Design and synthesis of novel tacrine–indole hybrids as potential multitarget-directed ligands for the treatment of Alzheimer's disease

Abstract

The authors report on the synthesis and biological evaluation of new compounds whose structure combines tacrine and indole moieties. Tacrine–indole heterodimers were designed to inhibit cholinesterases and β-amyloid formation, and to cross the blood–brain barrier. The most potent new acetylcholinesterase inhibitors were compounds 3c and 4d (IC₅₀ = 25 and 39 nM, respectively). Compound 3c displayed considerably higher selectivity for acetylcholinesterase relative to human plasma butyrylcholinesterase in comparison to compound 4d (selectivity index: IC₅₀ [butyrylcholinesterase]/IC₅₀ [acetylcholinesterase] = 3 and 0.6, respectively). Furthermore, compound 3c inhibited β-amyloid-dependent amyloid nucleation in the yeast-based prion nucleation assay and displayed no dsDNA destabilizing interactions with DNA. Compounds 3c and 4d displayed a high probability of crossing the blood–brain barrier. The results support the potential of 3c for future development as a dual-acting therapeutic agent in the prevention and/or treatment of Alzheimer's disease.

Alzheimer's disease (AD) is the most prevalent cause of dementia in older adults. Its pathogenesis involves multiple pathological mechanisms, including extracellular β-amyloid (Aβ) deposition, tau protein aggregation, oxidative stress, mitochondrial dysfunction and decreased levels of acetylcholine [1–5]. Consequently, compounds targeting several of these mechanisms simultaneously are more likely to prevent or cure AD than compounds active against single targets. This fact inspired the development of multitarget-directed ligands (MTDLs) – that is, small molecules that can hit multiple targets simultaneously – leading to the discovery of some promising compounds [6]. Tacrine (9-amino-1,2,3,4-tetrahydroacridine), a nonselective cholinesterase inhibitor, was the first therapeutic agent approved by the US FDA for the treatment of Alzheimer's disease [7]. Although tacrine later fell out of use because of its modest therapeutic efficacy and potential hepatotoxicity [8], it still represents an attractive structural moiety for AD drug design [9–13]. Its potential for future drug development is supported by studies that demonstrated successful lead optimization of tacrine, leading to a broader range of biological activities and reduced hepatotoxicity [14–16].

The indole nucleus is an important heterocyclic moiety found in many natural and synthetic molecules with interesting pharmacological activities [17,18]. This moiety is also frequently present in structures of acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibitors [19–25]. MTDLs containing an indole ring function as dual binding site AChE inhibitors, and the indole ring introduces disease-modifying properties into the molecule. This new activity is caused by inhibition of AChE-induced Aβ peptide aggregation via the ligand binding to peripheral anionic sites of the enzyme [23,26–28]. Furthermore, some natural marine compounds with an indole scaffold have been identified as the inhibitors of GSK-3β, DYRK1A, CLK1 and CK1δ, all of which are involved in the process of neurofibrillary tangle formation through phosphorylation of the microtubule-associated protein tau in the brain [29].

Another study reported on quinoline–indole derivatives as MTDLs for the treatment of AD [30]. These derivatives exerted multiple biological activities, including antioxidant effect, biometal chelation and modulation of Aβ aggregation as well as neurotrophic and neuroprotective properties. In the authors' previous work, we synthesized a series of tacrine–tryptophan heterodimers as multitarget agents with potential for AD drug development [31]. The compound S-K1035 (Figure 1) was the most potent inhibitor of both AChE and BChE with respective IC₅₀ values of 6.3 and 9.1 nM. In addition, S-K1035 inhibited Aβ₄₂ self-aggregation (58.6 ± 5.1% at 50 μM) as well as AChE-induced Aβ₄₀ aggregation (48.3 ± 6.3% at 100 μM). Protective effects of S-K1035 on cognition were demonstrated using the scopolamine-induced cognitive deficit rat model. Likewise, in vivo toxicological studies found generally lower toxicity for S-K1035 compared with tacrine. An interesting series of indole–tacrine heterodimers were synthesized by Muñoz-Ruiz et al. [32]. Tacrine–indole heterodimers, which contain a 6-chlorotacrine fragment, nonsubstituted indole and methylene tether with six or seven carbons, showed excellent inhibitory activity against human AChE (hAChE), human BChE (hBChE) and β-secretase 1 at both nanomolar and picomolar concentrations (IC₅₀ = 70 and 20 pM). Taken together, tacrine–indole derivatives can be promising compounds for Alzheimer's disease drug discovery and development.

Figure 1. . Design strategy for tacrine–indole heterodimers.

Design

In this work, the authors report on the synthesis and biological evaluation of new tacrine–indole heterodimers. This design was inspired by a study by Rodríguez-Franco et al., who synthesized a series of tacrine–melatonin heterodimers as novel subnanomolar to picomolar inhibitors of human AChE with antioxidant properties [33]. The tacrine–melatonin hybrid (Figure 1), derived from 6,8-dichlorotacrine and an unsubstituted indole with a 6-methylene linker, inhibited hAChE at IC₅₀ of 8 pM.

Recently, the authors reported on the synthesis of tacrine–tacrine homodimers [34] tethered with alkylene–thiourea linkers (Figure 1). These compounds inhibited AChE with high potency that reached even subnanomolar IC₅₀ values. Tacrine–tacrine homodimers with four and six methylene units in the linker were found to be particularly potent inhibitors of AChE, with IC₅₀ of 2 and 8 nM, respectively.

Based on these results, the authors expected that tacrine and indole moieties tethered through linkers of variable chain lengths would be promising molecules for AD therapeutic development. In these molecules, the tacrine moiety is expected to interact with aromatic residues in the catalytic active site of AChE. By contrast, indole is expected to adopt lodging in the peripheral anionic site (PAS), thus amplifying cholinesterase inhibition and preventing AChE-induced Aβ aggregation [30,33]. Optimization of linker length was expected to ensure appropriate ligand binding into the active site of AChE. In addition, the authors determined the DNA intercalating activity from a toxicological point of view, as acridine derivatives are known to target DNA through intercalation [35]. The authors have also predicted the ability of new compounds to cross the blood–brain barrier (BBB) as a key prerequisite for drugs targeting the brain. In general, the authors' goal was to develop novel ligands with broad therapeutic profile that overcome the limitations of current AD drugs.

Experimental section

Chemistry & chemical methods

All reagents used in the synthesis were obtained commercially and were used without further purification unless otherwise specified. The reactions were monitored by TLC using ALUGRAM-SIL G/UV254 TLC sheets (Macherey-Nagel GmbH & Co. KG, Düren, Germany). Purification by flash chromatography was performed using silica gel 60 Å 0.0040–0.063 mm (Merck KGaA, Darmstadt, Germany) with the indicated eluent. Melting points were determined by a Boetius hot-stage apparatus and are uncorrected. NMR spectra were recorded at room temperature on Varian Mercury Plus 400 MHz (Varian Medical Systems, Inc., CA, USA) operating at 400 MHz for ¹H and 100 MHz for ¹³C and Varian VNMRS 600 MHz (Varian Medical Systems, Inc.) operating at 600 MHz for ¹H and 150 MHz for ¹³C. Chemical shifts (δ in ppm) from internal solvent DMSO-d₆ are given. Infrared spectra were recorded in the range of 4000–500 cm^-1 on an IRAffinity-1 FTIR spectrophotometer (Shimadzu Corporation, Kyoto, Japan) using the KBr method.

General procedure for the preparation of 3a-g & 4a–f

To the solution of corresponding N-(1,2,3,4-tetrahydroacridin-9-yl)-1,n-alkanediamine 1 (0.1 mmol) in DCM (4 ml), isothiocyanate 2 (2a: 3-(2-ethylisothiocyanato)-1H-indol, 2b: 3-(2-isothiocyanatoethyl)-1-methoxy-1H-indole; 0.17 mmol) and TEA (0.1 mmol) were added. The reaction mixture was stirred at room temperature for 20–24 h. After completion of the reaction, the solvent was evaporated, and crude product was purified by column chromatography using MeOH as an eluent.

1-[2-(1H-indol-3-yl)ethyl]-3-[2-(1,2,3,4-tetrahydroacridin-9-ylamino)ethyl]thiourea (3a)

Yield 68%; yellow solid; Mp = 101–102°C; ¹H-NMR (400 MHz, DMSO): δ 1.74–1.88 (m, 4H, 2 × CH₂, H-2, 3), 2.71–2.77 (m, 2H, CH₂, H-1), 2.85–2.94 (m, 4H, 2 × CH₂, H-4, 8´), 3.52–3.75 (m, 6H, 3 × CH₂, H-2´, 3´, 7´), 5.46 (t, 1H, NH, H-1´, J = 5.6 Hz), 6.96 (t, 1H, CH, H-5´´, J = 7.0 Hz), 7.05 (ddd, 1H, CH, H-6´´, J = 1.2; 7.0; 8.0 Hz), 7.13 (d, 1H, CH, H-2´´, J = 2.0 Hz), 7.30–7.36 (m, 2H, 2 × CH, H-7, 7´´), 7.52 (ddd, 1H, CH, H-6, J = 1.2; 6.8; 8.2 Hz), 7.60 (d, 1H, CH, H-4´´, J = 7.6 Hz), 7.71 (dd, 1H, CH, H-5, J = 1.2; 7.6 Hz), 8.10 (d, 1H, CH, H-8, J = 8.4 Hz), 10.81 (bs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 25.0 (C-1, 8´), 33.4 (C-4), 43.8 (C-3´), 43.9 (C-7´), 48.2 (C-2´), 111.3 (C-7´´), 111.6 (C-3´´), 115.9 (C-9a), 118.2 (C- 5´´), 118.4 (C-4´´), 120.0 (C-8a), 120.9 (C-6´´), 122.7 (C-2´´), 123.1 (C-8), 123.2 (C-7), 127.2 (C-3a´´), 127.9 (C-6), 128.1 (C-5), 136.2 (C-7a´´), 146.7 (C-5a), 150.2 (C-9), 157.8 (C-4a), 181.5 (C = S). Infrared (IR) (KBr, selected bands): 3056, 2927, 2858, 2363, 2341, 1559, 1500, 1457, 1420, 1340, 1292, 1276, 1126, 744. Anal. Calcd for C₂₆H₂₉N₅S (443.61): C, 70.40; H, 6.59; N, 15.79. Found: C, 70.43; H, 6.50; N, 15.70.

1-[2-(1H-indol-3-yl)ethyl]-3-[3-(1,2,3,4-tetrahydroacridin-9-ylamino)propyl] thiourea (3b)

Yield 71%; white solid; Mp = 105–106°C; ¹H-NMR (400 MHz, DMSO): δ 1.71–1.85 (m, 6H, 3 × CH₂, H-2, 3, 3´), 2.68–2.80 (m, 2H, CH₂, H-1), 2.85–2.93 (m, 4H, 2 × CH₂, H-4, 9´), 3.30–3.47 (m, 4H, 2 × CH₂, H-2´, 4´), 3.53–3.70 (m, 2H, CH₂, H-8´), 5.46 (t, 1H, NH, H-1´, J = 6.0 Hz), 6.95 (t, 1H, CH, H-5´´, J = 7.2 Hz), 7.05 (t, 1H, CH, H-6´´, J = 7.2 Hz), 7.13 (d, 1H, CH, H-2´´, J = 2.2 Hz), 7.30–7.36 (m, 2H, 2 × CH, H-7, 7´´), 7.51 (ddd, 1H, CH, H-6, J = 1,2; 6,8; 8.2 Hz), 7.60 (d, 1H, CH, H-4´´, J = 7.6 Hz), 7.70 (dd, 1H, CH, H-5, J = 1.2; 7.6 Hz), 8.13 (d, 1H, CH, H-8, J = 8.4 Hz), 10.80 (šs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 25.1 (C-9´), 25.2 (C-1), 30.5 (C-3´), 33.5 (C-4), 45.0 (C-8´), 45.3 (C-2´, 4´), 111.3 (C-7´´), 111.6 (C-3´´), 115.9 (C-9a), 118.2 (C- 5´´), 118.4 (C-4´´), 120.2 (C-8a), 120.9 (C-6´´), 122.7 (C-2´´), 122.9 (C-8), 123.2 (C-7), 127.2 (C-3a´´), 127.8 (C-6), 128.3 (C-5), 136.2 (C-7a´´), 146.9 (C-5a), 150.1 (C-9), 157.9 (C-4a), 181.5 (C = S). IR (KBr, selected bands): 3056, 2926, 2855, 1560, 1501, 1457, 1422, 1354, 1294, 1127, 744. Anal. Calcd for C₂₇H₃₁N₅S (457.63): C, 70.86; H, 6.83; N, 15.30. Found: C, 70.84; H, 6.81; N, 15.34.

1-[2-(1H-indol-3-yl)ethyl]-3-[4-(1,2,3,4-tetrahydroacridn-9-ylamino)butyl]thiourea (3c)

Yield 44%; powder solid; Mp = 94–95°C; ¹H-NMR (400 MHz, DMSO): δ 1.44–1.60 (m, 4H, 2 × CH₂, H-3´, 4´), 1.74–1.88 (m, 4H, 2 × CH₂, H-2, 3), 2.72 (t, 2H, CH₂, H-1, J = 6.0 Hz), 2.85–2.93 (m, 4H, 2 × CH₂, H-4, 10´), 3.30–3.45 (m, 4H, 2 × CH₂, H-2´, 5´), 3.58–3.72 (m, 2H, CH₂, H-9´), 5.40 (t, 1H, NH, H-1´, J = 6.2 Hz), 6.96 (ddd, 1H, CH, H-5´´, J = 1.0; 7.0; 7.9 Hz), 7.05 (ddd, 1H, CH, H-6´´, J = 1.2; 7.0; 8.1 Hz), 7.13 (d, 1H, CH, H-2´´, J = 2.2 Hz), 7.33 (dd, 1H, CH, H-7´´, J = 0.7; 8.1 Hz), 7.34 (dd, 1H, CH, H-7, J = 1.3; 8.3 Hz), 7.41 (šs, 1H, NH, H-6´), 7.52 (ddd, 1H, CH, H-6, J = 1.2; 6.8; 8.2 Hz), 7.6 (d, 1H, CH, H-4´´, J = 7.7 Hz), 7.7 (dd, 1H, CH, H-5, J = 1.0; 8.4 Hz), 8.12 (d, 1H, CH, H-8, J = 8.0 Hz), 10.81 (šs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.8 (C-2, 3), 24.9 (C-10´), 25.1 (C-1), 26.2 (C-4´), 28.1 (C-3´), 33.5 (C-4), 43.2 (C-5´), 44.2 (C-9´), 47.7 (C-2´), 111.3 (C-7´´), 111.7 (C-3´´), 115.9 (C-9a), 118.2 (C-5´´), 118.5 (C-4´´), 120.3 (C-8a), 120.9 (C-6´´), 122.7 (C-2´´), 123.0 (C-8), 123.2 (C-7), 127.3 (C-3a´´), 127.9 (C-6), 128.2 (C-5), 136.2 (C-7a´´), 145.9 (C-5a), 150.3 (C-9), 157.9 (C-4a), 181.5 (C = S). IR (KBr, selected bands): 3344, 3226, 2939, 2860, 1560, 1508, 1436, 1357, 1279, 1236, 748, 742. Calcd for C₂₈H₃₃N₅S (471.66): C, 71.30; H, 7.05; N, 14.85. Found: C, 71.34; H, 7.09; N, 14.80.

1-[2-(1H-indol-3-yl)ethyl]-3-[5-(1,2,3,4-tetrahydroacridin-9-ylamino)pentyl]thiourea (3d)

Yield 48%; yellow solid; Mp = 82–83°C; ¹H-NMR (400 MHz, DMSO): δ 1.22–1.33 (m, 2H, CH₂, H-4´), 1.40–1.50 (m, 2H, CH₂, H-5´), 1.52–1.63 (m, 2H, CH₂, H-3´), 1.75–1.87 (m, 4H, 2 × CH₂, H-2, 3), 2.68–2.74 (m, 2H, CH₂, H-1), 2.84–2.93 (m, 4H, 2 × CH₂, H-4, 11´), 3.25–3.41 (m, 4H, 2 × CH₂, H-2´, 6´), 3.56–3.70 (m, 2H, CH₂, H-10´), 5.34 (t, 1H, NH, H-1´, J = 6.4 Hz), 6.96 (ddd, 1H, CH, H-5´´, J = 1.2; 7.2; 8.0 Hz), 7.05 (ddd, 1H, CH, H-6´´, J = 1.2; 7; 8.1 Hz), 7.13 (d, 1H, CH, H-2´´, J = 2.2 Hz), 7.30–7.35 (m, 2H, 2 × CH, H-7, 7´´), 7.38 (šs, 1H, NH, H-7´), 7.51 (ddd, 1H, CH, H-6, J = 1.2; 6.8; 8.2 Hz), 7.60 (d, 1H, CH, H-4´´, J = 8.0 Hz), 7.7 (dd, 1H, CH, H-5, J = 1.0; 8.4 Hz), 8.11 (d, 1H, CH, H-8, J = 8.0 Hz), 10,8 (šs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 23.7 (C-4´), 25.0 (C-11´), 25.1 (C-1), 28.5 (C-5´), 30.3 (C-3´), 33.5 (C-4), 43,5 (C-6´), 44.8 (C-10´), 47.9 (C-2´), 111.3 (C-7´´), 111.7 (C-3´´), 115.9 (C-9a), 118.1 (C-5´´), 118.4 (C-4´´), 120.3 (C-8a), 120.9 (C-6´´), 122.7 (C-2´´), 123.0 (C-8), 123.2 (C-7), 127.2 (C-3a´´), 127.8 (C-6), 128.3 (C-5), 136.2 (C-7a´´), 146.9 (C-5a), 150.3 (C-9), 157.9 (C-4a), 181.3 (C = S). IR (KBr, selected bands): 3257, 3057, 2930, 2858, 2358, 2334, 1560, 1500, 1355, 1293, 1275, 1127, 744. Calcd for C₂₉H₃₅N₅S (485.69): C, 71.72; H, 7.26; N, 14.42. Found: C, 71.70; H, 7.29; N, 14.40.

1-[2-(1H-indol-3-yl)ethyl]-3-[6-(1,2,3,4-tetrahydroacridin-9-ylamino)hexyl]thiourea (3e)

Yield 65%; yellow solid; Mp = 79–80°C; ¹H-NMR (400 MHz, DMSO): δ 1.20–1.33 (m, 4H, 2 × CH₂, H-4´, 5´), 1.37–1.50 (m, 2H, CH₂, H-6´), 1.50–1.60 (m, 2H, CH₂, H-3´), 1.75–1.84 (m, 4H, 2 × CH₂, H-2, 3), 2.68–2.75 (m, 2H, CH₂, H-1), 2.84–2.94 (m, 4H, 2 × CH₂, H- 4, 12´), 3.30–3.43 (m, 4H, 2 × CH₂, H-2´, 7´), 3.58–3.70 (m, 2H, CH₂, H-11´), 5.34 (t, 1H, NH, H-1´, J = 6.4 Hz), 6.96 (t, 1H, CH, H-5´´, J = 7.4 Hz), 7.05 (t, 1H, CH, H-6´´, J = 7.4 Hz), 7.13 (d, 1H, CH, H-2´´, J = 2.0 Hz), 7.30–7.37 (m, 2H, 2 × CH, H-7, 7´´), 7.40 (šs, 1H, NH, H-8´), 7.50 (t, 1H, CH, H-6, J = 7.6 Hz,), 7.60 (d, 1H, CH, H-4´´, J = 7.6 Hz), 7.70 (d, 1H, CH, H-5, J = 8.8 Hz), 8.11 (d, 1H, CH, H-8, J = 8.0 Hz), 10.8 (šs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 25.0 (C-12´, 1), 26.1 (C-4´, 5´), 28.7 (C-6´), 30.6 (C-3´), 43.5 (C-7´), 44.2 (C-11´), 47.9 (C-2´), 111.3 (C-7´´), 111.7 (C-3´´), 115.8 (C-9a), 118.1 (C-5´´), 118.4 (C-4´´), 120.2 (C-8a), 120.8 (C-6´´), 122.6 (C-2´´), 122.9 (C-8), 123.1 (C-7), 127.2 (C-3a´´), 127.7 (C-6), 128.3 (C-5), 136.2 (C-7a´´), 146.9 (C-5a), 150.2 (C-9), 157.9 (C-4a), 181.5 (C = S). Calcd for C₃₀H₃₇N₅S (499.71): C, 72.11; H, 7.46; N, 14.01. Found: C, 72.09; H, 7.43; N, 14.04.

1-[2-(1H-indol-3-yl)ethyl]-3-[7-(1,2,3,4-tetrahydroacridin-9-ylamino)heptyl]thiourea (3f)

Yield 45%; yellow solid; Mp = 74–75°C; ¹H-NMR (400 MHz, DMSO): δ 1.15–1.30 (m, 6H, 3 × CH₂, H-4´–6´), 1.34–1.44 (m, 2H, CH₂, H-7´), 1.48–1.58 (m, 2H, CH₂, H-3´), 1.74–1.84 (m, 4H, 2 × CH₂, H-2, 3), 2.65–2.72 (m, 2H, CH₂, H-1), 2.84–2.92 (m, 4H, 2 × CH₂, H- 4, 13´), 3.20–3.42 (m, 4H, 2 × CH₂, H-2´, 8´), 3.56–3.68 (m, 2H, CH₂, H-12´), 5.38 (t, 1H, NH, H-1´, J = 6.4 Hz), 6.94 (t, 1H, CH, H-5´´, J = 7,4 Hz), 7.04 (t, 1H, CH, H-6´´, J = 7,4 Hz), 7.12 (d, 1H, CH, H-2´´, J = 2.0 Hz), 7.30–7.35 (m, 2H, 2 × CH, H-7, 7´´), 7.37 (šs, 1H, NH, H-9´), 7.50 (t, 1H, CH, H-6, J = 7.6 Hz), 7.60 (d, 1H, CH, H-4´´, J = 7.6 Hz), 7.70 (d, 1H, CH, H-5, J = 8.4 Hz), 8.10 (d, 1H, CH, H-8, J = 8.4 Hz), 10.78 (šs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 25.0 (C-1, 13´), 26.3 (C-4´, 5´, 6´), 28.5 (C-7´), 30.5 (C-3´), 33.4 (C-4), 43.5 (C-8´), 44.0 (C-12´), 47.9 (C-2´), 111.3 (C-7´´), 111.7 (C-3´´), 115.7 (C-9a), 118.1 (C-5´´), 118.4 (C-4´´), 120.1 (C-8a), 120.9 (C-6´´), 122.7 (C-2´´), 123.0 (C-8), 123.2 (C-7), 127.2 (C-3a´´), 127.8 (C-6), 128.1 (C-5), 136.2 (C-7a´´), 146.7 (C-5a), 150.4 (C-9), 157.7 (C-4a), 181.5 (C = S). Calcd for C₃₁H₃₉N₅S (513.74): C, 72.47; H, 7.65; N, 13.63. Found: C, 72.40; H, 7.63; N, 13.59.

1-[2-(1H-indol-3-yl)ethyl]-3-[8-(1,2,3,4-tetrahydroacridin-9-ylamino)octyl]thiourea (3g)

Yield 79%; yellow solid; Mp = 69–70°C; ¹H-NMR (400 MHz, DMSO): δ 1.15–1.32 (m, 6H, 3 × CH₂, H-4´–7´), 1.36–1.48 (m, 2H, CH₂, H-8´), 1.48–1.60 (m, 2H, CH₂, H-3´), 1.75–1.84 (m, 4H, 2 × CH₂, H-2, 3), 2.67–2.74 (m, 2H, CH₂, H-1), 2.84–2.93 (m, 4H, 2 × CH₂, H- 4, 14´), 3.22–3.42 (m, 6H, 3 × CH₂, H-2´, 9´), 3.55–3.70 (m, 2H, CH₂, H-13´), 5.36 (t, 1H, NH, H-1´, J = 6.4 Hz), 6.96 (t, 1H, CH, H-5´´, J = 7.4 Hz), 7.05 (t, 1H, CH, H-6´´, J = 7.4 Hz), 7.14 (d, 1H, CH, H-2´´, J = 2.0 Hz), 7.30–7.37 (m, 2H, 2 × CH, H-7, 7´´), 7.40 (šs, 1H, NH, H-10´), 7.50 (t, 1H, CH, H-6, J = 7.4 Hz), 7.60 (d, 1H, CH, H-4´´, J = 7.4 Hz), 7.7 (d, 1H, CH, H-5, J = 8.4 Hz), 8.10 (d, 1H, CH, H-8, J = 8.0 Hz), 10.81 (šs, 1H, NH, H-1´´). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 25.5 (C-1), 25.4 (C-14´), 26.7, 29.1 (C-4´–8´), 30.5 (C-3´), 33.5 (C-4), 43.5 (C-9´), 44.3 (C-13´), 48.0 (C-2´), 111.0 (C-7´´), 111.7 (C-3´´), 116.0 (C-9a), 118.5 (C-4´´, 5´´), 120.2 (C-8a), 121.0 (C-6´´), 122.9 (C-2´´), 123.2 (C-8), 123.2 (C-7), 128.0 (C-6), 127.2 (C-3a´´), 128.2 (C-5), 136.2 (C-7a´´), 147.0 (C-5a), 150,5 (C-9), 158.0 (C-4a), 181.5 (C = S). Calcd for C₃₂H₄₁N₅S (527.77): C, 72.82; H, 7.83; N, 13.27. Found: C, 72.80; H, 7.85; N, 13.20.

1-[2-(1-methoxyindol-3-yl)ethyl]-3-[2-(1,2,3,4-tetrahydroacridin-9-ylamino)ethyl] thiourea (4a)

Yield 43%; yellow solid; Mp = 92–93°C; ¹H-NMR (400 MHz, DMSO): δ 1.71–1.82 (m, 4H, 2 × CH₂, H-2, 3), 2.72 (t, 2H, CH₂, H-1, J = 6.4 Hz), 2.82–2.90 (m, 4H, 2 × CH₂, H-4, 8´), 3.50–3.74 (m, 6H, 3 × CH₂, H-2´, 3´, 7´), 4.0 (s, 3H, OCH₃), 5.40 (t, 1H, NH, H-1´, J = 6.0 Hz), 6.96 (t, 1H, CH, H-5´´, J = 7.6 Hz), 7.10 (t, 1H, CH, H-6´´, J = 7.6 Hz), 7.13 (s, 1H, CH, H-2´´), 7.26 (d, 1H, CH, H-7´´, J = 8.2 Hz), 7.35 (t, 1H, CH, H-7, J = 7.9 Hz), 7.53 (d, 1H, CH, H-4´´, J = 7.7 Hz), 7.59 (t, 1H, CH, H-6, J = 7.7 Hz), 7.84 (šs, 1H, NH), 7.94 (d, 1H, CH, H-5, J = 8.5 Hz), 8.32 (d, 1H, CH, H-8, J = 8.6 Hz). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 24.8 (C-8´), 25.0 (C-1), 33.5 (C-4), 44.0 (C-3´, 7´), 48.0 (C-2´), 65.1 (OCH₃), 107.6 (C-7´´), 108.8 (C-3´´), 116.0 (C-9a), 118.6 (C-4´´), 118.9 (C- 5´´), 120.1 (C-8a), 120.9 (C-2´´), 121.8 (C-6´´), 123.1 (C-8), 123.5 (C-3a´´), 124.3 (C-7), 127.8 (C-6), 128.3 (C-5), 132,0 (C-7a´´), 146.9 (C-5a), 150.0 (C-9), 157.8 (C-4a), 181.5 (C = S). IR (KBr, selected bands): 3055, 2931, 2858, 2363, 2342, 1560, 1502, 1435, 1386, 1292, 1136, 759, 741. Calcd for C₂₇H₃₁N₅OS (473.63): C, 68.47; H, 6.60; N, 14.79. Found: C, 68.40; H, 6.59; N, 14.76.

1-[2-(1-methoxyindol-3-yl)ethyl]-3-[3-(1,2,3,4-tetrahydroacridin-9-ylamino)propyl] thiourea (4b)

Yield 57%; yellow solid; Mp = 84–85°C; ¹H-NMR (400 MHz, DMSO): δ 1.70–1.87 (m, 4H, 2 × CH₂, H-2, 3), 2.73 (t, 2H, CH₂, H-1, J = 6.0 Hz), 2.84–2.93 (m, 4H, 2 × CH₂, H-4, 9´), 3.40–3.46 (m, 4H, 2 × CH₂, H-2´, 4´), 3.57–3.68 (m, 2H, CH₂, H-8´), 4.00 (s, 3H, OCH₃), 5.45 (t, 1H, NH, H-1´, J = 6.0 Hz), 7.0 (t, 1H, CH, H-5´´, J = 7.6 Hz), 7.15–7.21 (m, 1H, CH, H-6´´), 7.30–7.36 (m, 1H, CH, H-7), 7.37–7.43 (m, 2H, 2 × CH, H-2´´, 7´´), 7.48–7.53 (m, 1H, CH, H-6), 7.61–7.65 (m, 1H, CH, H-4´´), 7.70 (d, 1H, CH, H-5, J = 8.8 Hz), 8.12 (d, 1H, CH, H-8, J = 8.4 Hz). ¹³C-NMR (DMSO): δ 22.4, 22.8 (C-2, 3), 24,6 (C-9´), 25.2 (C-1), 30.5 (C-3´), 33.5 (C-4), 44.0 (C-8´), 45.3 (C-2´, 4´), 65.5 (OCH₃), 108.1 (C-7´´), 108.9 (C-3´´), 116.0 (C-9a), 119.2 (C-4´´), 119.3 (C- 5´´), 120.3 (C-8a), 121.7 (C-2´´), 122.2 (C-6´´), 123.0 (C-8), 123.2 (C-7), 123.5 (C-3a´´), 127.8 (C-6), 128.3 (C-5), 132.0 (C-7a´´), 146.9 (C-5a), 150.1 (C-9), 157,9 (C-4a), 181.5 (C = S). IR (KBr, selected bands): 3054, 2928, 2856, 2365, 1560, 1500, 1450, 1294, 758, 740. Calcd for C₂₈H₃₃N₅OS (627.88): C, 68.96; H, 6.82; N, 14.36. Found: C, 68.90; H, 6.80; N, 14.30.

1-[2-(1-methoxyindol-3-yl)ehtyl]-3-[4-(1,2,3,4-tetrahydroacridin-9-ylamino)butyl] thiourea (4c)

Yield 46%; yellow solid; Mp = 66–67°C; ¹H-NMR (400 MHz, DMSO): δ 1.43–1.6 (m, 4H, 2 × CH₂, H-3´, 4´), 1.73–1.88 (m, 4H, 2 × CH₂, H-2, 3), 2.72 (t, 2H, CH₂, H-1, J = 6.4 Hz), 2.84–2.93 (m, 4H, 2 × CH₂, H-4, 10´), 3.30–3.45 (m, 4H, 2 × CH₂, H-2´, 5´), 3.57–3.68 (m, 2H, CH₂, H-9´), 4.00 (s, 3H, OCH₃), 5.40 (t, 1H, NH, H-1´, J = 6.4 Hz), 7.03 (t, 1H, CH, H-5´´, J = 7.6 Hz), 7.15–7.20 (m, 1H, CH, H-6´´), 7.33 (dd, 1H, CH, H-7, J = 1.2; 6.8 Hz), 7.37–7.43 (m, 2H, 2 × CH, H-2´´, 7´´), 7.51 (dd, 1H, CH, H-6, J = 1.2; 6.8 Hz), 7.62–7.65 (m, 1H, CH, H-4´´), 7.70 (dd, 1H, CH, H-5, J = 1.2; 8.8 Hz), 8.11 (d, 1H, CH, H-8, J = 8.4 Hz). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 24.8 (C-10´), 25.1 (C-1), 28.1 (C-3´, 4´), 33.5 (C-4), 44.0 (C-9´), 47.7 (C-2´, 5´), 65.5 (OCH₃), 108.1 (C-7´´), 108.9 (C-3´´), 116.0 (C-9a), 119.2 (C- 4´´, 5´´), 120.3 (C-8a), 121.7 (C-2´´), 122.1 (C-6´´), 123.0 (C-8), 123.2 (C-7), 123.5 (C-3a´´), 127.7 (C-6), 128.3 (C-5), 132.0 (C-7a´´), 146.9 (C-5a), 150.2 (C-9), 157.9 (C-4a), 181.5 (C = S). IR (KBr, selected bands): 3056, 2932, 2858, 1560, 1500, 1450, 1353, 1294, 759, 741. Calcd for C₂₉H₃₅N₅OS (501.69): C, 69.43; H, 7.03; N, 13.96. Found: C, 69.40; H, 7.01; N, 13.90.

1-[2-(1-methoxyindol-3-yl)ethyl]-3-[5-(1,2,3,4-tetrahydroacridin-9-ylamino)pentyl] thiourea (4d)

Yield 71%; yellow solid; Mp = 55–56°C; ¹H-NMR (400 MHz, DMSO): δ 1.20–1.35 (m, 2H, CH₂, H-4´), 1.34–1.45 (m, 2H, CH₂, H-5´), 1.50–1.62 (m, 2H, CH₂, H-3´), 1.74–1.87 (m, 4H, 2 × CH₂, H-2, 3), 2.65–2.77 (m, 2H, CH₂, H-1), 2.83–2.93 (m, 4H, 2 × CH₂, H-4, 11´), 3.22–3.45 (m, 4H, 2 × CH₂, H-2´, 6´), 3.55–3.70 (m, 2H, CH₂, H-10´), 4.0 (s, 3H, OCH₃), 5.34 (t, 1H, NH, H-1´, J = 6.0 Hz), 7.03 (t, 1H, CH, H-5´´, J = 7.6 Hz), 7.14–7.20 (m, 1H, CH, H-6´´), 7.30–7.36 (m, 1H, CH, H-7), 7.37–7.45 (m, 2H, 2 × CH, H-2´´, 7´´), 7.47–7.54 (m, 1H, CH, H-6), 7.61–7.67 (m, 1H, CH, H-4´´), 7.70 (d, 1H, CH, H-5, J = 8.4 Hz), 8.10 (d, 1H, CH, H-8, J = 8.4 Hz). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 24.2 (C-4´), 25.1 (C-11´), 25.5 (C-1), 28.9 (C-5´), 30.8 (C-3´), 34.0 (C-4), 44.0 (C-10´), 44.5 (C-6´), 48.4 (C-2´), 65.9 (OCH₃), 108.5 (C-7´´), 109.4 (C-3´´), 116.0 (C-9a), 119.7 (C- 4´´–6´´), 120.2 (C-8a), 122.2 (C-2´´), 123.2 (C-7), 123.5 (C-3a´´), 123.6 (C-8), 128.2 (C-6), 128.7 (C-5), 132.2 (C-7a´´), 147.0 (C-5a), 150.7 (C-9), 158.0 (C-4a), 181.3 (C = S). IR (KBr, selected bands): 3056, 2932, 2857, 1560, 1500, 1450, 1352, 1293, 758, 740. Calcd for C₃₀H₃₇N₅OS (515.71): C, 69.87; H, 7.23; N, 13.58. Found: C, 69.90; H, 7.26; N, 13.50.

1-[2-(1-methoxyindol-3-yl)ethyl]-3-[6-(1,2,3,4-tetrahydroacridin-9-ylamino)hexyl] thiourea (4e)

Yield 69%; yellow solid; Mp = 58–59°C; ¹H-NMR (400 MHz, DMSO): δ 1.2–1.35 (m, 4H, 2 × CH₂, H-4´, 5´), 1.35–1.49 (m, 2H, CH₂, H-6´), 1.50–1.60 (m, 2H, CH₂, H-3´), 1.75–1.87 (m, 4H, 2 × CH₂, H-2, 3), 2.66–2.76 (m, 2H, CH₂, H-1), 2.85–2.93 (m, 4H, 2 × CH₂, H-4, 12´), 3.25–3.42 (m, 4H, 2 × CH₂, H-2´, 7´), 3.58–3.70 (m, 2H, CH₂, H-11´), 4.0 (s, 3H, OCH₃), 5.34 (t, 1H, NH, H-1´, J = 6.4 Hz), 7.04 (t, 1H, CH, H-5´´, J = 7.6 Hz), 7.15–7.21 (m, 1H, CH, H-6´´), 7.30–7.36 (m, 1H, CH, H-7), 7.37–7.43 (m, 2H, 2 × CH, H-2´´, 7´´), 7.48–7.54 (m, 1H, CH, H-6), 7.62–7.66 (m, 1H, CH, H-4´´), 7.70 (d, 1H, CH, H-5, J = 8.4 Hz), 8.10 (d, 1H, CH, H-8, J = 8.0 Hz). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 24.6 (C-12´), 25.0 (C-1), 26.0, 26.1 (C-4´, 5´), 28.6 (C-6´), 30.6 (C-3´), 33.5 (C-4), 43.2 (C-7´), 44.0 (C-11´), 47.9 (C-2´), 65.5 (OCH₃), 108.1 (C-7´´), 108.9 (C-3´´), 116.0 (C-9a), 119.2 (C- 4´´, 5´´), 120.2 (C-8a), 121.7 (C-6´´), 122.1 (C-2´´), 123.0 (C-8), 123.2 (C-7), 123.5 (C-3a´´), 127.7 (C-6), 128.3 (C-5), 132.0 (C-7a´´), 146.9 (C-5a), 150.3 (C-9), 158.0 (C-4a), 181.5 (C = S). Calcd for C₃₁H₃₉N₅OS (529.74): C, 70.29; H, 7.42; N, 13.22. Found: C, 70.20; H, 7.40; N, 13.20.

1-[2-(1-methoxyindol-3-yl)ethyl]-3-[7-(1,2,3,4-tetrahydroacridin-9-ylamino)heptyl] thiourea (4f)

Yield 74%; yellow solid; Mp = 56–57°C; ¹H-NMR (400 MHz, DMSO): δ 1.15–1.30 (m, 6H, 3 × CH₂, H-4´, 6´, 7´), 1.35–1.47 (m, 2H, CH₂, H-5´), 1.50–1.60 (m, 2H, CH₂, H-3´), 1.74–1.87 (m, 4H, 2 × CH₂, H-2, 3), 2.66–2.72 (m, 2H, CH₂, H-1), 2.83–2.92 (m, 4H, 2 × CH₂, H-4, 13´), 3.25–3.45 (m, 4H, 2 × CH₂, H-2´, 8´), 3.58–3.70 (m, 2H, CH₂, H-12´), 4.0 (s, 3H, OCH₃), 5.32–5.41 (m, 1H, NH, H-1´), 7.0–7.10 (m, 1H, CH, H-5´´), 7.14–7.22 (m, 1H, CH, H-6´´), 7.30–7.46 (m, 3H, 3 × CH, H-7, 2´´, 7´´), 7.46–7.54 (m, 1H, CH, H-6), 7.60–7.73 (m, 2H, 2 × CH, H-5, 4´´), 8.10 (d, 1H, CH, H-8, J = 8.0 Hz). ¹³C-NMR (DMSO): δ 22.4, 22.7 (C-2, 3), 25.0 (C-13´), 25.1 (C-1), 26.3, 29.0 (C-4´–7´), 30.5 (C-3´), 33.5 (C-4), 44.3 (C-8´), 47.9 (C-2´), 65.5 (OCH₃), 108.1 (C-7´´), 108.9 (C-3´´), 115.8 (C-9a), 119.2 (C- 4´´, 5´´), 120.2 (C-8a), 121.7 (C-2´´), 122.2 (C-6´´), 123.0 (C-7, 8), 123.5 (C-3a´´), 127.8 (C-6), 128.3 (C-5), 132.0 (C-7a´´), 147.0 (C-5a), 150.3 (C-9), 158.0 (C-4a), 181.5 (C = S). Calcd for C₃₂H₄₁N₅OS (543.76): C, 70.68; H, 7.60; N, 12.88. Found: C, 70.59; H, 7.59; N, 12.84.

In vitro anticholinesterase assay

The inhibitory activities of prepared compounds 3, 4 and standards (tacrine, 7-methoxytacrine) against human recombinant AChE (E.C. 3.1.1.7) and human plasma BChE (E.C. 3.1.1.8) were measured by modified Ellman's method [36–38] and expressed as IC₅₀ (the concentration of the inhibitor required to decrease the activity of cholinesterase by 50%). Human recombinant AChE, phosphate buffer (PB) with a pH of 7.4, 5,5′-dithio-bis(2-nitrobenzoic) acid (Ellman's reagent, or DTNB), acetylthiocholine, butyrylthiocholine and other used compounds were purchased from Sigma-Aldrich (Prague, Czech Republic). Human plasma was used as a source of BChE and was prepared from heparinized human blood. Blood was centrifuged for 20 min at 4°C and 2300×g using a Universal 320R centrifuge (Andreas Hettich GmBH, DE, USA). Plasma was separated and stored at -80°C. Enzyme activity was measured in 96-well polystyrene microplates (Thermo Fisher Scientific, MA, USA). Stock solutions of cholinesterases in PB were diluted up to a final activity of 0.002 U/μl. The assay was performed in a volume of 100 μl and included the following: cholinesterase (10 μl), 0.01 M DTNB (20 μl) and 0.1 M PB (40 μl). Tested compounds (10 μl of different concentrations) were preincubated for 5 min in the assay medium, and solution of the substrate (20 μl of 0.01 M acetylthiocholine or butyrylthiocholine iodide solution) was added to start the reaction. The increase of absorbance was measured at 412 nm using the Synergy 2 multimode microplate reader (BioTek Instruments Inc., VT, USA). The percentage of inhibition was calculated using the following formula:

$$I=\left(1-\frac{\Delta A_i}{\Delta A_0}\right)\times 100$$

where ΔA_i indicates the change in absorbance in the presence of inhibitor and ΔA₀ indicates the change in absorbance when a solution of PB was added instead. The data were analyzed using Excel (Microsoft Corporation, WA, USA) and Prism 6.07 (GraphPad Software, CA, USA) for Windows (Microsoft Corporation).

Kinetic study of cholinesterase inhibition

A kinetic study of AChE and BChE inhibition was performed as described earlier using the modified Ellman's method. The values of V_max and K_m (from Michaelis–Menten kinetics) as well as K_i and K_i’ were determined by nonlinear regression analysis of the substrate velocity curves. Linear regression was employed to produce Lineweaver–Burk plots. All calculations were performed using Prism 6.07 (GraphPad Software) for Windows (Microsoft Corporation).

Inhibition of Aβ-dependent amyloid nucleation

The authors employed two haploid [psi^-] strains of different genotypic backgrounds designated GT17 [39] and GT409 [40]. They were transformed with a plasmid containing an HIS3 marker and expressing the chimeric construct SUP35N-Aβ42 from the galactose-inducible P_GAL promoter. Nucleation of the [PSI⁺] prion was detected by growth on a synthetic medium lacking adenine (–Ade) because of translational readthrough (nonsense suppression) of the ade1-14 (UGA) reporter allele as a result of partial inactivation of the translational termination activity of the Sup35 release factor in prion form [41]. Yeast cultures were grown overnight in a standard synthetic liquid medium with glucose lacking histidine (–His), washed and inoculated into –His medium containing 2% galactose instead of glucose for induction of P_GAL promoter. Compound 3c (140 μM) or solvent control (70% EtOH:DMSO in a 1:2.18 v/v ratio) was added at a starting cell density of OD₆₀₀ = 0.1. After a period of 16–36 h, as indicated, dilutions were either spotted or plated onto solid media with glucose lacking either only histidine (–His), for the detection of plasmid-containing cells, or both histidine and adenine (−His/−Ade), for the detection of cells with the [PSI⁺] prion; in some cases, spotting or plating onto complete organic yeast extract–peptone–dextrose medium, where all cell grow was performed in parallel. Concentrations of plasmid-containing cells were determined from numbers of colonies grown on –His medium (as detected after 3–4 days of incubation), whereas concentrations of plasmid-containing cells with the [PSI⁺] prion were determined from numbers of colonies grown on −His/−Ade medium (as detected after 10–14 days of incubation). The frequency of prion nucleation was determined as a ratio between the concentration of [PSI⁺] cells and the concentration of plasmid-containing cells (amyloid, nucleation frequency [ANF]).

DNA thermal denaturation assay

DNA thermal denaturation assay was performed as previously described [42]. Calf thymus DNA (Sigma-Aldrich, MO, USA) at 2.5 mg/ml in a buffer solution (10 mM Tris, 5 mM NaCl, 0.5 mM EDTA, pH 7.5) was reacted with 3c or solvent control for 48 h at 37°C. The molar ratio of 3c to DNA (per 1 bp) was 1:10. Reaction mixtures were used directly for melting analysis. Melting curves were recorded using a Cary 1E UV-visible spectrophotometer (Varian Medical Systems, Inc., CA, USA) with a Peltier temperature controller. Sample absorbance was measured in 1-cm path-length cells at 260 nm every 0.2°C as the temperature changed from 25°C to 95°C at a rate of 0.5°C/min. The temperature-dependent absorbance data were smoothed using the Loess method with four-point averages and a second-order polynomial. The first derivative melting curve (dA[T]/dT) was computed. Three independent melting experiments gave reproducible transitions.

Determination of in vitro BBB permeation

The parallel artificial membrane permeability assay (PAMPA) was used as an in vitro assay predicting BBB penetration. PAMPA was performed in coated 96-well membrane filters as previously described [43]. Briefly, the filter membrane of the donor plate was coated with polar brain lipid (Avanti Polar Lipids, AL, USA) in dodecane (4 μl of 20 mg/ml polar brain lipid in dodecane), and the acceptor well was filled with 300 μl of the phosphate-buffered saline buffer (pH 7.4, V_A). Tested compounds were dissolved first in the DMSO/phosphate-buffered 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. The concentration of DMSO did not exceed 0.5% v/v in the donor solution. The donor solution (300 μl, V_D) was added to the donor wells, and the donor filter plate was carefully put on the acceptor plate so that the coated membrane was ‘in touch’ with both donor solution and acceptor buffer. Test compounds were allowed to diffuse from the donor wells, through the polar brain lipid membranes (area = 0.28 cm²), to the acceptor wells. Concentrations of tested compounds in the donor and acceptor wells were determined by Synergy HT spectrophotometry (BioTek Instruments Inc.) at the maximum absorption wavelength of each compound after 3, 4, 5 and 6 h of incubation. The assays were performed in quadruplicate. Solutions at the theoretical equilibria were also prepared for each tested compound (i.e., the theoretical concentrations of the donor and acceptor compartments were simply combined) and measured as previously described. Concentrations in the donor and acceptor wells and equilibrium concentrations were calculated from the standard curves, and the permeability coefficient was calculated using the following formula [43]:

$$Pe=C\times -ln\left(1-\frac{{[drug]}_{acceptor}}{{[drug]}_{equilibrium}}\right)$$

where

$$C=\left(\frac{V_D\times V_A}{(V_D+V_A)\times Area \times T ime }\right)$$

Molecular modeling studies

Ligand preparation

The 3D models of compounds 3c and 4d were prepared within a building option in ACD/ChemSketch [44]. Optimization of model geometries was done with the Chimera software option using a minimization module [45].

Receptor preparation

Coordinates for the enzymes were obtained from the Protein Data Bank (PDB) database: PDB identifier (ID): 1B41 for hAChE and PDB ID: 1P0I for hBChE. The coordinates of missing enzyme residues were assigned using Modeller 9.11 within the Model/Refine Loops options in Chimera software [46]. Only nonterminal missing parts of the enzymes' x-ray structures were modeled. These enzyme structures were used for the next docking and molecular dynamic simulations.

Docking runs

AutoDock MGLTools 1.4.5 (revision 30) was used for preparation of the input files, which included the following: removal of nonenzymatic molecules and duplicate amino acid residues from the structures of enzymes, addition of hydrogen atoms and application of Gasteiger partial atomic charges for enzymes and ligands. Docking simulations were performed using AutoDock 4.2 software (revision 4.2.6) [47–49]. For the docking simulations within hAChE, the energy grid was set at coordinates x 116.4, y 104.3, z −130.6 with dimensions 60 × 60 × 60 points. For hBChE, the energy grid was set at coordinates x 137.8, y 122.7, z 38.8 with dimensions 60 × 60 × 60 points. Flexible ligand docking was employed for all ligand bonds, except for the thiourea NH–CS moieties, which were set as nonrotatable s-cis/s-trans conformations. A Lamarckian genetic algorithm was used for the docking runs, with spacing of 0.375 Å for both enzymes. A population of random ligand conformations in random orientation and at random translation was applied in each docking simulation. Docking simulations were derived from 100 different runs, terminating after a maximum of 5,000,000 energy evaluations or 27,000 generations. Population size was set at 300, and other parameters were applied as default.

Scoring of final structures

The final pose of the conformers with the lowest binding energy was chosen for the preparation of the interaction complex ligand – enzyme that was used in the next molecular dynamics simulation. The s-cis/s-trans thiourea moiety conformation of ligand 3c and the s-trans/s-cis conformer of ligand 4d were chosen. Pictures were prepared using Chimera software [45].

Molecular dynamics simulations

Ternary complex preparation

All molecular dynamics simulations were carried out in NAMD 2.13 using a generalized born implicit solvent for waters, parm99.dat parameters set for proteins (Amber FF14SB) and GAFF atom types for ligands [50–54]. The antechamber and xleap modules of the AmberTools 18 software package were applied to extrapolate missing ligand force field parameters and to derive charges using the AM1-BCC method.

Molecular dynamics production run

A ligand–receptor ternary complex was minimized using 1000 steps of the conjugate gradient. The system was then equilibrated at 300 K for 1 ns with harmonic restraints on nonhydrogen atoms, continuously decreasing in five steps, each 0.2 ns, with force constants of 50, 10, 5, 2 and 0.5 kcal mol^-1 Å^-2. The final production run was done at 300 K for 5 ns. Energy information, averages and coordinates were recorded every 500 steps, and a nonbonded list was updated every ten steps. The root-mean-square deviation (RMSD) of the ligand–enzyme complex was examined using an RMSD calculator extension module in the VMD 1.8.6 software package [55]. The hAChE binding pocket used for RMSD calculation consisted of amino acid residue numbers 72, 73, 74, 75, 76, 80, 82, 83, 86, 122,124,125, 202, 203, 284, 286, 288, 289, 296, 294, 295, 297, 337, 342, 343, 344, 447 and 449. The hBChE binding pocket used for RMSD calculation consisted of amino acid residue numbers 67, 68, 69, 70, 73, 74, 76, 78, 79, 81, 82, 83, 84, 88, 115, 116, 117, 120, 121, 122, 125, 128, 328, 329 and 331. A NAMD energy module within the VMD software was applied to calculate the interaction energy of the ligand–enzyme complex. The same software was used to evaluate a trajectory of the production run for the ligand–enzyme complex. Pictures were prepared using Chimera 1.13.1 software (build 41965) [45]. Maestro software (Schrödinger LLC, NY, USA) was used to prepare a ligand interaction diagram [56].

Results & discussion

Chemistry

Two series of tacrine–indole heterodimers 3 and 4 were synthesized by the reaction of indol-3-yl)ethyl isothiocyanates 2a and b with N-(1,2,3,4-tetrahydroacridin-9-yl)alkanediamines 1a–g in the presence of TEA in CH₂Cl₂ (Scheme 1). Initially, the reaction of 9-chloro-1,2,3,4-tetrahydroacridine with the corresponding α,ω-diaminoalkane produced N-(1,2,3,4-tetrahydroacridin-9-yl)alkanediamines 1a–g [57]. In addition, 3-(2-isothiocyanatoethyl)-1H-indole (2a) was prepared from tryptamine by reaction with CS₂, followed by the addition of Boc₂O and DMAP [58]. (1-substituted indol-3-yl)ethyl isothiocyanate 2b was obtained in 60% yield by treatment with 1-methoxytryptamine [59], freshly prepared by reduction of 1-methoxy-3-(2-nitrovinyl)indole [60], followed by treatment with thiophosgene in CH₂Cl₂/H₂O in the presence of CaCO₃ [61]. Final tacrine–indole heterodimers 3 and 4 were synthesized using the appropriate 3-(2-isothiocyanatoethyl)-1H-indole 2a, b and N-(1,2,3,4-tetrahydroacridin-9-yl)alkanediamines 1a–g in the presence of TEA in CH₂Cl₂ (Figure 2). All final compounds were characterized by ¹H, ¹³C-NMR, IR spectra and CHN analysis.

Figure 2. . Synthetic procedure for tacrine–indole heterodimers 3a–g and 4a–f.

Evaluation of AChE & BChE inhibitory activity

To determine the potential of tacrine–indole heterodimers for AD treatment, the authors evaluated their anticholinesterase profile (Table 1). The IC₅₀ values ranged from 160 to 25 nM and 160 to 39 nM for hAChE and hBChE, respectively. Of the most potent inhibitors of AChE 3c and 4d, compound 3c, with an IC₅₀ value of 25 nM, displayed approximately 20-fold higher potency than 7-methoxytacrine. Furthermore, 3c displayed considerably higher selectivity for AChE relative to human plasma BChE compared with compound 4d, which was the most potent BChE inhibitor from the tacrine–indole family (selectivity index: IC₅₀ [BChE]/IC₅₀ [AChE] = 3 and 0.6 for 3c and 4d, respectively). With regard to tether length, it cannot clearly determine any structure–activity relationship.

Table 1. . In vitro hAChE and hBChE inhibitory activity of tacrine–indole heterodimers 3a–g and 4a–f, reference compounds and selectivity index.

Compound	No.	R	IC50 ± SEM†(nM) hAChE	IC50 ± SEM†(nM) hBChE	Selectivity for hAChE‡
Tacrine			500 ± 100	23 ± 4	0.05
7-MEOTA			15,000 ± 2900	21,000 ± 3400	1.4
3a	2	H	160 ± 8.0	79 ± 2.5	0.5
3b	3	H	77 ± 2.0	85 ± 1.6	1.1
3c	4	H	25 ± 0.6	76 ± 2.4	3.0
3d	5	H	58 ± 2.2	71 ± 2.4	1.2
3e	6	H	76 ± 1.1	79 ± 1.2	1.0
3f	7	H	77 ± 4.2	110 ± 4	1.4
3g	8	H	49 ± 2.7	99 ± 4.9	2.0
4a	2	OMe	130 ± 8.0	160 ± 6.0	1.2
4b	3	OMe	71 ± 2.3	80 ± 3.8	1.1
4c	4	OMe	80 ± 1.7	58 ± 2.1	0.7
4d	5	OMe	70 ± 1.9	39 ± 1.8	0.6
4e	6	OMe	70 ± 1.4	57 ± 2.0	0.8
4f	7	OMe	99 ± 4.5	81 ± 2.1	0.8

^†Results are expressed as the mean of at least three experiments.

^‡Selectivity for hAChE is determined as the ratio IC₅₀ (hBChE)/IC₅₀ (hAChE).

hAChE: Human acetylcholinesterase; hBChE: Human butyrylcholinesterase; SEM: Standard error of the mean; 7-MEOTA: 7-methoxytacrine.

Compounds 4a–f were designed by introducing a substituent on indole nitrogen (R = OMe). The presence of a methoxy group did not significantly affect inhibitory activity against hAChE except for 4a and 4b bearing ethylene and propylene bridges, respectively, between tacrine and 1-methoxyindole moieties. Regarding inhibition of hBChE, the methoxy group on indole nitrogen enhanced the inhibitory activity of 4c–e with a chain length of n = 4–6.

The inhibitory activity of tacrine–tacrine homodimers (Figure 1) with thiourea linkers was reported in the authors' previous work [34]. Comparison of IC₅₀ values in the class of these compounds with tacrine–indole hybrids 3 and 4 revealed that three factors can greatly influence inhibition potential; namely, length of methylene spacer, presence of tacrine/indole unit and substitution at the 1-position of the indole moiety. Specifically, it was observed that the four-methylene spacer appeared to best suit hAChE inhibition (3c), whereas the six-methylene bridge conferred the best inhibition to hBChE (3e). Second, the tacrine fragment was crucial for retaining a high cholinesterase inhibition profile since its introduction to indole moiety reduced activity for both cholinesterases, especially BChE. Furthermore, introduction of the methoxy substituent on the indole nitrogen increased AChE and BChE inhibitory activity with regard to the two-methylene analogue (4a), but for derivatives 4c and e, with several four- and six-methylene groups, respectively, the inhibitory activity against both AChE and BChE decreased or did not change.

Kinetic study of AChE & BChE inhibition

To evaluate the mode of interaction of the heterodimer 3c with hAChE and hBChE, a kinetic study was performed. Inhibition kinetics were elucidated from enzyme velocity curves that were determined at several concentrations of the tested inhibitor and corresponding substrates. The mode of enzyme inhibition and corresponding kinetic parameters (K_i and K_i') were determined using nonlinear regression analysis. Results for each type of inhibition (competitive, noncompetitive, uncompetitive and mixed) were compared using the sum-of-squares F test. Statistical analysis found a mixed-mode AChE inhibition and a competitive-type BChE inhibition (p < 0.05). This is consistent with the Lineweaver–Burk plot used for graphical presentation of the data (Figure 3).

Figure 3. . Steady-state inhibition of AChE and BChE substrate hydrolysis by compound 3c at different concentrations.

Lineweaver–Burk plots of initial velocity at increasing substrate concentrations (AChE: 0.1563–1.250 mM; BChE: 2.5–20.0 mM) are presented. Lines were derived from a linear regression of the data points. Heterodimer 3c is a mixed-mode inhibitor of AChE and a competitive inhibitor of BChE.

AChE: Acetylcholinesterase; BChE: Butyrylcholinesterase.

The crossing of lines for inhibitor 3c was placed above the x-axis for hAChE, which shows a reversible binding mode of mixed type. In this type of inhibition, the inhibitor binds to both free enzyme and enzyme–substrate complex. Apparent V_max decreased with increasing concentrations of 3c, and the K_m value remained unchanged, providing evidence for affinity to both free hAChE and its complex with substrate. Compound 3c bound with higher affinity to free AChE (K_i < K_i'), and apparent V_max was reduced at higher concentrations of the inhibitor, whereas K_m was slightly increased. The inhibitor interacted with the enzyme's allosteric PAS, which caused conformational changes of the cholinesterase, yielding a change in its active site.

The crossing of lines was placed on the y-axis for hBChE, which shows a reversible binding mode to the active site of the enzyme. With growing concentrations of the inhibitor, apparent V_max remained unchanged and Km increased. A K_i of 20.7 ± 6.6 nM and K_i' of 21.9 ± 4.9 nM were measured for 3c on AChE, and a K_i of 20.4 ± 4.4 nM was determined for BChE.

Inhibition of Aβ-dependent amyloid nucleation

Nucleation is postulated to be an early step in the formation of amyloidogenic fibrils associated with numerous amyloid/prion diseases. The ability of compound 3c to inhibit amyloid nucleation was evaluated using the yeast-based prion nucleation assay [62].

The yeast assay is designed to detect the amyloid-nucleating abilities of the mammalian protein fused to the N-terminal prion domain (Sup35N) of the yeast prion protein Sup35. Sup35 protein is a yeast counterpart of eukaryotic release factor 3; conversion of Sup35 into the amyloid form leads to the formation of the yeast prion [PSI⁺] [41]. In strains containing the amyloid (prion) form of another protein (e.g., [PIN⁺], a prion form of Rnq1), Sup35 can convert into the amyloid form through heterologous cross-seeding, most efficiently if Sup35 or Sup35N is present at high levels [1]. However, such conversion does not occur in strains lacking any preexisting prions (designated [psi⁻ pin⁻]). With regard to a mammalian amyloidogenic protein (e.g., human Aβ₄₂ peptide, which is associated with AD) attached to Sup35N, such a chimeric construct is able to nucleate the [PSI⁺] prion in the absence of any preexisting prions. It has also been proven that the amyloid-nucleating ability in yeast cells is strictly dependent on the amyloidogenic properties of Aβ peptide [63].

Yeast culture lacking any preexisting prion ([psi⁻ pin⁻]) was transformed with the plasmid bearing the chimeric P_GAL-SUP35N-Aβ42 construct [62]. Plasmid-containing cells were grown on galactose medium, where P_GAL promoter was induced in either the presence or absence of 3c (140 μM), and plated onto glucose medium, where P_GAL promoter was repressed, lacking either only histidine (–His), for determining the concentration of plasmid-containing cells, or both histidine and adenine (–His/–Ade), allowing for the detection of Ade⁺ colonies resulting from the nucleation of the amyloid form of Sup35 protein, [PSI⁺].

The culture treated with 3c displayed reduced formation of His⁺ Ade⁺ colonies in comparison with the solvent control-treated culture (Figures 4A–C & 5). This decrease indicated inhibition of Aβ-dependent amyloid nucleation. The result was confirmed for two yeast strains of different phenotypic origins treated with 3c for two different periods of time, 36 h (Figure 4) and 16 h (Figure 5), respectively. Statistical significance of the differences shown in Figure 4 was confirmed by the t-test (two-tailed p = 0.0337).

Figure 4. . Semiquantitative detection of the inhibition of amyloid nucleation by 3c in the yeast assay.

Cells of the [psi⁻ pin⁻] strain GT17 expressing the chimeric Sup35N-Aβ42 construct from the galactose-inducible P_GAL promoter were grown in galactose medium in the presence of either compound 3c or solvent control for 36 h. (A) Serial tenfold dilutions of 3c-treated (upper line) and solvent-treated control (lower line) cultures were spotted on plasmid-selective medium (–His), complete organic medium (YPD) or plasmid-selective medium lacking adenine (–His/–Ade), which allowed for the detection of [PSI⁺] colonies originating from cells in which nucleation of an amyloid had occurred. (B & C) Equal dilutions of cultures treated with solvent or 3c were plated onto –His medium (1), −His/−Ade medium (2) and YPD medium (3) or velveteen replica-plated from −His medium to −His/−Ade medium (4).

YPD: Yeast extract–peptone–dextrose.

Figure 5. . Quantitative detection of the inhibition of amyloid nucleation by 3c in the yeast assay.

Cells of the [psi⁻ pin⁻] strain GT17 expressing the chimeric Sup35N-Aβ42 construct from the galactose-inducible P_GAL promoter were grown in galactose medium in the presence of either compound 3c or solvent control for 16 h, followed by plating cells onto –His and –His/–Ade media (as in Figure 4B & C). ANF was determined as the ratio of His⁺ Ade⁺ cell concentration to total His⁺ cell concentration. Means and standard deviations are shown for three biological replicates (p-value as per two-tailed t-test).

ANF: Amyloid nucleation frequency.

Extracellular deposits from aggregated Aβ peptide are a histopathological hallmark of AD. Aβ peptide aggregation is a critical step in the formation of neurotoxic oligomeric species, protofibrils and fibrils; therefore, its inhibition is a promising strategy for preventive or therapeutic pharmacological interventions [63]. However, drugs currently approved for the treatment of AD, which include inhibitors of AChE and an antagonist of NMDA receptors, do not target this process and fail to slow disease progression. As a result, new therapeutics that would modify critical pathological steps leading to AD are needed [64].

The authors' results (Figures 4 & 5) show that in the yeast assay, 3c inhibits nucleation of Sup35 amyloids, which in this particular experimental design specifically depends on the amyloidogenic properties of Aβ peptide attached to the yeast protein [62]. Thus, 3c is likely to inhibit an early triggering step in the Aβ cascade. Further experiments are needed to determine if 3c physically interferes with amyloid nucleation and whether or not this effect is specific to Aβ. However, even if 3c possesses a general antinucleation effect applicable to various amyloids, it does not invalidate its potential anti-Alzheimer's properties.

The authors' findings are especially promising in this regard because 3c is also a potent AChE inhibitor. Although this property of 3c cannot be tested in the yeast assay, it suggests that 3c can potentially exert beneficial dual action against AD both as an agent preventing disease development and as a symptom-ameliorating agent.

DNA thermal denaturation assay

Acridine derivatives are known to target DNA through intercalation [65], which results in distortion of dsDNA following the insertion of intercalating agents between the stacked DNA base pairs. Intercalating agents display mutagenic properties directly or indirectly through the facilitation of specific interactions of other chemically reactive molecules with DNA, which can lead to genotoxicity and carcinogenicity [66]. As a result, DNA intercalation can be, depending on the circumstances, an undesired property of pharmaceuticals. In this study, the authors tested the ability of 3c, the most potent inhibitor of AChE identified herein, to intercalate DNA by the method of DNA thermal denaturation assay. This assay measures temperature-dependent denaturation of dsDNA based on the hyperchromicity effect using absorbance spectroscopy at 260 nm. Melting temperature (Tm), at which 50% of the DNA becomes single-stranded, can be determined from the maximum of the first derivative of the DNA melting curves (dA260/dT vs temperature), where A260 is the absorbance at 260 nm. Intercalation results in an increase in Tm of DNA reacted with an intercalating agent compared with untreated DNA [67].

The authors' results demonstrated no difference between Tm values of calf thymus DNA reacted with 3c and unreacted control DNA (Figure 6) under the experimental conditions. Consequently, 3c does not seem to be a DNA intercalating agent or a compound inducing distortions of dsDNA via other types of interactions, which would be detectable through changes in Tm value. This result suggests the absence of mutagenicity and favorable toxicological properties of the tacrine–indole heterodimer 3c.

Figure 6. . First derivative of the melting curve of native CT DNA and CT DNA reacted with 3c.

Change in absorbance dA/dt is plotted against T. A comparison of these curves shows equal Tm values (81.5°C), which implies no interaction of 3c with CT DNA under the conditions of the experiment.

CT DNA: Calf thymus DNA; T: Temperature; Tm: Melting temperature.

In vitro BBB permeation

The BBB is one of the obstacles to delivering potential drugs to treat CNS diseases. The efficiency of studied tacrine–indole derivatives was predicted using PAMPA as an in vitro model of passive diffusion of the compound to the CNS [68]. All data obtained from PAMPA are listed in Table 2. Reference drugs with high BBB permeability and poor or no BBB permeability were used as positive and negative controls, respectively.

Table 2. . Permeability values from the PAMPA-BBB assay.

Ligand	Pe ± SEM (10-6 cm s-1)†	CNS predicted availability ‡
3a	6.65 ± 1.58	CNS+
3b	4.94 ± 1.15	CNS+
3c	4.44 ± 1.32	CNS+
3d	6.64 ± 0.69	CNS+
3e	9.18 ± 1.67	CNS+
3f	17.4 ± 3.96	CNS+
3g	31.2 ± 6.72	CNS+
4a	3.10 ± 0.17	CNS+/−
4b	6.70 ± 1.19	CNS+
4c	7.91 ± 1.59	CNS+
4d	9.09 ± 1.54	CNS+
4e	12.9 ± 3.19	CNS+
4f	12.9 ± 2.42	CNS+
Tacrine	6.0 ± 0.6	CNS+
Donepezil	21.9 ± 2.1	CNS+
Rivastigmine	20.0 ± 2.1	CNS+
Ibuprofen	18.0 ± 4.3	CNS+
Chlorothiazide	1.1 ± 0.5	CNS−
Furosemide	0.2 ± 0.07	CNS−
Ranitidine	0.04 ± 0.02	CNS−
Sulfasalazine	0.09 ± 0.05	CNS−

^†Results are expressed as the mean of at least three experiments ± SEM.

^‡Classification of the prediction to cross the BBB: CNS+, high BBB permeation predicted, with Pe (×10^-6 cm s^-1) >4.0; CNS−, low BBB permeation predicted, with Pe (×10^-6 cm s^-1) <2.0; CNS+/−, BBB permeation uncertain, with Pe (×10^-6 cm s^-1) from 4.0 to 2.0 [45].

BBB: Blood–brain barrier; PAMPA: Parallel artificial membrane permeability assay; Pe: Permeability coefficient; SEM: Standard error of the mean.

The tacrine–indole derivatives 3a–g and 4b–f, with permeability coefficient values over 4.0 × 10^-6 cm.s^-1, were considered ligands with a high probability of crossing the BBB. The tacrine–indole derivative 4a fell into a permeability coefficient interval ranging between 2.0 × 10^-6 and 4.0 × 10^-6 cm.s^-1, with uncertain BBB permeation.

Molecular modeling studies

Docking simulation was performed with compounds 3c and 4d, which were the most potent inhibitors of hAChE and hBChE. To allow sufficient space sampling, flexible ligand docking was done for the rotatable bonds of all ligands, whereas NH–CS–NH bonds were set as nonrotatable in the s-cis/s-trans conformation. Thus, four conformations were obtained for each ligand. Ligands with the trans-trans conformation were omitted from the next docking study because of a possible steric clash. Conformation of NH–CS–NH was set as rigid because when set as rotatable, docking simulation produced unrealistic results. To perform docking simulation, AutoDock software was used, and hAChE (PDB ID: 1B41) and hBChE (PDB ID: 1P0I) were used as receptors for molecular modeling. Modeller software was used to reconstruct missing parts of the enzymes' x-ray structures.

The docking runs for ligands 3c and 4d led to multiconformation binding clusters. Putative binding complexes with the lowest binding energy for each conformer of both ligands are depicted in Figure 7A & B & Supplementary Figures 5, 6, 12 & 13. Graphs of ligand interactions within an active gorge of enzymes are depicted in Supplementary Figures 9 & 16. Docking simulations proposed an interaction of the indole moiety of ligand 3c with the enzyme PAS as a main binding mode. However, this result should be interpreted carefully because the opposite direction of ligand 3c was also an outcome of docking simulation, although it was energetically less favorable.

Figure 7. . Molecular modeling visualizations.

(A) Top docking pose of derivative 3c, depicting its putative hydrogen bonds formed with amino acid residues in the active site gorge of hAChE. Conformations of nonrotatable bonds (N–S–N) are s-cis/s-trans. (B) Top docking pose of derivative 4d, depicting its putative hydrogen bonds formed with amino acid residues in the active site gorge of hBChE. Conformations of nonrotatable bonds (N–S–N) are s-trans/s-cis. (C & D) Color palette of atoms: C, gray; H, white; N, blue; O, red. Interaction energy (100 mA) during the simulation for the 3c–hAChE binding complex and the 4d–hBChE binding complex. Van der Waals energy is depicted in red, and electrostatic energy is depicted in blue.

hAChE: Human acetylcholinesterase; hBChE: Human butyrylcholinesterase.

To study the forces driving the formation of a ligand–enyzme complex, methods of molecular dynamics were employed. As an input geometry for ligand 3c, docking coordinates were used for a conformer with the overall lowest binding energy. In the case of ligand 4d, the ligand's coordinates were selected for the conformer with the second lowest binding energy because of the intramolecular ligand's loop formation. Molecular dynamics simulations were performed using a generalized born implicit solvent for water. An equilibrium run was performed at 300 K for 1 ns with harmonic restraints, and the final production run was done at 300 K for 5 ns. The result of the simulations was analyzed using the VMD software package.

Based on the change in the RMSD value of ligand 3c as well as amino acid residues in the active site of hAChE, an ongoing change in the active site during the simulation was clearly visible (Supplementary Figures 1 & 2). As depicted in Supplementary Figures 7 & 8, the position of ligand 3c within the active site of the enzyme varied, at 1 and 5 ns. These changes might be ascribed to a closely linked induction fit formation with no clear orientation of the indole core within a PAS. A graph of the ligand's interactions that stabilized the 3c–hAChE ligand binding complex is depicted in Supplementary Figures 10 & 11.

With regard to the hBChE–4d binding complex, the most pronounced changes in the RMSD value were connected to a rotation of the methoxy group of the ligand (Supplementary Figure 4). By contrast, the RMSD value for the binding pocket was relatively stable (Supplementary Figure 3). Putative conformations of ligand 4d within the active site of the enzyme varied, at 1 and 5 ns, as depicted in Supplementary Figures 14, 15, 17 & 18.

Based on the docking simulations and the molecular dynamics results, the authors inferred the multiconformation binding mode of ligands 3c and 4d within the active site of the cholinesterases. The authors' results suggested that the 3c–hAChE ligand binding complex was stabilized mainly by van der Waals forces (Figure 7C), and the same interpretation applied to the 4d–hBChE ligand binding complex (Figure 7D). It should be noted that this type of study should be optimally performed for more than two conformers for each ligand to obtain a better sampling of conformation space, but this was beyond the scope of this work.

Conclusion

Two series of new tacrine–indole ligands 3 and 4 were synthesized and evaluated for their biological activity. The results suggest that at least some of these compounds have the potential for future development as preventive and/or therapeutic agents against AD because of their ability to potently inhibit AChE and Aβ-dependent amyloid nucleation and cross the BBB as well as the absence of DNA intercalating activity that would be toxicologically unfavorable for these types of potential therapeutics.

Future perspective

Significant efforts have been made over the last three decades to identify the neurodegenerative processes behind AD [69]. Given the complexity of the disease and the fact that genetic and environmental factors can determine disease progression, single targeted therapies might be inadequate [70,71]. The high failure rate of AD drugs in clinical trials, together with alarming data related to not only the social impact but also the economic burden of this disease, has shifted the drug development paradigm to so-called multipotent or multitarget-directed ligands. Indeed, these compounds might better address the need to secure potent, disease-modifying therapy for AD. Our previous study identified new promising inhibitors from the group of tacrine–tacrine homodimers with a thiourea linker [34]. These compounds displayed an intriguing profile of activity in vitro as low or even subnanomolar scale inhibitors of cholinesterases. Inspired by these findings, we followed the success story by designing new tacrine–indole heterodimers. The newly synthesized tacrine–indole family is endowed with a promising profile of biological activities that include inhibition of cholinesterases, inhibition of Aβ-dependent amyloid nucleation, the potential to penetrate the BBB and the absence of toxicologically relevant DNA intercalation. Future application of tacrine–indole heterodimers is contingent on their in vivo validation, which needs to confirm their low acute and chronic toxicity, including hepatotoxicity; favorable pharmacokinetics and brain permeability; no or low CNS side effects; and in vivo efficacy using transgenic animal models of AD. Because of the limitations of existing AD drugs, the MTDL approach remains a promising avenue for future drug development. In line with previously published reports, we expect that the potency of our newly synthesized compounds can be further increased through substitution of the thiourea tether with a urea residue [72,73].

Summary points.

Cholinesterase inhibition

• Novel tacrine–indole heterodimers exhibit nanomolar inhibitory potency on human acetylcholinesterase and human butyrylcholinesterase.

• Kinetic study of acetylcholinesterase/butyrylcholinesterase inhibition of the most potent compound, 3c, confirmed its mixed mode of human acetylcholinesterase inhibition and competitive type of human butyrylcholinesterase inhibition.

β-amyloid inhibition

• Tacrine–indole ligand 3c inhibits β-amyloid-dependent amyloid nucleation of Sup35 amyloids.

Other observed significant properties

• According to parallel artificial membrane permeability assay results, all tacrine–indole heterodimers except one are predicted to cross the blood–brain barrier by passive diffusion.