Serotonin 5-HT6 Receptor Ligands and Butyrylcholinesterase Inhibitors Displaying Antioxidant Activity—Design, Synthesis and Biological Evaluation of Multifunctional Agents against Alzheimer’s Disease

Neurodegeneration leading to Alzheimer’s disease results from a complex interplay of a variety of processes including misfolding and aggregation of amyloid beta and tau proteins, neuroinflammation or oxidative stress. Therefore, to address more than one of these, drug discovery programmes focus on the development of multifunctional ligands, preferably with disease-modifying and symptoms-reducing potential. Following this idea, herein we present the design and synthesis of multifunctional ligands and biological evaluation of their 5-HT6 receptor affinity (radioligand binding assay), cholinesterase inhibitory activity (spectroscopic Ellman’s assay), antioxidant activity (ABTS assay) and metal-chelating properties, as well as a preliminary ADMET properties evaluation. Based on the results we selected compound 14 as a well-balanced and potent 5-HT6 receptor ligand (Ki = 22 nM) and human BuChE inhibitor (IC50 = 16 nM) with antioxidant potential expressed as a reduction of ABTS radicals by 35% (150 μM). The study also revealed additional metal-chelating properties of compounds 15 and 18. The presented compounds modulating Alzheimer’s disease-related processes might be further developed as multifunctional ligands against the disease.


Introduction
Alzheimer's disease (AD) is the most commonly diagnosed form of dementia with a significant impact on public health as well as the economy. According to the current data, more than 55 million people are living with dementia nowadays, and this number will almost triple by 2050 as a result of the demographic ageing phenomenon [1]. Patients suffering from AD manifest a gradual deterioration of cognitive functions and memory loss as well as accompanying behavioural impairments such as depression, psychosis, or anxiety. These symptoms arise from the emerging neurodegenerative changes within different neuron populations, especially cholinergic, serotonergic, and GABAergic. Among the factors that contribute to the development of AD, abnormal protein aggregates-amyloid-beta (Aβ) plaques and neurofibrillary tangles (NFTs) [2], oxidative stress, metal dyshomeostasis and inflammatory processes are of the utmost importance [3,4]. Their modulation may bring disease-modifying effects and can lead to effective therapeutics against AD.
Currently available first-line anti-AD drugs are acetyl-(AChE) and/or butyrylcholinesterase (BuChE) inhibitors that enhance disrupted cholinergic neurotransmission, temporarily relieving cognitive symptoms and slowing the disease progression [5]. This symptomrelieving function is essential in the treatment of patients with developed AD manifestations [6]. Therefore, many AD drug discovery programmes are based on the search for multifunctional ligands combining AChE/BuChE inhibitory activity with diseasemodifying effects [7][8][9][10]. Although AChE initially appeared to be the key enzyme in acetylcholine metabolism, over the years it has become apparent that in AD BuChE takes over this function and its increased expression, especially in the hippocampus and temporal cortex, contributes largely to the development of the disease [11,12]. Not only does it decrease cholinergic neurotransmission but it also contributes to Aβ accumulation and aggregation, as shown in the 5XFAD/BChE-KO mice model (rapidly developing severe amyloid pathology BuChE-knockouts) where a fibrillar Aβ is significantly reduced [13].
A rather modest effect of cholinesterase inhibitors has been enhanced by the simultaneous use of 5-HT 6 receptor (5-HT 6 R) antagonists. Idalopirdine and intepirdine demonstrated positive effects on cognition in phase II clinical trials as an adjunct to cholinesterase inhibitors in patients with moderate Alzheimer's dementia [14,15]. 5-HT 6 receptors are mainly distributed on neurons in the brain structures responsible for memory and learningstriatum, hippocampus and cerebral cortex-and are linked with multiple neurotransmitter systems including cholinergic, glutamatergic and GABAergic [16,17]. Its pharmacological blockade was shown to enhance neurotransmission and exert anxiolytic and antidepressant activity, thus alleviating behavioural and psychological symptoms in dementia [18,19]. Additionally, 5-HT 6 R antagonists may also have a disease-modulating effect by their influence on the reduction of Aβ (demonstrated in retrospective and prospective studies with SSRIs) [20] or modulation of synaptic plasticity [21], neuronal hyperexcitability and regulation of synaptic remodelling [22][23][24].
Oxidative stress plays a significant role in AD and other neurodegenerative diseases. It results from a disturbed balance between the production and accumulation of reactive oxygen species (ROS), leading to the progressive destruction of lipids, nucleic acids and proteins including enzymes involved in glucose metabolism [25]. Glucose dysmetabolism, energy deficiencies and resulting ions' gradient disturbances lead to disrupted production and propagation of action potentials, and accumulation of free Ca 2+ ions in cells contributing to neurons' dysfunction and death [26]. Several mechanisms contribute to the development of oxidative stress, among them mitochondrial dysfunction, overexpression of oxidase enzymes, and imbalance of redox-active transition metal homeostasis [27,28]. A mounting body of evidence points to the crucial role of iron overload in the pathomechanism of AD. Increased concentrations of Fe 2+ , but also Cu 2+ , Al 3+ and Zn 2+ ions were observed post-mortem in AD patients' brain samples and also in vivo [28][29][30][31] in amyloid plaques and the surrounding tissues. Increased levels of iron ions have been reported to have a multidirectional influence on the pathomechanism of AD by the generation of ROS, promotion of Aβ aggregation to oligomers and plaques, and aggregation of tau protein into neurofibrillary tangles [32]. Recently, this has received a lot of attention due to the definition of ferroptosis-a phenomenon of nonapoptotic programmed cell death driven by lethal lipid peroxidation resulting from an imbalance in redox homeostasis [33][34][35][36][37].
The complexity of AD etiopathogenesis and the variety of contributing factors makes it difficult to select biological targets whose modulation would bring beneficial and effective therapeutic results. However, it seems obvious that effective therapy of the disease needs to address the modulation of several factors involved in the development of the disease. In our research, we focused on the development of multifunctional ligands to address both the crucial processes involved in the development of AD and the symptoms resulting from neurodegeneration: antioxidant and metal-chelating activity combined with inhibition of cholinesterases and antagonism of the 5-HT 6 receptors.

Design
Following the idea of multifunctional ligands, we continue our research on 5-HT 6 R antagonists and cholinesterase inhibitors targeting processes underlying the pathomechanism of AD [38][39][40][41][42]. Previously, we have reported a series of indole-based compounds with antioxidant activity [42] and in the ongoing project, we selected three 2-aminoethoxy-substituted indole derivatives known for their 5-HT 6 R antagonistic properties-1-(phenylsulfonyl)-1H-indole, 1-(phenylsulfonyl)indoline and 1-benzyl-1,3-dihydro-2H-benzo[d]imidazol-2one [40,43]-and connected them with the AChE-aiming 1-benzylpiperidine moiety via aliphatic linkers of different lengths ( Figure 1). The 1-benzylpiperidine fragment is derived from the marketed acetylcholinesterase inhibitor donepezil and plays a crucial role in the interaction pattern of this drug with the active site of AChE [44,45]. In addition to AChE inhibitory activity, donepezil displayed antioxidant activity in vivo by reducing the brain concentrations of malondialdehyde in scopolamine and lipopolysaccharide models [46]. disease needs to address the modulation of several factors involved in the development of the disease. In our research, we focused on the development of multifunctional ligands to address both the crucial processes involved in the development of AD and the symptoms resulting from neurodegeneration: antioxidant and metal-chelating activity combined with inhibition of cholinesterases and antagonism of the 5-HT6 receptors.

Design
Following the idea of multifunctional ligands, we continue our research on 5-HT6R antagonists and cholinesterase inhibitors targeting processes underlying the pathomechanism of AD [38][39][40][41][42]. Previously, we have reported a series of indole-based compounds with antioxidant activity [42] and in the ongoing project, we selected three 2aminoethoxy-substituted indole derivatives known for their 5-HT6R antagonistic properties-1-(phenylsulfonyl)-1H-indole, 1-(phenylsulfonyl)indoline and 1-benzyl-1,3dihydro-2H-benzo[d]imidazol-2-one [40,43]-and connected them with the AChE-aiming 1-benzylpiperidine moiety via aliphatic linkers of different lengths (Figure 1). The 1benzylpiperidine fragment is derived from the marketed acetylcholinesterase inhibitor donepezil and plays a crucial role in the interaction pattern of this drug with the active site of AChE [44,45]. In addition to AChE inhibitory activity, donepezil displayed antioxidant activity in vivo by reducing the brain concentrations of malondialdehyde in scopolamine and lipopolysaccharide models [46].

Synthesis
Key intermediates for the synthesis of the designed compounds were (1benzylpiperidin-4-yl)alkyl aldehydes (7 and 8) and aryloxyethanamines (9-11). Aldehydes 7 and 8 were prepared according to a previously published method presented in Scheme 1 [47]. In this procedure, the Horner-Westworth-Emmons reaction between commercially available ketone 1 or aldehyde 2 and diethyl (cyanomethyl) phosphonate was used for the synthesis of acrylonitriles 3 and 4. The olefins' double bond was reduced

Synthesis
Key intermediates for the synthesis of the designed compounds were (1-benzylpiperidin-4-yl)alkyl aldehydes (7 and 8) and aryloxyethanamines (9)(10)(11). Aldehydes 7 and 8 were prepared according to a previously published method presented in Scheme 1 [47]. In this procedure, the Horner-Westworth-Emmons reaction between commercially available ketone 1 or aldehyde 2 and diethyl (cyanomethyl) phosphonate was used for the synthesis of acrylonitriles 3 and 4. The olefins' double bond was reduced with magnesium and I 2(cat.) in MeOH and the obtained alkanenitriles 5 and 6 were further reduced by DIBAL-H and hydrolysed to give the aldehydes 7 and 8 (Scheme 1).
We prepared aryloxyethanamines 9-11 according to the previously described procedures [40,48]. The final compounds 12-19 were obtained in the reductive amination reaction between the aldehydes 2 (commercially available), 7 or 8 and appropriate amines (9)(10)(11) in the presence of NaBH3CN as the reducing agent in the 1:1 mixture of methanol and THF (Scheme 2).

5-HT6R binding affinity and cholinesterase (AChE/BuChE) inhibitory activity.
We evaluated the pharmacological properties of the final compounds 12-19 in vitro according to the previously described, well-established protocols. Their affinities for recombinant human 5-HT6 receptors (h5-HT6R) were tested in a radioligand binding assay and the inhibitory potencies against human recombinant AChE (hAChE) and BuChE isolated from human plasma (hBuChE) in the spectrophotometric Ellman's assay [38,49]. Table 1 lists the potencies expressed as Ki and IC50 values determined for the compounds and references: donepezil, tacrine and mianserin.

Biological Evaluation, SAR Analysis
5-HT 6 R binding affinity and cholinesterase (AChE/BuChE) inhibitory activity. We evaluated the pharmacological properties of the final compounds 12-19 in vitro according to the previously described, well-established protocols. Their affinities for recombinant human 5-HT 6 receptors (h5-HT 6 R) were tested in a radioligand binding assay and the inhibitory potencies against human recombinant AChE (hAChE) and BuChE isolated from human plasma (hBuChE) in the spectrophotometric Ellman's assay [38,49]. Table 1 lists the potencies expressed as K i and IC 50 values determined for the compounds and references: donepezil, tacrine and mianserin. All the compounds displayed affinities for 5-HT 6 receptors ranging from 17 nM to 598 nM. Molecular modelling studies on compound 14 disclosed that it adopts elongated conformation and a binding mode which is in line with the currently approved paradigm based on the published simulation on 5-HT 6 R homology models [50]. The interactions conserved for monoamine GPCRs-Asp3.32 salt bridge/H-bond and Phe6.52 CH-π stacking are the main anchoring forces. The orthosteric binding pocket interactions: CH-π stacking Phe5.38, Asn6.55 are unique for subtype 6 serotonin receptors. The salt bridge between the protonated amine and Asp7.36 has been described for tacrine-related 5-HT 6 R inhibitors ( Figure 2) [38]. As we observed in our previous studies [38,40], the 1-(phenylsulfonyl)-1Hindole fragment, as in compounds 12-14, ensures the highest potencies (17-22 nM).

Mianserin
0.056 ± 0.012 -a Ki values are expressed as the mean ± standard error of the mean (SEM) of at least three experiments. b IC50 values are expressed as the mean ± standard error of the mean (SEM) of at least three experiments. c % of inhibition of the enzymes observed at 10 μM concentration of 17. h5-HT6R-recombinant human 5-HT6 receptors, hAChE-human recombinant AChE, hBuChE-human recombinant BuChE.
All the compounds displayed affinities for 5-HT6 receptors ranging from 17 nM to 598 nM. Molecular modelling studies on compound 14 disclosed that it adopts elongated conformation and a binding mode which is in line with the currently approved paradigm based on the published simulation on 5-HT6R homology models [50]. The interactions conserved for monoamine GPCRs-Asp3.32 salt bridge/H-bond and Phe6.52 CH-π stacking are the main anchoring forces. The orthosteric binding pocket interactions: CHπ stacking Phe5.38, Asn6.55 are unique for subtype 6 serotonin receptors. The salt bridge between the protonated amine and Asp7.36 has been described for tacrine-related 5-HT6R inhibitors ( Figure 2) [38]. As we observed in our previous studies [38,40], the 1-(phenylsulfonyl)-1H-indole fragment, as in compounds 12-14, ensures the highest potencies (17-22 nM).  The tested compounds showed similar potencies against both hAChE and hBuChE, except for compound 14, which displayed 60-fold higher potency against hBuChE (hAChE IC 50 = 0.930 µM vs. hBuChE IC 50 = 0.016 µM). The predicted binding mode of compound 14 in BuChE is presented in Figure 2. To investigate the exceptional potency of compound 14 and differences in the activity between the related compounds, we performed MD simulations in Desmond MD System (D. E. Shaw Research, New York, NY, 2020, Maestro-Desmond Interoperability Tools, Schrödinger, New York, NY, 2020-4 Release). Interactions lasting more than 30% of the simulation time were documented ( Figure 3). The conformation of the molecules corresponds with the co-crystal structures of the related compounds [40]. Indole moiety in 14, accountable for the recognition by the 5-HT 6 R, was sandwiched between the acyl loop and the Gly116/117 oxyanion hole, and its position was preserved through MD simulation. The compound was bound in the catalytic active site of hBuChE and formed stable interactions, crucial for the inhibitory activity of: (1) conserved π-π and cation-π interactions between Trp82 and the protonated piperidine moiety, maintained through the entire MD simulation; (2) the salt bridge and H-bond between Asp70 and the protonated secondary amine in the linker part of the molecule, being the major interactions in the peripheral anionic site PAS; (3) aromatic π-π stacking interactions with Trp231 and Phe329 in the acyl pocket, observed and described in co-crystal studies of propidium [51,52], contributing to the binding stabilisation. position was preserved through MD simulation. The compound was bound in the catalytic active site of hBuChE and formed stable interactions, crucial for the inhibitory activity of: (1) conserved π-π and cation-π interactions between Trp82 and the protonated piperidine moiety, maintained through the entire MD simulation; (2) the salt bridge and H-bond between Asp70 and the protonated secondary amine in the linker part of the molecule, being the major interactions in the peripheral anionic site PAS; (3) aromatic ππ stacking interactions with Trp231 and Phe329 in the acyl pocket, observed and described in co-crystal studies of propidium [51,52], contributing to the binding stabilisation. Notable dissimilarities regarding the pivotal salt bridge between Asp70 in the PAS and the secondary amine of the linker as well as the π-π stacking formed by Trp82 in the mid-gorge of BuChE are presented. The persistence of each contact is represented as the participation percentage of the whole simulation time, seen as a value above the arrow depicting particular interaction. SAR analysis showed that the longer the linker, the higher the potency against both hAChE and hBuChE. This observation is consistent in all the series with the difference mostly pronounced for 12 and 14 in terms of hBuChE inhibition (IC 50 = 2.3 µM vs. 0.016 µM, respectively). The shorter linker (methylene vs. propylene) was responsible for compromised flexibility, causing the unfavourable shift of benzyl piperidine moiety, interruption, and permanent breakage of major binding interaction, which is a salt bridge with Asp70 ( Figure 3).
Antioxidant activity. As mentioned before, oxidative stress is one of the components of AD pathomechanism and is characterised by an imbalance in the production of ROS. One of the mechanisms by which antioxidants reduce the concentration of ROS and therefore prevent or slow the cell damage caused by free radicals is scavenging them [53]. Thus, we tested the scavenging ability of compounds 12-19 in the spectrophotometric ABTS assay that we used before [54]. In this assay, the ability of compounds to scavenge ABTS • radicals, which are generated from 2,2 -azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS), is measured. We tested the compounds at three concentrations: 30, 75 and 150 µM after 5, 30 and 60 min of incubation using Trolox, a water-soluble vitamin E analogue, as a reference (Table S1 in Supporting information).
The results showed that the antioxidant capacity of these compounds depends on both incubation time and their concentration. After 60 min of incubation, all the compounds reduced ABTS • by 6 to 27% at 30 µM, 11 to 36% at 75 µM and 24 to 43% at 150 µM (Figure 4), while Trolox reduced ABTS • by 59% at 30 µM and 100% at higher concentrations. Almost all the compounds showed at least one-third of Trolox activity at 150 µM. Additionally, imidazoline-2-one-based compounds 17-19 kept this activity also at the lowest tested concentration. Antioxidant properties of the presented series of compounds, together with their other activities, may contribute to the prevention of cell damage in AD.
Thus, we tested the scavenging ability of compounds 12-19 in the spectrophotometric ABTS assay that we used before [54]. In this assay, the ability of compounds to scavenge ABTS • radicals, which are generated from 2,2'-azino-bis(3-ethylbenzothiazoline-6sulphonic acid) (ABTS), is measured. We tested the compounds at three concentrations: 30, 75 and 150 μM after 5, 30 and 60 min of incubation using Trolox, a water-soluble vitamin E analogue, as a reference (Table S1 in Supporting information).
The results showed that the antioxidant capacity of these compounds depends on both incubation time and their concentration. After 60 min of incubation, all the compounds reduced ABTS • by 6 to 27% at 30 μM, 11 to 36% at 75 μM and 24 to 43% at 150 μM (Figure 4), while Trolox reduced ABTS • by 59% at 30 μM and 100% at higher concentrations. Almost all the compounds showed at least one-third of Trolox activity at 150 μM. Additionally, imidazoline-2-one-based compounds 17-19 kept this activity also at the lowest tested concentration. Antioxidant properties of the presented series of compounds, together with their other activities, may contribute to the prevention of cell damage in AD.  Metal-chelating properties. The homeostasis of metal ions in the brain is one of the important factors that determine its proper functioning. Disturbed levels of metal ions can lead to overproduction and aggregation of amyloid beta tau proteins and an increase in oxygen radicals production, leading to a degeneration of neurons observed in AD [55]. Additionally, it was observed that iron accumulation in microglia triggers a cascade of events leading to the production of a more sustained inflammatory response with increased reactive oxygen species (ROS) production and increased glycolysis [56,57]. Adjusting the balance of metal ions by supplementing or chelating may be beneficial in alleviating AD pathology. Thus, we tested the metal-chelating properties of compounds 12-19 against 10 metal ions, recognised as potentially involved in AD. The formation of ligand-ion (Al 3+ , Ca 2+ , Co 2+ , Cu 2+ , Fe 2+ , Fe 3+ , Mg 2+ , Ni 2+ , Pb 2+ or Zn 2+ ) complexes was measured spectrophotometrically (250-400 nm with 4 nm interval) at 50 µM concentrations of both, according to the protocol reported previously [54]. Metal-chelating ability was calculated as a shift of absorbance between the absorbance of the ligand-ion mixture and the sum of the absorbance of the ligand and ion separately. Compounds 15 and 18 were identified in the screening as Fe 2+ ions chelators ( Figure 5). This might be an advantageous feature of potential anti-neurodegenerative agents considering the correlation between iron overload, inflammation and glycolysis being inherent components of AD pathogenesis [58]. Therefore, the growing number of iron-chelating compounds with potential disease-modifying efficacy and the availability of MRI and CSF biomarkers of iron load encourage a deeper exploration of this therapeutic class in AD [59]. advantageous feature of potential anti-neurodegenerative agents considering the correlation between iron overload, inflammation and glycolysis being inherent components of AD pathogenesis [58]. Therefore, the growing number of iron-chelating compounds with potential disease-modifying efficacy and the availability of MRI and CSF biomarkers of iron load encourage a deeper exploration of this therapeutic class in AD [59].

Preliminary In Vitro ADMET Studies
The search for biologically active substances should be accompanied by the assessment of their ADMET parameters, describing the absorption, distribution, metabolism, excretion and toxicity of novel compounds. Such an approach allows for the assessment of the potential of compounds for further development and minimises the risk of failure while progressing to in vivo testing. In our project, as a preliminary screening of ADMET parameters, we tested metabolic stability, interaction with cytochrome P450 isoenzymes (CYP3A4 and CYP2D6) and potential hepatotoxicity of the selected compounds 14, 16 and 19 (each with a different 5-HT6-aiming fragment).
Metabolic stability. As the liver is the primary site of drug metabolism, we assessed the in vitro metabolic stability of compounds 14, 16 and 19 on human and mouse liver microsomes (HLMs, MLMs). The compounds and verapamil, used as a reference, were incubated with HLMs and MLMs for 2 h and then the resulting mixtures were analysed with UPLC-MS. The compounds showed high stability, with at least 75% of each left in the unaffected form after incubation, while the figure with verapamil was ca. 27% (Table  2). Based on the analysis of the MS spectra of the metabolites supported by the in silico predictions by MetaSite 8.0.1, we concluded that the main metabolic pathway was oxidative deamination, taking place on the nitrogen atom either in the oxyethylamine

Preliminary In Vitro ADMET Studies
The search for biologically active substances should be accompanied by the assessment of their ADMET parameters, describing the absorption, distribution, metabolism, excretion and toxicity of novel compounds. Such an approach allows for the assessment of the potential of compounds for further development and minimises the risk of failure while progressing to in vivo testing. In our project, as a preliminary screening of ADMET parameters, we tested metabolic stability, interaction with cytochrome P450 isoenzymes (CYP3A4 and CYP2D6) and potential hepatotoxicity of the selected compounds 14, 16 and 19 (each with a different 5-HT 6 -aiming fragment).
Metabolic stability. As the liver is the primary site of drug metabolism, we assessed the in vitro metabolic stability of compounds 14, 16 and 19 on human and mouse liver microsomes (HLMs, MLMs). The compounds and verapamil, used as a reference, were incubated with HLMs and MLMs for 2 h and then the resulting mixtures were analysed with UPLC-MS. The compounds showed high stability, with at least 75% of each left in the unaffected form after incubation, while the figure with verapamil was ca. 27% (Table 2). Based on the analysis of the MS spectra of the metabolites supported by the in silico predictions by MetaSite 8.0.1, we concluded that the main metabolic pathway was oxidative deamination, taking place on the nitrogen atom either in the oxyethylamine moiety (comp. 14) or in the piperidine ring (compounds 16 and 19). For detailed results, including UPLC chromatograms and MS spectra, see the Supporting information (Figures S1-S16).
Influence on CYP3A4 and CYP2D6 activity. Possible drug-drug interactions (DDIs) are an important issue for elderly patients, who are usually on multiple medications. They may result, among other things, from the interactions (inhibition or activation) of the cytochrome P450 isoenzymes, especially CYP3A4 and CYP2D6, which play the most significant role in drugs' biotransformation [60]. Therefore, to predict the potential influence of the studied compounds (14, 16 and 19) on the metabolism of other drugs, we used the CYP450 inhibition luminescence assay (Promega). None of the compounds showed any effect on CYP3A4 at 0.1 µM concentration. We observed inhibition of the enzyme starting from 1 µM concentration but it is worth noting that in all the concentrations (1, 10 and 25 µM) it was significantly lower than that of 1 µM of ketoconazole, used as a reference in the test ( Figure 6A). The more pronounced effect of the compounds was observed for the CYP2D6 isoform. While at the concentration of 0.1 µM, compound 19 did not influence the activity of CYP2D6, compounds 14 and 16 inhibited the enzyme by ca. 20% ( Figure 6B). The effect increased with the increasing concentrations of the compounds, with one exception: compound 19 at 1 µM showed a CYP2D6-activating effect. Nevertheless, none of the compounds was nearly as potent an inhibitor as quinidine, a drug used here as a reference.  Hepatotoxicity. To further investigate the safety of the compounds, we determined their hepatotoxicity. We used the MTS test that determines cell viability to assess the effect of compounds 14, 16 and 19 on HepG2 (human hepatocellular carcinoma) cells [61]. The compounds were tested at 1, 10, 50 and 100 μM, and cytostatic drug doxorubicin (DX) at 1 μM was used as a reference compound. While none of the tested compounds showed a cytotoxic effect at the concentration of 1 μM, they caused a significant reduction in cell viability at higher concentrations (10-100 μM, Figure 7). For a more precise evaluation of the effect, we determined the IC50 value for compound 14, which equalled 2.16 μM. This was a concentration at least 98-fold higher than the Ki for 5-HT6R and IC50 for hBuChE, which indicates a broad safety margin for the compound. Hepatotoxicity. To further investigate the safety of the compounds, we determined their hepatotoxicity. We used the MTS test that determines cell viability to assess the effect of compounds 14, 16 and 19 on HepG2 (human hepatocellular carcinoma) cells [61]. The compounds were tested at 1, 10, 50 and 100 µM, and cytostatic drug doxorubicin (DX) at 1 µM was used as a reference compound. While none of the tested compounds showed a cytotoxic effect at the concentration of 1 µM, they caused a significant reduction in cell viability at higher concentrations (10-100 µM, Figure 7). For a more precise evaluation of the effect, we determined the IC 50 value for compound 14, which equalled 2.16 µM. This was a concentration at least 98-fold higher than the K i for 5-HT 6 R and IC 50 for hBuChE, which indicates a broad safety margin for the compound.
1 μM was used as a reference compound. While none of the tested compounds showed a cytotoxic effect at the concentration of 1 μM, they caused a significant reduction in cell viability at higher concentrations (10-100 μM, Figure 7). For a more precise evaluation of the effect, we determined the IC50 value for compound 14, which equalled 2.16 μM. This was a concentration at least 98-fold higher than the Ki for 5-HT6R and IC50 for hBuChE, which indicates a broad safety margin for the compound.

Chemistry
Commercially available reagents were purchased from Merck, Aldrich, Acros or ChemPur and were used without additional purification, except for triethylamine (TEA),

Chemistry
Commercially available reagents were purchased from Merck, Aldrich, Acros or ChemPur and were used without additional purification, except for triethylamine (TEA), which was purified by vacuum distillation. Anhydrous solvents were obtained by drying over sodium (THF) or CaH 2 (DCM) followed by distillation under argon immediately before use. Microwave irradiation was carried out using the Discover LabMate reactor (CEM Corporation, USA). Reactions were monitored by thin-layer chromatography (TLC) (aluminium sheets precoated with silica gel 60 F 254 (Merck)). The spots were visualised with UV light (254 nm) and by staining with a 0.5% solution of ninhydrin in propan-1-ol or a solution of 5% (NH 4 ) 6 Mo 7 O 24 , 0.2% Ce(SO 4 ) 2 , and 5% H 2 SO 4 in water. Purification by column chromatography was carried out using silica gel mesh 0.063-0.200 mm (Sigma-Aldrich) as the stationary phase. 1 H NMR and 13 C NMR spectra were recorded on JEOL ECA400II or ECX500 at magnetic field strengths of 11.75 T, corresponding to 1 H and 13 C resonance frequencies of 500.16 MHz and 125.77 MHz at ambient temperature (25 • C). Chemical shifts (δ) are reported in parts per million (ppm) and referenced to the residual solvent signals (CDCl 3 1 H: 7.26 ppm, 13 C: 77.16 ppm; CD 3 OD 1 H: 3.31 ppm, 13 C: 49.00 ppm). Coupling constants (J) are reported in hertz (Hz). The purity and identity of the final compounds were confirmed with an ultra-performance liquid chromatography instrument (Waters ACQUITY equipped with a UPLC BEH C18 column, 1.7 µm, 2.1 100 mm; Acquity (10 min. gradient MeCN/H 2 O + 0.1% HCOOH, Q = 0.3 mL/min,) coupled to a Waters TQD mass spectrometer (ESI-tandem quadrupole) (Waters, Milford, MA, USA)).

Previously Reported Compounds
The following compounds were synthesized following previously published procedures To a solution of NaH (60% dispersion in mineral oil, 1.1 eq.) in dry DMF, diethyl(cyanomethyl) phosphonate (1.1 eq.) was added dropwise under argon atmosphere at 0 • C and the mixture was stirred for 20 min at 0 • C. Then, a solution of carbonyl compound 1 or 2 (1 eq.) in dry DMF at 0 • C was added. The resulting mixture was warmed up to room temperature and stirred overnight. The reaction was quenched by adding water and extracted with Et 2 O three times. The organic extracts were combined, dried with anhydrous MgSO 4 , filtered and concentrated in a vacuum. The crude product was purified by column chromatography (DCM/EtOAc, 3:1).

2-(1-Benzylpiperidin-4-ylidene)acetonitrile (3)
Following GP1, compound 3 was prepared using NaH (960 mg, 60% dispersion in mineral oil, 24 mmol), diethyl(cyanomethyl)phosphonate (3.88 mL, 24 mmol), 1-benzylpiperidin-4-one (1) (4.124 g, 3.89 mL, 21.8 mmol) in 42 mL of dry DMF (32 + 10 mL). Yield: 2.5 g (80%), orange oil. 1  To a solution of nitrile 3 or 4 (1 eq.) in MeOH, Mg (20 eq.) and a catalytic amount of I 2 were added and the mixture was sonicated in an ultrasonic bath until the nitrile was consumed (TLC). After the conversion, the reaction mixture was cooled in an ice bath and conc. HCl was added. When all material was dissolved, the mixture was neutralised to pH = 8, by adding 30% NaOH. The mixture was then extracted with EtOAc (3 × 50 mL) and the extracts were combined and dried over Na 2 SO 4 , filtered and concentrated under vacuum. The product was used in the next step without further purification. The solution of nitrile 5 or 6 (1 eq.) in dry toluene in an oven-dried and argon-purged flask was cooled to -78 • C, and 1 M DIBAL-H solution in toluene (2 eq.) was added dropwise. The reaction mixture was stirred at -78 • C for 1.5 h. Then, MeOH was added, the solution was poured into 5% H 2 SO 4 (aq) and finally basified with conc. NH 3 (aq). The precipitate was filtered off and washed with EtOAc. The filtrate was extracted with AcOEt three times. The combined organic extracts were washed with brine, dried with anhydrous MgSO 4 and purified by column chromatography (hexane/EtOAc, 2:1).

Molecular Modelling
All simulations, analyses and visualisations were conducted with Maestro software (Schrödinger LLC, New York, NY, 2020-4 Release).
Molecules underwent energy minimisation with LigPrep script and the most viable conformations were used in docking studies with Glide script. Molecular modelling was performed with the OPLS_2005 force field. Crystal structure of BuChE was extracted from PDB (ID: 7awh) and refined with Protein Preparation Wizard script, no constraints were applied in the docking process. The homology model of 5-HT 6 used in docking studies was developed with well-validated method and successfully utilised in previously published studies [38]. The only grid-based constraint applied in the docking processes was H-bond formation with Asp3.32.
To investigate activity differences in BuChE between related molecules (12,14,16,19), the MD simulations were conducted with Desmond MD System (D. E. Shaw Research, New York, NY, 2020, Maestro-Desmond Interoperability Tools, Schrödinger, New York, NY, 2020-4 Release). Systems were solvated in 10Å cubic box, TIP3P solvent model in 0.15 mM KCl and neutralised by the addition of 6 Cl − ions, with OPLS_2005 force field. The default settings were applied for MD simulations; the time of simulation was 10 ns. Interactions lasting more than 30% of the simulation time were documented.

Preparation of Solutions of Test and Reference Compounds
We prepared 10 mM of stock solutions of tested compounds in DMSO. Serial dilutions of compounds were prepared in a 96-well microplate in assay buffers using the automated pipetting system epMotion 5070 (Eppendorf). Each compound was tested in 8 concentrations from 10 −5 to 10 −12 M (final concentration).

5-HT 6 Receptor Binding Assay
The radioligand binding assay was performed using membranes from CHO-K1 cells stably transfected with the human 5-HT 6 receptor (PerkinElmer) according to the previously described procedure [38]. All assays were carried out in duplicates. We prepared 50 µL working solution of the tested compounds, 50 µL [ 3 H]-LSD (final concentration 2.5 nM) and 150 µL diluted membranes (15 µg protein per well) in assay buffer (50 mM Tris, pH 7.4, 10 mM MgCl 2 , 0.1 mM EDTA) and transferred them to polypropylene 96-well microplates using a 96-well pipetting station Rainin Liquidator (MettlerToledo). Methiothepin (10 µM) was used to define nonspecific binding. The microplate was covered with sealing tape, mixed and incubated for 60 min at 27 • C. The reaction was terminated by rapid filtration through GF/B filter mate presoaked with 0.5% polyethyleneimine for 30 min. Ten rapid washes with 200 µL 50 mM Tris buffer (4 • C, pH 7.4) were performed using the automated harvester system Harvester-96 MACH III FM (Tomtec). The filter mates were dried at 37 • C in a forced air fan incubator and then solid scintillator MeltiLex was melted on filter mates at 90 • C for 5 min. Radioactivity was counted in MicroBeta2 scintillation counter (PerkinElmer). Data were fitted to a one-site curve-fitting equation with Prism 6 (GraphPad Software) and K i values were estimated from the Cheng−Prusoff equation.

Human Acetylcholinesterase and Butyrylcholinesterase In Vitro Inhibitory Activity
The assay was based on spectrophotometric Ellman's method [49] and was performed in 96-well microplates with the reagents purchased from Sigma-Aldrich (Steinheim, Germany) according to the previously described procedure [38]. hBuChE was kindly donated by Vivonics (Bedford, MA, USA). Stock solutions of the tested compounds (1.14 mM) and reagents were prepared in DMSO and water, respectively. Before starting the enzymatic reaction, the test compound or water or mixture of DMSO/water in appropriate ratio (0.025 mL) were incubated in a phosphate buffer (0.2 mL, 100 mM, pH 8.0) with Ellman's reagent (0.02 mL; 2.5 mM) and hAChe or hBuChE (0.02 mL; 0.384 U/mL) at 36 • C for 5 min. Then, substrate solutions (0.02 mL, 3.75 mM) were added: acetylthiocholine for hAChE or butyrylthiocholine for hBuChE. After 5 minutes of the enzymatic reaction, the changes in absorbance were measured at 412 nm, using a microplate reader (SPECTROstar Nano; BMG Labtech, Ortenberg, Germany). First, compounds were tested at the screening concentration of 10 µM. Based on the equation 100 − (S/B) × 100 (where S and B were the respective enzyme activities with and without the test sample, respectively) the percentage inhibition of each enzyme for each compound was calculated. Next, compounds with enzyme inhibitory activity higher than 50% were tested again at seven different concentrations to determine their IC 50 values. The IC 50 values were calculated using nonlinear regression (GraphPad Prism 5; GraphPad Software, San Diego, CA, USA) by plotting the residual enzyme activities against the applied inhibitor concentration. Donepezil and tacrine were used as the reference compounds. All experiments were performed in triplicate.

Free Radical Scavenging In Vitro Activity
The spectrophotometric ABTS assay was performed in 96-well microplates with the reagents purchased from Sigma-Aldrich (Steinheim, Germany) according to the previously described procedure [54]. Stock solutions of the tested compounds (6.4 mM) and reagents were prepared in DMSO and water, respectively. 2,2 -Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) free radical (ABTS • ) was prepared by mixing the stock solution of ABTS (7 mM) and K 2 S 2 O 3 (2.45 mM) in ratio 1:1. The mixture was kept in the dark at 4 • C for 16 h. The obtained solution of (ABTS • ) was diluted with water 20 times. Then, the test compound or water or mixture of DMSO/water in the appropriate ratio (0.02 mL) was added to ABTS • solution (0.18 mL). The three concentrations of the test compounds and Trolox were tested in triplicate: 0.15, 0.075 and 0.03 mM. Changes in absorbance were measured at 734 nm after 5, 30 and 60 min, using a microplate reader (SPECTROstar Nano; BMG Labtech, Ortenberg, Germany). Based on the equation ((A 0 − A)/A 0 ) × 100% (where A 0 and A were the absorbances without and with the test compound, respectively) the percentage of free radical scavenging activity was calculated. The results are presented as mean ± SD.

Metal-Chelating In Vitro Activity
The spectrophotometric assessment of metal-chelating properties was performed in UV-transparent 96-well microplates (Corning ® Costar) with the reagents purchased from Sigma-Aldrich (Steinheim, Germany) according to the previously published procedure [54].
Stock solutions of the tested compounds (6.4 mM) were prepared in DMSO and diluted with water to the final concentration of 0.5 mM. Stock solutions of cation salts (0.5 mM)-ZnCl 2 , CoCl 2 , AlCl 3 , NiCl 2 , CaCl 2 , MgCl 2 , Pb(CH 3 COO) 2 , FeSO 4 , FeCl 3 -were prepared in water. The test compound, or the cation salt or water or mixture of DMSO/water in the appropriate ratio (0.02 mL), or the mixture of the test compound and the cation salt (0.04 mL, ratio 1:1) were incubated in HEPES buffer (0.18 mL or 0.16 mL for the mixture of compound and the cation salt, 20 mM, pH 7.4) at room temperature for 30 min. Then, absorption spectra in the range of 250-400 nm (4 nm interval) were recorded using a microplate reader (SPECTROstar Nano; BMG Labtech, Ortenberg, Germany). A shift of absorbance was calculated as differences between the absorbance of the mixture of the compound and the cation salt and the sum of the absorbances of the compound and the cation salt.

ADME-Tox Parameters
The ADME-Tox parameters were determined in vitro as described previously [61]. Statistical significances were analysed by GraphPad Prism ™ 8 software using one-way ANOVA and Bonferroni's multiple comparison post-test. All reference compounds (caffeine, doxorubicin, ketoconazole, quinidine and verapamil) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All UPLC/MS analyses were done by Waters ACQUITY TQD system with the TQ Detector (Waters, Milford, USA). The absorbance and luminescence were measured using a microplate reader (EnSpire Multimode; PerkinElmer, Waltham, MA, USA).

Metabolic Stability in Human Liver Microsomes
The metabolic stability was evaluated by incubation of compounds with human liver microsomes (Sigma-Aldrich, St. Louis, MO, USA) in 10 mM Tris-HCl buffer (pH 7.4) at 37 • C for 120 min in the presence of NADPH Regeneration System (Promega, Madison, WI, USA). The percentage of the remaining substrate and the most probable metabolic pathways were estimated by UPLC-MS analyses of the reaction mixtures.

Influence on CYP P450 Activity
The influence on CYP 3A4 and 2D6 activity was estimated by respective luminescent CYP P450-Glo™ assay (Promega ® , Madison, WI, USA). The compounds were tested in the concentration range of 0.1-25 µM. The results were compared to the respective reference CYP inhibitors: ketoconazole and quinidine.

Hepatotoxicity in HepG2 Cells
The HepG2 (ATCC ® HB-8065™) cell line was cultivated according to the procedure provided by ATCC. The tested compounds were incubated with cells for 72 h in the following concentrations: 1, 10, 50 and 100 µM. To evaluate the cells' viability, CellTiter 96 ® AQ ueous Non-Radioactive Cell Proliferation Assay (MTS, Promega ® , Madison, WI, USA) was performed. The antiproliferative drug doxorubicin was used as a positive control at the concentration of 1 µM. For determination of the IC 50 value for compound 14, additional study was performed. The tested compound was incubated with the HepG2 cells for 72 h in the following nine concentrations: 0.001, 0.01, 0.1, 1, 5, 10, 25, 50 and 100 µM. The cells' viability was estimated by MTS test. IC 50 was calculated by Graph Pad Prism 8.0.1 software.

Conclusions
Therapeutic options in AD treatment are limited to drugs that temporarily relieve the symptoms of the disease-cholinesterase inhibitors (donepezil, rivastigmine and galantamine) and NMDA receptor antagonists (memantine). In 2021 the U.S. Food and Drug Administration approved aducanumab-an amyloid beta-directed monoclonal antibodythe first drug with disease-modifying potential. Aducanumab was approved under the accelerated approval pathway, based on effects on a surrogate endpoint, which in this case was a reduction of amyloid beta plaques in the brains of patients with AD. Its efficacy in terms of clinical benefits is yet to be determined [62,63]. Hence, there is still an unmet need to search for effective therapy for AD, primarily with disease-modifying effects but preferably with a symptoms-relieving component for patients with the already developed disease. Following this idea, we have been working on multifunctional ligands that address the neurotransmission disorders and oxidative stress underlying AD. Within the synthesised and evaluated compounds, we selected compound 14, with an excellent and balanced potency against h5-HT 6 R (K i = 22 nM) and hBuChE (IC 50 = 16 nM) and free radicals scavenging activity. The compound is metabolically stable both on human and mouse liver microsomes and safe in the concentrations required for modulation of h5-HT 6 R and hBuChE. Therefore, compound 14 may serve as a promising candidate for further development of multitarget-directed ligands aiming at symptoms and causes of Alzheimer's disease.

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