e) −6.10 −0.218 −0.661 −1.408 S-(2f) −6.50 −0.241 −0.722 −1.513 R-(2f) −6.47 −0.240 −0.719 −1.505 S-(2 g) −6.35 −0.244 −0.723 −1.491 R-(2 g) −6.38 −0.245 −0.727 −1.498 S-(2 h) −8.66 −0.333 −0.987 −2.035 R-(2 h) −6.83 −0.263 −0.778 −1.603 S-(2i) −6.97 −0.268 −0.794 −1.636 R-(2i) −6.90 −0.265 −0.786 −1.619 S-(2j) −5.55 −0.205 −0.616 −1.291 R-(2j) −6.50 −0.241 −0.723 −1.514 S-(3a) −8.33 −0.278 −0.863 −1.893 R-(3a) −9.51 −0.317 −0.985 −2.160 S-(3b) −6.98 −0.258 −0.775 −1.624 R-(3b) −9.40 −0.348 −1.045 −2.189 S-(3c) −7.14 −0.275 −0.814 −1.677 R-(3c) −7.46 −0.287 −0.850 −1.751 S-(3d) −7.28 −0.270 −0.809 −1.694 R-(3d) −8.15 −0.302 −0.906 −1.898 S-(3e) −9.75 −0.325 −1.010 −2.216 R-(3e) −7.57 −0.252 −0.784 −1.720 S-(3f) −4.77 −0.170 −0.517 −1.101 R-(3f) −5.40 −0.193 −0.585 −1.245 S-(3 g) −6.21 −0.230 −0.690 −1.446 R-(3 g) −5.65 −0.209 −0.628 −1.316 S-(3 h) −7.47 −0.277 −0.830 −1.739 R-(3 h) −9.10 −0.337 −1.011 −2.119 S-(3i) −6.99 −0.259 −0.777 −1.628 R-(3i) −9.78 −0.362 −1.087 −2.278 S-(3 k) −4.38 −0.156 −0.475 −1.011 R-(3 k) −8.58 −0.307 −0.931 −1.982 S-(4a) −6.65 −0.237 −0.721 −1.535 R-(4a) −5.32 −0.190 −0.577 −1.229 R-(4a) −5.32 −0.190 −0.577 −1.229 R-(4b) −7.02 −0.260 −0.780 −1.634 S-(4c) −7.60 −0.253 −0.787 −1.726 R-(4c) −7.35 −0.245 −0.761 −1.670 S-(4d) −6.67 −0.256 −0.760 −1.566 R-(4d) −6.22 −0.239 −0.709 −1.461 S-(4e) −6.87 −0.245 −0.745 −1.585 R-(4e) −6.12 −0.219 −0.664 −1.413 S-(4f) −6.81 −0.252 −0.756 −1.585 R-(4f) −7.02 −0.260 −0.780 −1.633 S-(4 g) −6.96 −0.240 −0.738 −1.594 R-(4 g) −7.37 −0.254 −0.780 −1.687 S-(4 h) −7.11 −0.263 −0.790 −1.654 R-(4 h) −6.32 −0.234 −0.702 −1.471 S-(4i) −9.72 −0.324 −1.006 −2.207 R-(4i) −10.54 −0.351 −1.092 −2.395 S-(5c) −7.82 −0.223 −0.730 −1.716 R-(5c) −8.02 −0.229 −0.749 −1.760 S-(7c) −9.54 −0.341 −1.034 −2.201 R-(7c) −8.80 −0.314 −0.954 −2.030 Compounds with docking scores <-8.0 kcal/mol are shaded in light brown. Not all the compounds screened in silico were biologically assayed. T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 8 Assay Interference Compounds (PAINS) or in other words, suggesting no interactions with multiple targets with a possible negative influence on any screening experiment (Table 5). All the synthesized compounds were predicted to be amenable to gastrointestinal absorption, as well as predicted to cross the blood–brain barrier (BBB). Most of the products were predicted to be moderately soluble in water, and a few (5b), (5c), Fig. 5. 2D schematic representation of predicted binding poses in the binding site of hBuChE with: A) (R)-(5c), B) (R)-(7c), C) (S)-(5c), and D) (S)-(7c). Note: the protonated form of (7c) was considered. Fig. 6. Top: Reference 1 H NMR spectrum of compound (2c) (5.5 mM) with eqBuChE enzyme (1.3 μ M). Bottom: the corresponding STD-NMR spectrum with 4 s of saturation. The H6 proton was set to 100%. The NMR spectra were recorded at 25 ◦C. T. Hofmanova et al. Results in Chemistry 6 (2023) 101032 9 (5d), (6a), and (6b) poorly soluble [27,28] (Table 5). Antiproliferative tests in SH-SY5Y human neuroblastoma cell line The in vitro antiproliferative activity of the six most active compounds resulting from the BuChE tests (Table 1) was studied against the Alzheimer’s disease model human neuroblastoma cell line, SH-SY5Y line using our implementation of the NCI protocol [29]. Thus, compounds (2c), (4c), (4i), (5c), (7n) and-(7o) were screened against the neuroblastoma cell line SH-SY5Y. This serves as a model for neurodegenerative disorders and is widely used in neurotoxicity studies [29]. The results of the 50% growth inhibition values after 48 h of exposure (GI 50 ) are given in Table 6. The results show that compounds (5c), (7n) and (7o) were inactive (GI 50 >100 μ M) against this neuroblastoma cell-line. In contrast, compounds (2c), (4c) and (4i) were cytotoxic, with GI 50 values in the range 19–46 μ M. Label-free continuous live cell imaging Live imaging enables one to observe cytostatic effects [30] or cell death of treated cells in real time and to identify several apoptotic hallmarks [32–35]. For characterization of the phenotypic effects of BuChE inhibitors on neuroblastoma cells, the evolution of the culture was monitored every 5 min using label-free holotomographic 3D microscopy. SH-SY5Y cells were exposed to compounds (2c), (5c) and (7o) for 20 h at 100 μ M (Supplementary material, videos S1–4). The selection was done considering the BuChE tests (Table 1) and the antiproliferative activity against this neuroblastoma cell line (Table 6). Untreated and treated cells moved freely and divided, indicating that the exposure to the compounds apparently did not affect cell growth and motility. A more precise analysis after segmentation of the single cells d6 Fig. 7. TOP: Reference 1 H NMR spectrum of compound (5c) (2.3 mM) with eqBuChE enzyme (1.9 μ M). Bottom: the corresponding STD-NMR spectrum with 4 s of saturation (bottom). The H3 proton was set to 100%. The NMR spectra were recorded at 25 ◦C. Table 5 Calculated molecular properties for selected compounds. Compound Mw (g/mol) MLOGP [28] LogS [27] # H-bond acceptors # H-bond donors TPSA PAINS # alerts (1d) 336.38 2.97 −4.77 4 0 58.97 0 (2c) 388.46 3.32 −5.47 3 1 58.64 0 (4c) 402.49 3.53 −5.86 3 1 58.64 0 (4i) 402.49 3.53 −5.86 3 1 58.64 0 (5a) 414.50 3.70 −5.73 3 1 58.64 0 (5b) 428.52 3.90 −6.11 3 1 58.64 0 (5c) 464.55 4.31 −7.03 3 1 58.64 0 (5d) 448.94 4.17 −6.38 3 1 58.64 0 (5e) 444.52 3.36 −5.90 4 1 67.87 0 (5f) 432.49 4.07 −5.84 4 1 58.64 0 (6a) 428.52 3.90 −6.11 3 1 58.64 0 (6b) 442.55 4.10 −6.49 3 1 58.64 0 (7j) 342.43 3.61 −4.52 2 1 46.33 0 (7n) 367.24 3.37 −4.23 2 1 46.33 0 (7o) 302.37 2.99 −3.89 2 1 46.33 0 (10a) 429.51 3.23 −4.61 3 1 61.88 0 (11a) 352.43 3.75 −4.12 2 1 32.34 0 (11b) 402.49 4.38 −5.42 2 1 32.34 0 MW =molecular weight; # H-bond =number of hydrogen bond; TPSA =topological surface area (Å 2 ). Table 6 Antiproliferative activity (GI 50 ) against the SH-SY5Y human neuroblastoma cell line. a. Compound SH-SY5Y Cell line GI 50 values (2c) 37 ±7.0 (4c) 19 ±5.2 (4i) 21 ±5.1 (5c) >100 (7n) >100 (o) >100 CDDP 14 ±2.4 a GI 50 values are given in μ M. Standard deviation was calculated from two to four independent experiments. Cisplatin (CDDP) was used as a reference cytotoxic drug. T. Hofmanova et al.