Antioxidant, Anti-inflammatory and Neuroprotective Profiles of Novel 1,4-Dihydropyridine Derivatives for the Treatment of Alzheimer’s Disease

Alzheimer’s disease is a chronic and irreversible pathological process that has become the most prevalent neurodegenerative disease. Currently, it is considered a multifactorial disease where oxidative stress and chronic neuroinflammation play a crucial role in its onset and development. Its characteristic neuronal loss has been related to the formation of neurofibrillary tangles mainly composed by hyperphosphorylated tau protein. Hyperphosphorylation of tau protein is related to the over-activity of GSK-3β, a kinase that participates in several pathological mechanisms including neuroinflammation. Neuronal loss is also related to cytosolic Ca2+ homeostasis dysregulation that triggers apoptosis and free radicals production, contributing to oxidative damage and, finally, neuronal death. Under these premises, we have obtained a new family of 4,7-dihydro-2H-pyrazolo[3–b]pyridines as multitarget directed ligands showing potent antioxidant properties and able to scavenge both oxygen and nitrogen radical species, and also, with anti-inflammatory properties. Further characterization has demonstrated their capacity to inhibit GSK-3β and to block L-type voltage dependent calcium channels. Novel derivatives have also demonstrated an interesting neuroprotective profile on in vitro models of neurodegeneration. Finally, compound 4g revokes cellular death induced by tau hyperphosphorylation in hippocampal slices by blocking reactive oxygen species (ROS) production. In conclusion, the multitarget profile exhibited by these compounds is a novel therapeutic strategy of potential interest in the search of novel treatments for Alzheimer’s disease.


Introduction
Alzheimer's disease (AD) is a chronic neurodegenerative disease characterized by a progressive memory loss and cognitive dysfunction. Currently, 46.8 million people suffer AD and its prevalence is forecasted to increase to 152 million by 2050 [1]. The World Health Organization considers AD as l-type VDCC antagonists, such as 1,4-dihydropyridines (1,, have been intensively pursued as potential agents to decrease [Ca 2+ ] c overload, Aβ production and NFT formation, a combination of targets that may stop the progression of the disease [46]. Based on the complexity and connectivity of the different pathological pathways involved in AD onset and development, it is described as a multifactorial disease [47] whose treatment should be addressed by acting on several targets simultaneously. These reports have boosted the development of novel small molecules designed to act on several targets at the same time, the so-called "multitarget strategy", to find a neuroprotective drug with disease modifying capacity. Considering the implication of exacerbated OS and neuroinflammation in the onset and development of AD, we report here the synthesis and pharmacological evaluation of a new family of 4,7-dihydro-2H-pyrazolo [3,4-b]pyridine derivatives, based on the 1,4-DHP core. As mentioned above, 1,4-DHPs are well-characterized selective l-type VDCC blockers with antioxidant and neuroprotective properties [48]. Based on previous results, we have included a pyrazole ring moiety in the structure of new compounds in order to implement GSK-3β inhibitory capacity to decrease tau hyperphosphorylation and, given the relationship of GSK-3β and OS with neuroinflammation, the potential inclusion of anti-inflammatory properties. The obtained derivatives showed potent free radical scavenger capacity, exerted an interesting anti-inflammatory effect, GSK-3β inhibitory capacity at low micromolar range, and moderate VDCC antagonist capacity. Additionally, these compounds showed potent neuroprotective properties against toxicity induced by OS, tau hyperphosphorylation and [Ca 2+ ] c overload. Finally, a selected compound reduced cell death in a model of hippocampal slices treated with okadaic acid (OA) by reducing OS.

Oxygen Radical Absorbance Capacity (ORAC) Assay
ORAC test developed by Cao et al. [49] and modified by Ou et al. [50] was carried out to evaluate the oxygen free radical scavenger capacity of the compounds. (±)-6-Hydroxy-2,5,7,8tetramethylchromane-2-carboxylic acid (Trolox) (1, 2, 4, 6 and 8 µM) was used as reference compound and melatonin as positive control. Compounds 4a-l were diluted (0.03, 0.1, 0.3, 1, 3 and 10 µM) in phosphate buffered saline (PBS) buffer (10 mM, pH 7.4) at 37 • C and placed a 96-well black microplate (COSTAR 3904). 150 µL of fluorescein (70 nM, final concentration), 25 µL of PBS buffer for blank, 25 µL of Trolox solution for standard and 25 µL of melatonin or compound were added to each well. A fluorescence measurement was done first to determine the basal signal. Then, 25 µL of 2,2 -azobis-amidinopropane dihydrochloride (AAPH) (12 mM, final concentration) were quickly added, since the reaction starts immediately after addition. All samples were carried out in duplicate in three different experiments. Fluorescence was recorded every min for 90 min at 37 • C to obtain the area under the fluorescence decay curve (fluorescence vs time) (AUC) in a FLUOstar Optima plate reader (BMG Labtech, Offenburg, Germany) with 485 nm excitation and 520 nm emission filters. After blank correction and plotting the net AUC vs antioxidant concentration, linear regressions were calculated for all the samples. Final results were expressed in Trolox equivalents (T.eq.), where the ORAC value of Trolox was taken as 1.0.

Mixed Primary Glial Culture
Experimental procedures performed following the Guide for the Care and Use of Laboratory Animals and were previously approved by the Ethics Committee of the Autonomous University of Madrid, Spain (PROEX: 252/16). These procedures were in agreement with the European Guidelines for the use and care of animals for research in accordance with the European Union Directive of 22 September 2010 (2010/63/UE) and with Spanish Royal Decree of 1 February 2013 (53/2013). Mixed glial cultures were obtained from cerebral cortex of 2-5-days-old Sprague-Dawley rats. After the removal of blood vessels and meninges, the forebrains were dissociated mechanically in DMEM/F12 medium. Then, cells were plated (3 × 10 5 cells/well) in DMEM/F12 medium with 20% fetal bovine serum (FBS) and 1% penicillin/streptomycin (10,000 units), at 37 • C and in a 5% CO 2 -supplemented air atmosphere. After 5 days, medium was substituted by DMEM/F12 medium with 10% FBS. Cells were cultured for 7-10 days before treatment.

Measurement of the Anti-inflammatory Capacity Measured as Nitrite Production Reduction in Mixed Primary Glial Cultures
Cells were pre-incubated with increasing concentrations (0.1, 1, 3, 10 and 30 µM) of compounds for 24 h. Then, treatments were removed and cells were incubated with Lipopolysaccharide (LPS, O26:B6 serotype, 1 µg/mL) in presence of each compound at the desired concentration. Nitrite production was assessed 18 h later by modified Griess assay. Briefly, samples (150 µL) were mixed with 4,4 -diamino-diphenylsulfone (75 µL) and n-(1-naphthyl)ethylenediamine (75 µL), and the mixture was incubated at room temperature for 5 min. Light absorption was measured at 550 nm in a microplate reader (Labtech, Offenburg, Germany). All data were normalized to basal nitrite production, considering this value as 100% of nitrite production. EC 50 values were calculated from dose-response curves represented as percentage of nitrite production reduction induced by the different concentrations of each compound.

GSK-3β Inhibitory Capacity
The method of Baki et al. [52], with some modifications, was followed to analyze the ability of the compounds to inhibit GSK-3β. Tested compounds were dissolved in assay buffer, containing 40 mM Tris (pH 7.5), 20 mM MgCl 2 , 0.1 mg/mL bovine serum albumin (BSA) and 50 µM dithiothreitol (DTT), to achieve final reaction concentrations of 0.1, 1, 10 and 30 µM. Compound SB216763 was used as reference compound. The GSK-3β inhibition assay was performed in white 96-well plates, in duplicates. Firstly, 10 µL of the compounds at the desired concentration were added to each well, followed by 10 µL of enzyme (10 ng). After 30 min of incubation, 20 µL of a mixture of adenosine triphosphate (ATP) and GSK-3β-peptide substrate were added to each well. After a 60 min incubation at 30 • C, the enzymatic reaction was stopped, and the remaining ATP concentration was measured by adding 40 µL/well of Kinase-Glo reagent. IC 50 values were calculated by non-linear regression analysis of individual concentration-response curves using GraphPad Prism software (San Diego, CA, USA). Ki values were calculated using the Chen-Prusoff equation simplified: Being S the substrate concentration (ATP) in the experiment, and K m the Michaelis-Menten constant of the substrate for the enzyme (K m = 2.345 µM for ATP).

SH-SY5Y Neuroblastoma Cell Line Culture
SH-SY5Y human neuroblastoma cells (obtained from ATCC, Manassas, VA, United States, CRL-2266) were cultured according to supplier directions in a 1:1 mixture of F12 (Ham 12) and Eagle's Minimum Essential Medium (MEM), supplemented with 15% non-essential amino acids, 10% heat-inactivated FBS, 0.5 mM sodium pyruvate, 24 mM NaHCO 3 , 100 µg/mL streptomycin and 100 units/mL penicillin. Cells were maintained at 37 • C in a humidified atmosphere of 95% air and 5% CO 2 . For experimental procedures, cells were cultured in 96-well plates (6 × 10 4 cells per well). Treatments were carried out in 1% FBS medium unless other concentration was specified. Cells were used from 4 up to 13 passages.

Voltage Dependent Calcium Channels Blockade Assay
Free cytosolic Ca 2+ was measured using the fluorescence Ca 2+ indicator Fluo-4/AM. SH-SY5Y cells were seeded in 96-well black bottom transparent plates at a density of 60,000 cells per well and cultured until they reach confluence. Cells were incubated with the Ca 2+ sensitive dye Fluo-4/AM at 5 µM in Krebs-HEPES buffer (KH) (145 mM NaCl, 4.9 mM KCl, 1.2 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES, 11 mM glucose, pH 7.4), and 0.05% pluronic acid, for 1 h at 37 • C in the dark. Then, cells were washed twice with KH solution to remove the excess of probe and compounds in solution were added. Test compounds at 10 µM were incubated 10 min before K + 70 mM was applied to trigger cytosolic Ca 2+ increase. Fluorescence was measured in a fluorescence microplate reader (FLUOstar Optima, BMG, Ortenberg, Baden-Wuerttemberg, Germany) at excitation and emission wavelengths 485 and 520 nm, respectively. To normalize Fluo-4 signals, Triton X-100 (5%) and subsequently MnCl 2 1 M were applied to the cells to register maximal and minimal fluorescence, respectively. Data were calculated as a percentage of Fluorescence increment with respect to F max − F min . The percentage of response was calculated considering basal conditions as 100% response.

Blood-Brain Barrier Permeation Assay (PAMPA)
Prediction of the capacity of compounds to cross the blood-brain barrier (BBB) by passive diffusion was evaluated using a Parallel Artificial Membrane Permeation Assay (PAMPA), in a similar manner as described previously [53]. The effective permeability of the compounds was measured in duplicate at an initial concentration of 100 µM, including reference positive and negative compounds (see Supporting Information, Table SI1). The compounds of interest were dissolved in 10 mM phosphate-buffered saline (PBS) buffer (pH 7.4) to the desired concentration in the donor well (Multiscreen IP sterile clear plate PDVF membrane, pore size 0.45 µM). The acceptor 96-well plate (Multiscreen, Millipore Corp. Burlington, MA, United States) was filled with 180 µL of PBS (VA). The filter membrane of the donor 96-well plate was impregnated with 4 µL of porcine brain lipid (PBL) (Avanti Polar Lipids, Inc., Alabaster, AL, United States) in dodecane (20 mg/mL) (Sigma-Aldrich, Madrid, Spain), and after 5 min, 180 µL of each compound solution was added to determine their ability to pass the brain barrier (VD). Then, the donor filter plate was carefully put on the acceptor plate to form a "sandwich", which was left undisturbed for 4 h at 25 • C. After incubation, the donor plate was carefully removed. An ultraviolet (UV) plate reader determined the concentration of compounds and commercial drugs in both the acceptor and the donor wells (150 µL/well) as the maximum absorption wavelength of each compound. Concentration of the compounds in the donor and acceptor well and equilibrium concentration were calculated from the standard curve and expressed as the permeability (Pe), according the equation [1] (see supporting information). Results are expressed as mean ± SEM of three different experiments in duplicate.

Molecular Docking Calculations
GSK-3β structure was prepared and minimized using MacroModel (v2018-1, Schrödinger, Inc., New York, NY, United States) with Optimized Potentials for Liquid Simulations 3 (OPLS3) force field. Next, molecular docking was performed using AutoDock Vina (v1.1.2, Molecular Graphics Lab, Scripps Research Institute, San Diego, CA, United State [54] with center of the box located on the ATP binding site, and the best nine poses were considered. The selected 4,7-dihydro-2H-pyrazolo [3,4-b]pyridine 4f (R)-and (S)-enantiomers were submitted to molecular docking with the previous specifications [28]. The same docking strategy was used for predictions with Ca V 1.2 L-type channel and selected 4,7-dihydro-2H-pyrazolo [3,4-b]pyridine compound 4f in the 1,4-DHP binding site. As a crystal structure of this subtype is not available, we used the Ca V 1.2 l-subtype VDCC model developed by Denis Tikhonov and Boris S. Zhorov and kindly shared with us by Prof. Zhorov [55].

Neuroprotection Studies in the SH-SY5Y Human Neuroblastoma Cell Line
Cells were pre-incubated with the corresponding compound at 1 µM in neuroblastoma culture medium. After 24 h, the medium was removed and replaced with 1% FBS neuroblastoma culture media containing the corresponding compound at 1 µM and the toxic stimuli, namely a mixture of rotenone and oligomycin A (R/O, 30 µM/10 µM, respectively), okadaic acid (OA) at 20 nM or a high concentration of KCl (70 mM). Cells were co-incubated for further 24 h with the R/O mixture or KCl solution, or 18 h with the OA solution. Melatonin (1 µM) or nimodipine (1 µM) were used as positive control and reference compounds in the R/O, OA or high concentration of potassium models, respectively. Control cells were incubated with the same amount of dimethyl sulfoxide (DMSO) without any drug. After the co-incubation period, cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction method [56]; basal condition was considered as 100% survival.

Measurement of ROS Production in Rat Hippocampal Slices
ROS quantification was performed using the fluorescent probe 2´,7´-dichlorodihydrofluorescein diacetate (H 2 DCFDA) (λ Ex/Em = 495/529 nm) (Thermo Fisher, Waltham, MA, USA). Slices were incubated with the probe for 45 min at a concentration of 10 µL/mL, in the presence of 1 µL/mL of Hoechst at 37 • C before finalizing the experiment. Fluorescence from the cornu Ammonis 1 (CA 1 ) was measured as mean intensity at an excitation wavelength of 480 nm and an emission of 520 nm for DCFDA in an inverted Nikon Eclipse T2000-U microscope (Nikon, Tokyo, Japan), with a 2X objective. Results were normalized to basal condition, which was considered as 100 %. NIS-Elements BR 4.10.04 64-bit software (NIS BR 4.10.04, Nikon, Tokyo, Japan) was used to analyze the images.

Statistical Analysis
All values are expressed as mean ± S.E.M. IC 50 and LD 50 values were calculated by non-linear regression analysis of individual concentration-response curves using GraphPad Prism software (v6.0, GraphPad, San Diego, CA, USA). Analysis of the results was performed by comparison between experimental and control data using one-way ANOVA followed by Newman-Keuls post-hoc test when three groups were implicated. Differences were considered to be statistically significant when p ≤ 0.05; "n" represents the number of different cultures used or enzyme inhibition assays performed.

Synthesis of 4,7-Dihydro-2H-Pyrazolo[3,4-b]Pyridine Derivatives
The synthesis of the 4,7-dihydro-2H-pyrazolo[3,4-b]pyridine derivatives 4 was performed by using a Hantzsch-like multicomponent-reaction approach from 3-methyl-1H-pyrazol-5-amine 1, aromatic aldehydes 2a-l and malononitrile 3 [28]. This mixture was refluxed in dry ethanol in the presence of dry ammonium acetate (Scheme 1) to obtain the desired compounds 4a-l. Under these conditions, we were able to obtain a library of substituted Ph derivatives bearing electron-withdrawing substituents at the aromatic ring (compounds 4a-i), as well as three compounds bearing heterocyclic substituents (4j-l). Regarding reactions involving aromatic aldehydes with electron-donating groups, we obtained only traces of the corresponding pyrazolo[3,4b]pyridine derivatives from highly complex reaction mixtures, due to full oxidation to the completely Under these conditions, we were able to obtain a library of substituted Ph derivatives bearing electron-withdrawing substituents at the aromatic ring (compounds 4a-i), as well as three compounds bearing heterocyclic substituents (4j-l). Regarding reactions involving aromatic aldehydes with electron-donating groups, we obtained only traces of the corresponding pyrazolo[3,4-b]pyridine derivatives from highly complex reaction mixtures, due to full oxidation to the completely aromatic pyridine system.

Antioxidant and Anti-inflammatory Properties of Novel 4,7-Dihydro-2H-Pyrazolo[3,4-b]Pyridine Derivatives 4a-l
As described previously, OS is considered a prominent pathological factor, not only in the progression of AD, but also in the onset and initial phases of the disease. Particularly, 1,4-DHPs are considered a specific type of antioxidants due to their ability to inhibit free radical reactions by acting as hydrogen donors and their capacity to be oxidized under biological conditions [48]. Thus, we tested the potential antioxidant effect of our 4,7-dihydro-2H-pyrazolo [3,4-b]pyridine derivatives as direct ROS scavengers using the ORAC test, and melatonin, a potent natural antioxidant molecule, as positive control. As reported in Table 1, melatonin successfully scavenged oxygen free radicals being 2.14-fold more potent than trolox. Compounds 4a-l demonstrated scavenger properties showing ORAC values between 0.63 T.eq. (4d, Ar = 3-ClPh) and 2.93 T.eq. (4j, Ar = 2-thienyl), the latter being more potent than melatonin. Considering phenyl derivatives, only compound 4b showed higher antioxidant capacity than trolox (1.86 T. eq.) being the most potent of the carbocyclic derivatives. To delve into the antioxidant characterization of compounds 4a-l, we evaluated their potential capacity to scavenge RNS by using the DPPH method. Ascorbic acid and melatonin were included as positive control and reference for comparative purposes. In general, compounds 4a-l showed increased scavenger effect towards RNS compared to melatonin, being in all cases more potent. Furthermore, they showed almost 3-fold higher antioxidant capacity than ascorbic acid, a potent RNS scavenger [57]. RNS scavenger capacity was similar in all cases with IC 50 values ranging from 17.1 µM (4k, Ar = 3-Py) to 25.2 µM (4a, Ar = Ph). This result indicates that the scavenger effect is directly related to the 4,7-dihydro-2H-pyrazolo [3,4-b]pyridine core, as it has been shown for different 1,4-DHPs described previously [58,59].
Neuroinflammation is a widely described pathological pathway present also at the initial phases of AD [60,61]. Furthermore, as depicted in the introduction section, extensive OS induces glial activation by raising the production of pro-inflammatory mediators. These mediators augment ROS and RNS to generate a positive feedback loop leading to chronic neuroinflammation, a highly toxic environment for neurons described as one of the pathological hallmarks of AD. With these premises, we tested the potential anti-inflammatory capacity of compounds 4a-l in primary rat glial cultures stimulated with LPS. TLR4/2 activation by LPS induces iNOS overexpression that increases NO production. Compounds 4a-l were evaluated at increasing concentrations (0.1, 1, 3, 10 and 30 µM) and EC 50 values were calculated from non-lineal regression of nitrite production reduction. Sulforaphane, a potent anti-inflammatory agent was included as positive control and for comparative purposes. Interestingly, all compounds showed anti-inflammatory capacity with EC 50 values ranging from 780 nM of compound 4d (Ar = 3-ClPh) to 22.3 µM of compound 4k (Ar = 3-Py) ( Table 1). With respect to the heterocyclic derivatives, only the 2-thienyl derivative showed an EC 50 lower than 10 µM.

GSK-3β Inhibition, Voltage-Dependent Ca 2+ Channels Blockade and BBB Permeation Properties
Among the pro-inflammatory pathways triggered before and after disease onset, GSK-3β is extensively described as an important kinase related to tau hyperphosphorylation and pro-inflammatory signal transduction [62,63]. Thus, we envisaged studying the GSK-3β inhibitory capacity of derivatives 4a-l. To this end, we used the luminescence Kinase-Glo assay at four different concentrations (0.1, 1, 10 and 30 µM). IC 50 values and their related apparent equilibrium dissociation constants (K i ) are summarized in Table 2, including compound SB216763, a potent GSK-3β inhibitor, as a positive control. Compound 4a showed an improved inhibitory capacity compared to its pyrano [2,3-c]pyrazole isoster [28], indicating an important role of the 1,4-DHP ring in the interaction with GSK-3β. Interestingly, the inclusion of substituents at the aromatic moiety at C-4 position increased their capacity to inhibit GSK-3β, showing IC 50 values from 820 nM of derivative 4h (Ar = 3-NO 2 Ph) to 10.2 µM of compound 4l (Ar = 4-Py). Considering substituent patterns, the tendency indicated that GSK-3β inhibitory potency increased when the substituent was inserted in ortho and meta positions compared to para position. Focusing on heterocyclic derivatives, the addition of an electron-rich heterocycle (thiophene) increased the inhibitory potency, compared to an electron-deficient ring (pyridine), the thienyl derivative being 4j two-fold more potent than 4k (Ar = 3-Py) and 5-fold more potent than 4l (Ar = 4-Py). Free radical scavenger effect was evaluated from dose-response curves of compounds evaluating their ability to reduce reactive oxygen species (ROS) derived free radicals (ORAC) and reactive nitrogen species (RNS) derived free radicals (DPPH). Nitrite production reduction was evaluated by the Griess method in primary glial cultures stimulated with LPS (1 µg/mL). *-not present or not measured. As discussed, inflammatory pathway activation is partially mediated by GSK-3β; thus, the inhibitory properties of this kinase by compounds 4a-l might be related to their anti-inflammatory capacity. Indeed, the anti-inflammatory properties of ClPh and NO 2 Ph derivatives correlated with GSK-3β inhibitory capacity, being meta derivatives most potent in both assays. Similar conclusions can be extracted considering the heterocyclic substituents. In this case, the most potent GSK-3β inhibitor, compound 4j (Ar = 2-thienyl, IC 50  These results indicate that the anti-inflammatory properties of compounds 4a-l could be partially dependent on the GSK-3β inhibitory capacity, and that their antioxidant properties participate in its anti-inflammatory mechanism of action. As previously described, specific Ca 2+ overload through L-type VDCC has been directly related to OS generation [42,45] and neuronal death. In this regard, the 1,4-DHP ring has specific activity towards L-type VDCCs. Thus, compounds 4a-l were tested as potential VDCC antagonist at 10 µM and the results are shown as percentage of response after membrane depolarization induced by 70 mM K + stimuli in the SH-SY5Y neuronal cell line. Nimodipine was included as positive control at the same concentration. In general, compounds 4a-l showed interesting antagonistic activity reducing Ca 2+ entrance to 85.6 % (4k) up to only 6.46 % response (4e). Compounds 4e and 4f (Ar = 4-ClPh and Ar = 2-BrPh derivatives, respectively) showed a highly effective blockade of [Ca 2+ ] c elevation, being more effective than nimodipine at the same concentration. Interestingly, they contain a halogen substitution, as other 1,4-DHPs used in the clinic as VDCC antagonists that are being tested for mild cognitive impairment [66].
Finally, considering the potential use of our compounds as AD treatment, we explored the BBB penetration using the parallel artificial membrane permeation assay (PAMPA) prediction method. The in vitro permeabilities (Pe) of compounds 4a-l were calculated using the lipid extract of porcine brain ( Table 2). We have used a range of positive and negative controls (see Supporting Information, Table SI1). The results depicted in Table 2 indicate that, excluding 4h (Ar = 3-NO 2 Ph), 4i (Ar = 4-NO 2 Ph) and 4k (Ar = 3-Py), all compounds are predicted to be able to cross the BBB by passive diffusion with Pe values higher than 4, a widely accepted value that marks the limit for passive diffusion [53]. In order to understand the GSK-3β inhibition ability of compounds 4a-l, we studied the binding mode of the most potent GSK-3β inhibitor, derivative 4f (Ar = 2-BrPh), to GSK-3β ( Figure 1) by means of molecular docking experiments. To this end, we evaluated both enantiomers at the ATP binding site of the enzyme. Results afforded similar poses for the (R)-and (S)-enantiomers of 4f, overlapping the pyridine ring with a 180 • flip, leaving the pyrazole core on opposite sides. Both enantiomers form stable hydrogen bonds with Q 185 and N 186 residues ( Figure 1A,B). Remarkably, the hydrogen bond between the NH group of the 1,4-DHP ring of both enantiomers and N 186 emphasizes the importance of this group for its inhibitory activity. Additionally, both enantiomers showed hydrogen bonding with Q 185 . The (R)-enantiomer forms this interaction through the amino group at the 1,4-DHP ring ( Figure 1A), while the (S)-enantiomer uses the NH moiety at the pyrazole ring ( Figure 1B), in line with the predicted 180 • flip binding mode shown in Figure 1. Compound 4f predicted poses were located in the surroundings of D 200 , key residue for ATP stabilization in the substrate cavity (see Figure 1 and Supporting Information, Figure SI1).
The in vitro permeabilities (Pe) of compounds 4a-l were calculated using the lipid extract of porcine brain ( Table 2). We have used a range of positive and negative controls (see Supporting Information, Table SI1). The results depicted in Table 2 indicate that, excluding 4h (Ar = 3-NO2Ph), 4i (Ar = 4-NO2Ph) and 4k (Ar = 3-Py), all compounds are predicted to be able to cross the BBB by passive diffusion with Pe values higher than 4, a widely accepted value that marks the limit for passive diffusion [53].

Molecular Docking: Proposed Binding Mode of Compound 4f at the ATP Binding Site of GSK-3β and the 1,4-DHP Binding Site of L-Type VDCC
In order to understand the GSK-3β inhibition ability of compounds 4a-l, we studied the binding mode of the most potent GSK-3β inhibitor, derivative 4f (Ar = 2-BrPh), to GSK-3β ( Figure 1) by means of molecular docking experiments. To this end, we evaluated both enantiomers at the ATP binding site of the enzyme. Results afforded similar poses for the (R)-and (S)-enantiomers of 4f, overlapping the pyridine ring with a 180° flip, leaving the pyrazole core on opposite sides. Both enantiomers form stable hydrogen bonds with Q 185 and N 186 residues (Figures 1A,B). Remarkably, the hydrogen bond between the NH group of the 1,4-DHP ring of both enantiomers and N 186 emphasizes the importance of this group for its inhibitory activity. Additionally, both enantiomers showed hydrogen bonding with Q 185 . The (R)-enantiomer forms this interaction through the amino group at the 1,4-DHP ring ( Figure 1A), while the (S)-enantiomer uses the NH moiety at the pyrazole ring ( Figure 1B), in line with the predicted 180° flip binding mode shown in Figure 1. Compound 4f predicted poses were located in the surroundings of D 200 , key residue for ATP stabilization in the substrate cavity (see Figure 1 and Supporting Information, Figure SI1). Considering that compound 4f showed a potent L-type VDCC blockade capacity (% response = 7.39%), we also carried out molecular docking studies to determine the binding mode and potential interactions with the CaV1.2 L-type VDCC. The L-type VDCC crystal structure is not available; thus, we used a structural homology model of the channel developed by Zhorov et al. [55]. As previously done, we performed docking studies of both enantiomers of 4f at the 1,4-DHP L-type VDCC binding site located at the interface of domains III and IV of the α1 pore-constituent subunit (Figure 2). Docking results predicted a binding mode similar to that previously described for nimodipine (see Supporting Information, figure SI2) [55]. The results obtained indicate that compound 4f places the 1,4-DHP ring of both enantiomers in a similar position to (S)-nimodipine, although they are slightly displaced towards the I 3i14 and Y 3i10 residues ( Figure 2 and Supporting Information, Figure SI2). In this sense, the stern and bowsprit positions of 4f are close to those of (S)-nimodipine (see Supporting Information, figure SI2). Both enantiomers showed a similar pose with a 180° flip generating similar interactions with key residues known to be involved in 1,4-DHPs stabilization from experimental data [67], such as (i) halogen bond between bromine present in the bowsprit and Y 4i11 ; (ii) hydrogen Considering that compound 4f showed a potent L-type VDCC blockade capacity (% response = 7.39%), we also carried out molecular docking studies to determine the binding mode and potential interactions with the Ca V 1.2 L-type VDCC. The L-type VDCC crystal structure is not available; thus, we used a structural homology model of the channel developed by Zhorov et al. [55]. As previously done, we performed docking studies of both enantiomers of 4f at the 1,4-DHP L-type VDCC binding site located at the interface of domains III and IV of the α1 pore-constituent subunit ( Figure 2). Docking results predicted a binding mode similar to that previously described for nimodipine (see Supporting Information, Figure SI2) [55]. The results obtained indicate that compound 4f places the 1,4-DHP ring of both enantiomers in a similar position to (S)-nimodipine, although they are slightly displaced towards the I 3i14 and Y 3i10 residues (Figure 2 and Supporting Information, Figure SI2). In this sense, the stern and bowsprit positions of 4f are close to those of (S)-nimodipine (see Supporting Information, Figure SI2). Both enantiomers showed a similar pose with a 180 • flip generating similar interactions with key residues known to be involved in 1,4-DHPs stabilization from experimental data [67], such as (i) halogen bond between bromine present in the bowsprit and Y 4i11 ; (ii) hydrogen bond between N atom of pyrazole core from starboard side of (S)-enantiomer and Q 3o18 . Interestingly, compound 4f also shows a π-π stacking interaction with F 3i22 , defined as one of the major contributors to the ligand-receptor energy stabilization of 1,4-DHPs at the Ca V 1.2 l-subtype VDCC model [55]. In summary, the predicted binding mode for compound 4f is analogous to other 1,4-DHPs such as (S)-nimodipine, showing interactions with key DHP-sensing residues of Ca V 1.2 l-subtype VDCCs.
bond between N atom of pyrazole core from starboard side of (S)-enantiomer and Q 3o18 . Interestingly, compound 4f also shows a π-π stacking interaction with F 3i22 , defined as one of the major contributors to the ligand-receptor energy stabilization of 1,4-DHPs at the CaV1.2 L-subtype VDCC model [55]. In summary, the predicted binding mode for compound 4f is analogous to other 1,4-DHPs such as (S)-nimodipine, showing interactions with key DHP-sensing residues of CaV1.2 L-subtype VDCCs.

Neuroprotective Properties of Derivatives 4a-l towards Tau-Hyperphosphorylation, [Ca 2+ ]c Overload and Oxidative Stress
As mentioned earlier, OS strongly contributes to AD onset and development, being a key pathological factor in the neurodegeneration observed in this disease [68]. Compounds 4a-l exert significant antioxidant properties (Table 1); thus, we tested the potential neuroprotection towards OS toxicity induced by rotenone (30 µM) and oligomycin A (10 µM) (R/O). Compounds 4a-l were tested at 1 µM concentration including melatonin as positive control and for comparative purposes. Results are expressed as percentage of survival considering non-treated cells as 100% survival. As shown in Figure 3A, all compounds significantly increased SH-SY5Y cells survival compared to the toxic challenge, eliciting protection values ranging from to 32.0% for compound 4d (Ar = 3-ClPh) to 71.6 % of compound 4l (Ar = 4-Py). Interestingly, compounds 4b (Ar = 4-FPh), 4g (Ar = 4-BrPh) and 4l (Ar = 4-Py), with protective percentages of 62.7%, 59.7% and 71.6%, respectively, showed a better neuroprotective effect than melatonin (52.5%). In this case, considering that the R/O combination induces high amounts of ROS, we observed an interesting correlation between the neuroprotective ability of compounds 4a-l and their capacity to scavenge ROS. In general, the neuroprotective effect increased with higher scavenger capacity, except in the case of compound 4j. Compounds with ORAC values lower than 0.9 T.eq. showed the lowest percentages of survival (% survival < 75.7%), while compounds showing higher ORAC values presented percentages of survival higher than 78% in all cases. For example, chlorine derivatives 4c, 4d and 4e (o-, m-and p-ClPh, respectively) showed scavenger capacities of 0.94, 0.63 and 0.66 T.eq., respectively, directly correlated with percentages of survival of 81.3%, 72.3% and 73.6%. In this line, derivative 4b (Ar = 4-FPh) with ROS scavenger ability of 1.86 showed an improvement on cell viability to 84.3%. The same trend can be extracted for bromophenyl, nitrophenyl and pyridyl derivatives.
As summarized in the introduction, GSK-3β is the main kinase related to the formation of neurotoxic hyperphosphorylated tau, the core component of NFTs in AD. Considering the antioxidant and GSK-3β inhibitory properties of compounds 4a-l, we envisaged their evaluation as potential neuroprotective agents in an in vitro model of tau hyperphosphorylation induced by OA. OA inhibits protein phosphatases 1 and 2A [69], increasing the phosphorylation status of tau [70], being an accepted in vitro model of AD. The toxic cascade activated by OA includes increased levels of OS that also contributes to cellular death [71]. All compounds were tested at 1 µM concentration (S)-enantiomer represented as white and green sticks, respectively. Hydrogen bond, halogen bond and π-π stacking interactions showed as yellow, purple and blue dotted lines, respectively. Ca V 1.2 L-type channel structure represented as colored cartoon with key residues as sticks.

Neuroprotective Properties of Derivatives 4a-l towards Tau-Hyperphosphorylation, [Ca 2+ ] c Overload and Oxidative Stress
As mentioned earlier, OS strongly contributes to AD onset and development, being a key pathological factor in the neurodegeneration observed in this disease [68]. Compounds 4a-l exert significant antioxidant properties (Table 1); thus, we tested the potential neuroprotection towards OS toxicity induced by rotenone (30 µM) and oligomycin A (10 µM) (R/O). Compounds 4a-l were tested at 1 µM concentration including melatonin as positive control and for comparative purposes. Results are expressed as percentage of survival considering non-treated cells as 100% survival. As shown in Figure 3A, all compounds significantly increased SH-SY5Y cells survival compared to the toxic challenge, eliciting protection values ranging from to 32.0% for compound 4d (Ar = 3-ClPh) to 71.6% of compound 4l (Ar = 4-Py). Interestingly, compounds 4b (Ar = 4-FPh), 4g (Ar = 4-BrPh) and 4l (Ar = 4-Py), with protective percentages of 62.7%, 59.7% and 71.6%, respectively, showed a better neuroprotective effect than melatonin (52.5%). In this case, considering that the R/O combination induces high amounts of ROS, we observed an interesting correlation between the neuroprotective ability of compounds 4a-l and their capacity to scavenge ROS. In general, the neuroprotective effect increased with higher scavenger capacity, except in the case of compound 4j. Compounds with ORAC values lower than 0.9 T.eq. showed the lowest percentages of survival (% survival < 75.7%), while compounds showing higher ORAC values presented percentages of survival higher than 78% in all cases. For example, chlorine derivatives 4c, 4d and 4e (o-, mand p-ClPh, respectively) showed scavenger capacities of 0.94, 0.63 and 0.66 T.eq., respectively, directly correlated with percentages of survival of 81.3%, 72.3% and 73.6%. In this line, derivative 4b (Ar = 4-FPh) with ROS scavenger ability of 1.86 showed an improvement on cell viability to 84.3%. The same trend can be extracted for bromophenyl, nitrophenyl and pyridyl derivatives.
As summarized in the introduction, GSK-3β is the main kinase related to the formation of neurotoxic hyperphosphorylated tau, the core component of NFTs in AD. Considering the antioxidant and GSK-3β inhibitory properties of compounds 4a-l, we envisaged their evaluation as potential neuroprotective agents in an in vitro model of tau hyperphosphorylation induced by OA. OA inhibits protein phosphatases 1 and 2A [69], increasing the phosphorylation status of tau [70], being an accepted in vitro model of AD. The toxic cascade activated by OA includes increased levels of OS that also contributes to cellular death [71]. All compounds were tested at 1 µM concentration and melatonin was included as positive control [28] and for comparative purposes ( Figure 3B). Successfully, compounds 4a-l showed neuroprotective capacity ranging from 34.6% protection of compound 4h (Ar = 3-NO 2 Ph) to 77.9% protection of compound 4g (Ar = 4-BrPh). Compounds 4c (Ar = 2-ClPh), 4f (Ar = 2-BrPh), 4g (Ar = 4-BrPh), 4k (Ar = 3-Py) and 4l (Ar = 4-Py) showed neuroprotective percentages over 50%, being the compounds 4f, 4g and 4l more potent than the reference compound, melatonin. In this particular model, the neuroprotective effect appears to be dependent on both, the antioxidant effect and the GSK-3β inhibitory potency.
Regarding the second target pursued in compounds 4a-l, VDCC antagonism, we evaluated their potential neuroprotective profile against [Ca 2+ ] c overload induced by high K + (70 mM). Nimodipine, a potent VDCC antagonist, was included as positive control at 1 µM. As shown in Figure 3C, compounds 4a-l exhibited an interesting neuroprotective profile with survival percentages ranging from 61.4 % of compound 4d (Ar = 3-ClPh) to 78.9 % for compound 4l (Ar = 4-Pyridyl). In this case, neuroprotection capacity was similar for all compounds with percentages of survival in the range of 70%, except in the case of compound 4l. Compounds 4e and 4f showed highly potent VDCC antagonism capacity. This potency correlated with interesting neuroprotection capacity (67.6% and 68.0%, respectively); however, they were moderated neuroprotectants compared to compound 4l. Regarding their antioxidant properties, in this case, the correlation between their neuroprotective capacity against [Ca 2+ ] c overload and their scavenger capacity was not clearly correlated. However, the most potent neuroprotectant in this model, compound 4l, showed moderate VDCC antagonist capacity and high antioxidant ability (ORAC = 1.1 T.eq.) indicating a potential implication in its neuroprotective mechanism of action. Recently, we described the potential dependence of the neuroprotective effect in this toxicity model with the antioxidant effect [72] following results previously published and demonstrating the relationship between Ca 2+ entry through VDCC and increased levels of free radicals [42,45,73].
In addition to compounds 4a-l neuroprotective profile, we evaluated their safety profile by testing their neurotoxicity in the SH-SY5Y neuroblastoma cell line and their gliotoxicity in primary mixed glial cultures. Our results indicate that compounds 4a-l are neither neurotoxic nor gliotoxic with LD 50 values higher than 100 µM (Table SI2, Supporting Information).

Compound 4g Reduces Cellular Death Induced by Okadaic Acid in Hippocampal Slices by Reducing Oxidative Stress
These encouraging results prompted us to test our compounds in a more complex model of AD. For this, we selected an acute treatment with OA of hippocampal slices. Recently, we have demonstrated the deleterious effect of OA (1 µM) in hippocampal slices after 6 h treatment [74]. It increases ROS production, decreases cellular viability and deregulates the autophagy system; thus, it was described as an ex vivo model of tauopathy mimicking AD pathology. For this purpose, we selected compound 4g (Ar = 4-BrPh) due to its overall pharmacological profile as a potent antioxidant able to scavenge ROS and RNS free radicals, a potent anti-inflammatory properties at low micromolar range, a GSK-3β inhibitor in the same low micromolar range and a moderate VDCC antagonist. The multitarget profile integrated into compound 4g afforded potent neuroprotective effect against OS and [Ca 2+ ] c overload, and it was the most potent compound against tau hyperphosphorylation. Compound 4g was tested at 10 µM, a concentration established for control and reference compounds used in these ex vivo models. For these experiments, we included melatonin, known to be highly effective in this model, as a reference and positive control.
As shown in Figure 4A, OA reduced cell viability by 30% (p < 0.001) measured by the MTT reduction assay. Successfully, compound 4g recovered cell viability to 92.1%, being a slightly more potent neuroprotector than melatonin, although the differences were not significant ( Figure 4B). As demonstrated previously, treatment with OA induces a significant increase in free radical production [74], measured as H 2 DCFDA fluorescence intensity increase (p < 0.01), and this ROS increase is related to the toxicity elicited by OA. Under these experimental conditions, compound 4g effectively reduced free radical production to basal conditions (p < 0.01) with an equal potency to melatonin, a well-known antioxidant molecule ( Figure 4A,C). As summarized above, we have demonstrated the scavenger capacity of this compound, and thus this antioxidant capacity should be part of its protective mechanism of action. and [Ca 2+ ]c overload, and it was the most potent compound against tau hyperphosphorylation. Compound 4g was tested at 10 µM, a concentration established for control and reference compounds used in these ex vivo models. For these experiments, we included melatonin, known to be highly effective in this model, as a reference and positive control.
As shown in Figure 4A, OA reduced cell viability by 30% (p < 0.001) measured by the MTT reduction assay. Successfully, compound 4g recovered cell viability to 92.1%, being a slightly more potent neuroprotector than melatonin, although the differences were not significant ( Figure 4B). As demonstrated previously, treatment with OA induces a significant increase in free radical production [74], measured as H2DCFDA fluorescence intensity increase (p < 0.01), and this ROS increase is related to the toxicity elicited by OA. Under these experimental conditions, compound 4g effectively reduced free radical production to basal conditions (p < 0.01) with an equal potency to melatonin, a wellknown antioxidant molecule ( Figure 4A,C). As summarized above, we have demonstrated the scavenger capacity of this compound, and thus this antioxidant capacity should be part of its protective mechanism of action.

Conclusions
AD is currently the most common type of dementia affecting an increasing number of patients and is one of the most important unmet clinical needs for our society. The development of an effective therapy with real disease modifying capacity is an urgent goal for medicinal chemists and biomedical research. The complexity of AD has led to the proposal of the "multitarget hypothesis" for drug development in which a unique chemical entity combines different activities to tackle the pathological pathways involved in AD onset and development. Following this hypothesis, we have obtained a new set of multitarget compounds based on the chemical core of the 1,4-DHPs due to their antioxidant properties and well-known VDCC antagonism capacity. In addition, the fusion of a pyrazole ring generated a suitable scaffold able to inhibit GSK-3β, a kinase related to tau Cell viability was then evaluated by MTT reduction assay and normalized to basal conditions (100%). (C) Quantitation of H 2 DCFDA fluorescence intensity as free radical production measurement after the experimental conditions described in (B). Data are expressed as mean ± S.E.M of four different experiments. ## p < 0.01, ### p < 0.001 compared to basal condition; ** p < 0.01, *** p < 0.001, compared to OA treated hippocampal slices.

Conclusions
AD is currently the most common type of dementia affecting an increasing number of patients and is one of the most important unmet clinical needs for our society. The development of an effective therapy with real disease modifying capacity is an urgent goal for medicinal chemists and biomedical research. The complexity of AD has led to the proposal of the "multitarget hypothesis" for drug development in which a unique chemical entity combines different activities to tackle the pathological pathways involved in AD onset and development. Following this hypothesis, we have obtained a new set of multitarget compounds based on the chemical core of the 1,4-DHPs due to their antioxidant properties and well-known VDCC antagonism capacity. In addition, the fusion of a pyrazole ring generated a suitable scaffold able to inhibit GSK-3β, a kinase related to tau hyperphosphorylation and pro-inflammatory pathways. The combination of pyrazole and 4-aryl-1,4-dihydropyridine moieties into a single molecular entity was achieved through a Hantzsch-like three-component reaction from 3-methyl-1H-pyrazol-5-amine, aromatic aldehydes and malononitrile. Compounds 4a-l, thus generated, have demonstrated ROS and RNS antioxidant capacity, anti-inflammatory properties, GSK-3β inhibitory capacity, at low micromolar range and l-type VDCC blockade with micromolar potency. Moreover, the multi-target directed ligands developed were found to lack neurotoxicity or gliotoxicity. They also showed good neuroprotective activity towards OS, tau-hyperphosphorylation and [Ca 2+ ] c overload induced toxicity in cell cultures, with interesting correlations between their primary targets and this neuroprotective effect indicating their participation in the neuroprotective mechanism of action. From our set of compounds we have identified 6-amino-4-(4-bromophenyl)-3-methyl-4,7-dihydro-2H-pyrazolo [3,4-b]pyridine-5-carbonitrile (4g) as an interesting multitarget directed ligand for further optimization. Compound 4g showed good antioxidant activity (ORAC = 0.96 T.eq.; DPPH IC 50 = 24.1 µM), potent anti-inflammatory capacity (EC 50 = 4.50 µM), GSK-3β inhibitory capacity (IC 50 = 2.35 µM) and VDCC antagonist activity (percentage response reduced to 62%) and it is predicted to cross the BBB. Moreover, it showed potent neuroprotective properties against OS, tau hyperphosphorylation and [Ca 2+ ] c overload. When tested in a more complex ex vivo model of AD, the acute treatment of hippocampal slices with OA, it demonstrated an extreme reduction of ROS production and also showed potent cytoprotecting capacity. Thus, this compound can be regarded as a good lead for future efforts to increase its potency towards primary targets in a hit to lead project that is currently under development in our laboratories following the promising and novel multitarget approach described here.