All chemicals were obtained from Sigma Aldrich and Alfa Aesar, and used as received. Hydroxytyrosol was obtained as reported in literature (Gambacorta et al., 2007). MW-assisted reactions, performed in the high energy density laboratory CEM Discover Microwave were run on dynamic mode, applying a specified amount of power to reach the desired temperature and pressure, working with borosilicate glass vessels equipped with silicon cap with septum. Reactions were monitored by TLC using silica plates 60-F264 on alumina, commercially available from Merk. Liquid flash chromatography was performed on a Supelco VERSA FLASH HTFP station using silica cartridges commercially available from Supelco. ¹H-NMR and ¹³C-NMR spectra were recorded on a Bruker WM 300 instrument on samples dissolved in CDCl₃(compounds HT1a, HT3, HT4, HT3a, HT4a) or in d₆-DMSO (compounds HT1 and HT2). Chemical shifts are given in parts per million (ppm) from tetramethylsilane as the internal standard (0.0 ppm). HRMS measurements were realized on a Thermo Scientific QExactive (Thermo Fisher, Milan, Italy) mass spectrometer working in positive mode at 35,000 resolving power, operating in SIM mode by flow injection (flow rate 15 μl/min for each stock solution). Stock solutions at a concentration of 1 mg/ml were prepared in UHPLC-MS grade MeOH for each analyte separately. Prior to analysis, each solution was diluted 1:1,000 (v/v) in a vial to obtain a concentration of 1 mg/L. Detection of the targeted compounds was based on theoretical exact mass. Data were evaluated by Xcalibur 2.2.SP1 (Thermo Fisher Scientific, Bremen, Germany). The mass accuracy, directly calculated from Xcalibur, is defined by the formula Δ(ppm) = [(theoretical mass − measured mass)/theoretical mass] × 1.000.000. The ORAC assay employed a microplate fluorimeter (Victor 2 multilabel counter, PerkinElmer Life Sciences, Waltham, MA, United States; excitation at 495 nm; emission at 535 nm; 0.5 s per reading from the bottom of the plate; 1 reading per min) and 24-well plates. Absorption measurements for the CUPRAC assay and the metal-chelating properties characterization were performed on a Perkin-Elmer Lambda 35 spectrophotometer.

The method for the ester hydrolysis was adapted from the procedure developed by Theodorou et al. (2007). Ethyl N-benzylpiperidine-4-carboxylate (1.08 g, 4.37 mmol) was dissolved in 20 ml of dichloromethane:methanol (9:1). Then, 3.0 ml of NaOH in methanol (2 N) were added, and the mixture is reacted for 24 h at room temperature and monitored by TLC. After 24 h, 1.0 ml of the basic solution was added, and the mixture was stirred for another 24 h. At the end of the reaction, the solvent was removed under reduced pressure, the residue was diluted with water and extracted three times with diethyl ether. The water phases were collected, cooled at 0°C, and acidified with HCl (6 N) till pH = 2 ÷ 3. Finally, the water phase was extracted three times with ethyl acetate. The organic phases were collected, dried with Na₂SO₄, filtered, and evaporated under reduced pressure. The product was used without further purification procedures.

The best reaction conditions developed for the synthesis of donepezil hybrids are described below. To a solution of N-benzylpiperidine-4-carboxylic acid (50 mg, 0.23 mmol) THF (5 ml), the desired polyphenolic alcohol (10 equivalents) and trimethylsilyl chloride (CTMS, 2.5 equivalents) were added in a 2-necked round bottom flask equipped with a magnetic stirrer and a condenser. The mixture was heated at 40°C for 6 days. Subsequently, the solvent was removed under reduced pressure, the crude was diluted in dichloromethane and was extracted three times with NaHCO₃ (saturated solution). The organic phases were collected and dried with Na₂SO₄, filtered, and evaporated under reduced pressure. The product is separated from the crude by flash chromatography (CH₂Cl₂/MeOH 9.5:0.5 v/v). All compounds (HT1, HT3, and HT4) were identified by HRMS and characterized by ¹H-NMR and ¹³C-NMR (See Supplementary Material).

The microwave-assisted synthesis of the acetylated compounds (HT1a, HT3a, and HT4a) were done as reported in literature (Oliverio et al., 2016). All compounds were identified by HRMS and characterized by ¹H-NMR and ¹³C-NMR (See Supplementary Material).

The synthesis of the nitro-derivative was obtained with the procedure reported in literature (Trujillo et al., 2014). Briefly, HT1 (55 mg, 0.155 mmol) was dissolved in AcOH/AcONa buffer (pH = 3.75, 31 ml). NaNO₂ (21.4 mg, 0.31 mmol) was added, and the mixture was stirred at room temperature in the dark for 30 min. Then, the aqueous solution was extracted three times with ethyl acetate. The organic phases were collected and dried with Na₂SO₄, filtered, and evaporated under reduced pressure. The product is separated from the crude by flash chromatography (CH₂Cl₂/MeOH/TEA 9.5:0.4:0.1 v/v).

The Orac assay was performed as reported in literature (Nardi et al., 2017) with slight modifications on the solvent solution. Briefly, 7% randomly methylated β-cyclodextrin (RMCD) solution (acetone/water/DMSO, 49.5:49.5:1 v/v) was used as the solvent for the Trolox standards and as a blank. The unprotected donepezil hybrids (10 μM) in this 7% RMCD solution were used as samples. To each well, 80 μl of sample, blank or Trolox (1.25, 6.25, 12.5, 25, 50 and 100 μM in 7% RMCD solution) were added to 800 μl of disodium fluorescein solution (FL, 14 μM in PBS, pre-incubated at 37°C for 15 min). Each solution was analysed in duplicate in a “forward-then-reverse” order as described in the literature (Ou et al., 2001). Reactions were initiated by the addition of 300 μl of AAPH (31.7 mM in PBS). The measurement temperature was set at 37°C. A regression equation was built by comparing the net area under the FL decay curve and the Trolox concentration according to the following equation: Y ( µM Trolox ) = a + b X ( AUC − n ) and the final ORAC_FL values were expressed as Trolox equivalents (μmol g⁻¹). The area under the curve was calculated with the following equation: AUC = 1 + ∑ i = 1 i = 45 f 1 / f 0

The CUPRAC test was performed as reported by Ozyürek et al. (2008) on the new compounds in the unacetylated form.

The MTT test is based on the observation that live cells can reduce the water-soluble yellow tetrazolium salt [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] into water-insoluble blue/magenta formazan crystals, by dehydrogenase of the active mitochondria. Dissolved formazan crystals can be quantified using a spectrophotometer and the values obtained are in direct correlation to the number of metabolically active cells. For this reason, this colorimetric assay is used to evaluate cell viability.

The MTT assay was performed as reported in literature (Corasaniti et al., 2007). SH-SY5Y were placed in 96-well microplates at a density of 6 × 10³ and, the next day, were treated with the donepezil hybrids for 24 h. Subsequently, the medium was replaced with a phenol red-free medium containing MTT solution (0.5 mg/ml) and, after 4 h incubation, 100 μl of 10% SDS was added to each well to solubilize the formazan crystals.

The optical density was measured at wavelengths of 540 and 690 nm using a spectrophotometer (X MARK Spectrophotometer Microplate Bio-Rad).

In the last experiment, SH-SY5Y cells were treated with donepezil hybrids for 21 h at the concentrations shown in the Figure 5C. After that, the cells were exposed to H₂O₂ (100 μM, 3 h), and cell viability was measured through the MTT assay as mentioned above.

H2DCF-DA is a molecule that easily diffuses into cells and, through intracellular esterases, is split into H2DCF following the loss of the acetate group. H2DCF remains trapped within cells and is oxidized by intracellular ROS to form the highly fluorescent DCF compound. Spectrofluorimetric quantification of the DCF probe, provides the content of the ROS in the cell.

SH-SY5Y cells were plated in 96-well microplates at a density of 6 × 10⁴ and, the following day, were treated with donepezil hybrids for 24 h. At the end of the treatment period, to ensure that the compounds used were not degraded, cells were exposed to fresh donepezil hybrids for an additional 3 hours. Subsequently, the growth medium was replaced with a fresh phenol red-free medium containing H2DCF-DA (25 μM). After 30 min at 37°C, the cells were washed twice to remove the extracellular H2DCF-DA, centrifuged, re-suspended in PBS, exposed to H₂O₂ (100 μM, 30 min) and the fluorescence was evaluated by flow cytometric analysis using a FACS Accury laser flow cytometer (Becton Dickinson). The results were expressed by setting the control equal to 1 and reporting all other values. H₂O₂ alone was used as a positive control.

The ORAC test for the unprotected hybrids was performed to evaluate the scavenging activity of samples against radical species. HT alone was also analyzed for comparison. The oxygen-radical absorbance capacity measured by fluorescence is a method that evaluates the antioxidant activity through a hydrogen atom transfer (HAT) mechanism. The results reported in Table 2 demonstrate that all compounds are more active than Trolox, a synthetic hydrophilic vitamin E analogue used as the reference compound. In particular, compound HT1 (entry 1, Table 2) is less active than HT (entry 8, Table 2), the derived alcohol, while the presence of a nitro group in HT2 reduced the antioxidant capacity by almost 3 times (entry 3, Table 2) compared to HT1. It is worth noting that the higher value of the Trolox equivalent was founded for the compound HT3, with the homovanillyl synthon in the structure (entry 4, Table 2).

An evaluation of the reducing activity for the unprotected compounds was also performed. The cupric reducing antioxidant capacity test is defined as an electron transfer (ET)-based assay for the evaluation of antioxidant action. As for the ORAC test, the antioxidant capacity of the tested hybrids was experimentally measured as Trolox equivalents, and HT was also analysed for comparison. As shown in Table 2, all hybrids demonstrated antioxidant activities comparable or higher than Trolox. HT1 was the more active compound (entry 1, Table 2), showing an antioxidant capacity higher than HT on its own (entry 8, Table 2). Otherwise, as shown in the ORAC test, the reduction in the antioxidant capacity for the nitro hybrid HT2 (entry 3, Table 2) is less pronounced, and its activity is almost high. Among the hybrids, the derivative HT4 possesses the lower antioxidant activity (entry 6, Table 2), although it is comparable with Trolox.

The human neuroblastoma cell line SH-SY5Y was exposed to increasing concentrations of these compounds (0.25–100 μM) for 24 h and, under these experimental conditions, all hybrids showed negligible cell death under 1 μM (Figure 4). Furthermore, data showed that all compounds are not toxic in the 1–10 μM range, and are stable in these experimental conditions. In particular, a reduction in cell viability was found for compounds HT3a and HT4a at concentrations higher than 10 μM, probably due to the higher permeability of the acetylated form.

Preliminary experiments were carried out in order to evaluate if donepezil hybrids had pro-oxidant effects. For this reason, SH-SY5Y cells were treated with the indicated compounds for 24 h and endogenous ROS were measured. H₂O₂ was used as positive control. As can be seen in Supplementary Figure S3, no compound has resulted in an accumulation of ROS at the indicated concentrations (0.25–100 μM). Subsequently, the antioxidant properties of the compounds were evaluated and conditions of oxidative stress in neuroblastoma cells were simulated by using H₂O₂ as toxic insult. The treatment time, with H₂O₂ used to produce an accumulation of ROS (30 min), did not cause reduction in cell viability. In order to obtain a significant reduction in viability we need to treat SH-SY5Y cells for 3 hours or more (see Supplementary Figure S4). Thus, exposure to H₂O₂ for 30 min causes only oxidative damage. In particular, Figures 5A,B shows the protective effect against H₂O₂, and among all concentrations used, those with the most protective effects were highlighted. The ROS values corresponding to the other concentrations are summarised in Table 2. The nitro hybrid HT2 was the more effective one to reduce the ROS amount to the physiological values. Furthermore, it is noteworthy that for the compounds HT3, the higher antioxidant effect it was measured in the acetylated form (HT3a). Then, the IC₅₀ values are reported in Table 3, together with the potential therapeutic index of these hybrids (expressed by the ratio between the concentration of the hybrid that causes a toxic effect and the concentration that causes an antioxidant effect in neuronal cells). As shown in Table 3, all hybrids possess a neurotoxic concentration higher than the concentration showing antioxidant activity. Compound HT1 showed the best therapeutic index ratio, while for compound HT3a the lower value was calculated, although it is equal to 10. Finally, to confirm the promising results obtained with the evaluation of antioxidant activity, we investigated whether donepezil hybrids were able to reduce the proliferative damage induced by exposure to H₂O₂. As shown in Figure 5C the co-treatment with donepezil hybrids and H₂O₂ had shown a protective effect. In particular, HT2, HT3a and HT4a showed a statistically significant increase in cell viability compared to the effect caused by treatment with hydrogen peroxide alone.