Discovery of Azaindolin-2-One as a Dual Inhibitor of GSK3β and Tau Aggregation with Potential Neuroprotective Activity

The inhibition of glycogen synthase kinase 3β (GSK3β) activity through pharmacological intervention represents a promising approach for treating challenging neurodegenerative disorders like Alzheimer’s disease. Similarly, abnormal tau aggregate accumulation in neurons is a hallmark of various neurodegenerative diseases. We introduced new dual GSK3β/tau aggregation inhibitors due to the excellent clinical outcome of multitarget drugs. Compound (E)-2f stands out among the synthesized inhibitors as a promising GSK3β inhibitor (IC50 1.7 µM) with a pronounced tau anti-aggregation effect in a cell-based model of tauopathy. Concurrently, (E)-2f was demonstrated to be non-toxic to normal cells, making it a promising neuroprotective lead compound that needs further investigation.


Introduction
Alzheimer's disease (AD) is a neurological disorder manifested with memory loss as well as cognitive and linguistic impairment [1]. Despite several theoretical ideas and models, there is no complete explanation for AD etiology, which has become a critical constraint in developing innovative AD therapies. However, AD has two neuropathological characteristics: the β-amyloid plaques and the neurofibrillary tangles (NFTs), both of which are prevalent in the hippocampus region [2,3]. When these two conditions are combined, cholinergic neurons are gradually lost, resulting in cognitive impairment.
According to the Amyloid Cascade Theory, AD amyloid-β (Aβ) deposition in the brain is the first event. However, there is mounting evidence that this explanation falls short of explaining many aspects of AD pathogenesis. Increased inflammatory markers in Alzheimer's patients and identifying AD risk genes linked to innate immune processes denote that neuroinflammation plays a crucial role in AD's etiology [4]. Glial cells, including compound III demonstrated excellent GSK3β inhibitory action (IC 50 = 34 nM) [24]. Gandini et al. identified compound IV as the first multitargeted scaffold that has the potential to inhibit GSK3β and tau-cascade [25]. Natural compounds like Staurosporine (II) have been extensively investigated [23]. In 2011, Chen et al. repositioned compound III as an ATP-competitive GSK3β inhibitor rather than a protein kinase C inhibitor, as previously stated. In vitro, the bisindolyl-pyrazolone-based compound III demonstrated excellent GSK3β inhibitory action (IC50 = 34 nM) [24]. Gandini et al. identified compound IV as the first multitargeted scaffold that has the potential to inhibit GSK3β and tau-cascade [25]. Bourahla et al. developed compound V as a rhodamine-based GSK3β inhibitor with pyridyl substituent [26]. In 2019, Lozinskaya et al. developed a novel ATP competitive GSK3β inhibitor VI with a 3-pyridyl 2-oxindole ring as a scaffold [27]. Compound VII, reported by Chu et al., showed triple inhibitory actions at nanomolar concentrations. In vitro binding assays confirmed the critical efficiency of compound VII on alpha-synuclein (α-syn), beta-amyloid (Aβ), and tau fibrils [28]. Tong et al. developed an azaindole-pyridine-based GSK3β inhibitor (compound VIII) and identified the azaindole moiety as the hinge-binding motif [29]. Interestingly, Selenica M-L et al. demonstrated in vivo that six different classes of small-molecules GSK3β inhibitors effectively inhibit tau phosphorylation in regions relevant to the pathogenesis of AD and at brain concentrations within their IC50 range [30]. Therefore, the inhibition of GSK3β by small molecules is a promising strategy for therapeutic targeting of AD.
Since the MTDD technique is ideal for designing ligands for diseases with complicated pathology and lacks definitive treatment, we used the MTDD technique to design ligands bound to several targets, all associated with the AD etiology. Our ligands act as inhibitors for GSK3β as well as Aβ/tau/α-syn aggregation.

Design of Dual GSK3β/tau Aggregation Inhibitors
It is challenging to design a ligand that can bind with two different targets with no homologous binding sites, such as GSK3β and tau protein [31]. More specifically, combining the molecular scaffolds of two central nervous system (CNS)-directed single-target ligands in a single small molecule may also be difficult, particularly in pharmacokinetics  Bourahla et al. developed compound V as a rhodamine-based GSK3β inhibitor with pyridyl substituent [26]. In 2019, Lozinskaya et al. developed a novel ATP competitive GSK3β inhibitor VI with a 3-pyridyl 2-oxindole ring as a scaffold [27]. Compound VII, reported by Chu et al., showed triple inhibitory actions at nanomolar concentrations. In vitro binding assays confirmed the critical efficiency of compound VII on alpha-synuclein (α-syn), beta-amyloid (Aβ), and tau fibrils [28]. Tong et al. developed an azaindolepyridine-based GSK3β inhibitor (compound VIII) and identified the azaindole moiety as the hinge-binding motif [29]. Interestingly, Selenica M-L et al. demonstrated in vivo that six different classes of small-molecules GSK3β inhibitors effectively inhibit tau phosphorylation in regions relevant to the pathogenesis of AD and at brain concentrations within their IC50 range [30]. Therefore, the inhibition of GSK3β by small molecules is a promising strategy for therapeutic targeting of AD.
Since the MTDD technique is ideal for designing ligands for diseases with complicated pathology and lacks definitive treatment, we used the MTDD technique to design ligands bound to several targets, all associated with the AD etiology. Our ligands act as inhibitors for GSK3β as well as Aβ/tau/α-syn aggregation.

Design of Dual GSK3β/tau Aggregation Inhibitors
It is challenging to design a ligand that can bind with two different targets with no homologous binding sites, such as GSK3β and tau protein [31]. More specifically, combining the molecular scaffolds of two central nervous system (CNS)-directed single-target ligands in a single small molecule may also be difficult, particularly in pharmacokinetics adjustment [32]. Many traditional GSK3 inhibitors have five-membered heterocyclic rings, such as 2-iminothiazolidin-4-one, hydantoin, or thiazolidinediones; these rings are essential to hydrogen bond acceptors and donors, which are critical to be active against GSK3β [33,34]. Instead, oxindole and its N bioisostere have attracted attention as new It was hypothesized that 5-arylidene substitution plays a vital role in the activity of 2-iminothiazolidin-4-one based GSK3β inhibitors. 5-Arylidene not only enhances GSK3β affinity but also enhances GSK3β selectivity as they do not fit in similar regions of homologous kinases [34]. This design would also allow the preservation of planarity and aromaticity, which are essential for tau fibril interaction [35]. Based on numerous reported examples of GSK3β inhibitors (Compounds V, VI, and VIII) and docking studies, we predicted that hybridizing pyridyl ring with oxindole or its N bioisostere would provide us with dual-acting anti-AD chemotypes, particularly as anti-tau and GSK3β inhibitors ( Figure 2). such as 2-iminothiazolidin-4-one, hydantoin, or thiazolidinediones; these rings are es tial to hydrogen bond acceptors and donors, which are critical to be active against GS [33,34]. Instead, oxindole and its N bioisostere have attracted attention as new fragm with the same hydrogen bond functionalities. Compounds VI and VIII are typical ex ples of GSK3β inhibitors with oxindole (compound VI) or its N bioisostere azain (Compound VIII). The rationale for selecting these two nuclei is the extra ability to vent the development of tau aggregates and destabilize preexisting tau aggregates dose-dependent manner.
It was hypothesized that 5-arylidene substitution plays a vital role in the activi 2-iminothiazolidin-4-one based GSK3β inhibitors. 5-Arylidene not only enhances GS affinity but also enhances GSK3β selectivity as they do not fit in similar regions of ho ogous kinases [34]. This design would also allow the preservation of planarity and maticity, which are essential for tau fibril interaction [35]. Based on numerous repo examples of GSK3β inhibitors (Compounds V, VI, and VIII) and docking studies, we dicted that hybridizing pyridyl ring with oxindole or its N bioisostere would provid with dual-acting anti-AD chemotypes, particularly as anti-tau and GSK3β inhibitors ure 2).

Results
We tested the following to determine if the synthesized compounds could hit targets at the same time: (a) early GSK3β activity screening of all produced compou (b) Molecular docking study, (c) Selective cytotoxicity on cancer cells and normal b cells, (d) Anti-tau aggregation in a tau pathology cell model.

Chemistry
The target compounds were synthesized in two stages, as depicted in Figures 3 4. In Figure 3, to structurally optimize the first part (oxindole or its pyrrolo [2,3-b] dine-2-one bioisostere), we have fixed the aryl substitution with the methylene linker 2-vinyl pyridine substituent on the indole ring. As demonstrated in Figure 3

Results
We tested the following to determine if the synthesized compounds could hit both targets at the same time: (a) early GSK3β activity screening of all produced compounds, (b) Molecular docking study, (c) Selective cytotoxicity on cancer cells and normal blood cells, (d) Anti-tau aggregation in a tau pathology cell model.

Chemistry
The target compounds were synthesized in two stages, as depicted in Figures 3 and 4. In Figure 3, to structurally optimize the first part (oxindole or its pyrrolo [2,3-b] pyridine-2one bioisostere), we have fixed the aryl substitution with the methylene linker as a 2-vinyl pyridine substituent on the indole ring. As demonstrated in Figure 3  To optimize the aryl part structurally, we kept the first part as 1,3-dihydro-2H rolo [2,3-b] pyridine-2-one, as depicted in Figure 4. The second series was synthesiz 24 h refluxing of 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one with different ary  To optimize the aryl part structurally, we kept the first part as 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one, as depicted in Figure 4. The second series was synthesized by 24 h refluxing of 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one with different aryl substituents in ethanol and piperidine to get three novel molecules (3a,b and 3d). In comparison, compound 3c was obtained without reflux by stirring in an ice bath for two hours then at room temperature for 12 h.

GSK3β Inhibition Assay
Screening of GSK3β inhibition activity for our new library was conducted using the Kinase-Glo luminescence assay to determine how much ATP is reduced after a kinase reaction [3]. All compounds were first tested at 10 µM and hit with a percentage inhibition of less than 50%, then were examined further to derive IC50 values, which were calculated using the linear regression parameter. Staurosporine (II) was used as a reference compound; it showed 98.46% inhibition at 10 µM. Excitingly, one of our first version compounds (E)-2f demonstrated 76.4% inhibition, while two compounds ((E)-and (E)-2c) showed percent inhibition at 45% and the others ((E)-2d and (E)-2e) showed less than 30%.
As (E)-2f has the most potency, we tried to optimize structurally (E)-2f by using different pyridyl substituents, as shown in Figure 4. Upon measuring the IC50 of these compounds, none had IC50 values better than (E)-2f, which revealed an IC50 of 1.7 µM (Figure To optimize the aryl part structurally, we kept the first part as 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one, as depicted in Figure 4. The second series was synthesized by 24 h refluxing of 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one with different aryl substituents in ethanol and piperidine to get three novel molecules (3a,b and 3d). In comparison, compound 3c was obtained without reflux by stirring in an ice bath for two hours then at room temperature for 12 h.

GSK3β Inhibition Assay
Screening of GSK3β inhibition activity for our new library was conducted using the Kinase-Glo luminescence assay to determine how much ATP is reduced after a kinase reaction [3]. All compounds were first tested at 10 µM and hit with a percentage inhibition of less than 50%, then were examined further to derive IC 50 values, which were calculated using the linear regression parameter. Staurosporine (II) was used as a reference compound; it showed 98.46% inhibition at 10 µM. Excitingly, one of our first version compounds (E)-2f demonstrated 76.4% inhibition, while two compounds ((E)-and (E)-2c) showed percent inhibition at 45% and the others ((E)-2d and (E)-2e) showed less than 30%.
As (E)-2f has the most potency, we tried to optimize structurally (E)-2f by using different pyridyl substituents, as shown in Figure 4. Upon measuring the IC 50 of these compounds, none had IC 50 values better than (E)-2f, which revealed an IC 50 of 1.7 µM ( Figure 5). However, the novel compounds had IC 50 of no more than 3 µM, except 3b and 3d, as shown in Table 1.
Pharmaceuticals 2022, 15, x FOR PEER REVIEW 6 of 20 5). However, the novel compounds had IC50 of no more than 3 µM, except 3b and 3d, as shown in Table 1.

Molecular Docking Study
In this study, the GSK3β protein structure (PDB ID: 4IQ6) was used because its cocrystallized ligand has azaindole chemotype, which is structurally similar to our azaindolin-2-one scaffold [29].
Prior to docking our compounds, the Molecular Operating Environment (MOE) was able to reproduce the pose of the native ligand within the active site of GSK3β protein as demonstrated in Figure 6A, where both the native and re-docked ligands made almost the same types of interactions with the critical amino acid residues. The re-docked ligand's binding free energy (∆G) equals −7.4 Kcal/mol with an acceptable Root Mean Square Deviation (RMSD) value of 1.4Å (Table 1). dolin-2-one scaffold [29].
Prior to docking our compounds, the Molecular Operating Environment (MOE) was able to reproduce the pose of the native ligand within the active site of GSK3β protein as demonstrated in Figure 6A, where both the native and re-docked ligands made almost the same types of interactions with the critical amino acid residues. The re-docked ligand's binding free energy (ΔG) equals −7.4 Kcal/mol with an acceptable Root Mean Square Deviation (RMSD) value of 1.4Å (Table 1).  Remarkably, only the E configuration of azaindolin-2-one based compounds (2f and 3a-d) can make these two strong interactions. In contrast, their Z configuration produces weak interactions due to incorporating amidic carbonyl group in the interaction rather than the strong H-bonds of the N atom of azaindolin-2-one moiety ( Figure 8A-F).
The top-ranked docking poses of indolin-2-one based compounds (2a-e) are depicted in Figure 7A-E. The NH group in the indolin-2-one moiety makes a weak H-bond with the critical residue (Asp 133) at the ATP binding site at the hinge region with the binding free energy −0.5 to −3.2 Kcal/mol. These binding energies are feeble compared to the native ligand, where its NH group makes a strong H-bond with the same residue (Asp133) with a binding free energy of −4.5 Kcal/mol. Remarkably, only the E configuration of azaindolin-2-one based compounds (2f and 3a-d) can make these two strong interactions. In contrast, their Z configuration produces weak interactions due to incorporating amidic carbonyl group in the interaction rather than the strong H-bonds of the N atom of azaindolin-2-one moiety ( Figure 8A-F).

Tau Aggregation Inhibition in a Cell Model of Tauopathy and Western Blot Analysis
Aggregation of Tau is observed using Hela cells transduced with the expression vector of tau 2N4R isoform carrying P301L mutation and fibrils of tau protein with P301L mutation as seed [36,37]. We tested the anti-aggregation effects of (E)-2f and 3a at four different concentrations, namely (1 µM, 3 µM, 10 µM, 30 µM), the measurements combined with negative seed (negative control = normal aggregation level), and positive seed (positive control = AD aggregation levels). In other words, seed-and + indicate the absence and presence of fibrils of tau protein with P301L mutation as seed, respectively. At 10 µM concentration, compound (E)-2f showed a significant anti-aggregation effect and even had less fibrillization than the negative control, while at 3 µM concentration it has moderate efficacy, similar to standard compound 3a, and demonstrated no efficacy at 1 µM (Figure 9). Surprisingly, at 30 µM concentrations, both (E)-2f and 3a completely protected Hela cells from tau aggregates ( Figure S41). (E) (F) Remarkably, only the E configuration of azaindolin-2-one based compounds (2f and 3a-d) can make these two strong interactions. In contrast, their Z configuration produces weak interactions due to incorporating amidic carbonyl group in the interaction rather than the strong H-bonds of the N atom of azaindolin-2-one moiety ( Figure 8A

Tau Aggregation Inhibition in a Cell Model of Tauopathy and Western Blot Analysis
Aggregation of Tau is observed using Hela cells transduced with the expression vector of tau 2N4R isoform carrying P301L mutation and fibrils of tau protein with P301L mutation as seed [36,37]. We tested the anti-aggregation effects of (E)-2f and 3a at four different concentrations, namely (1 µM, 3 µM, 10 µM, 30 µM), the measurements combined with negative seed (negative control = normal aggregation level), and positive seed (positive control = AD aggregation levels). In other words, seed-and + indicate the absence and presence of fibrils of tau protein with P301L mutation as seed, respectively. At 10 µM concentration, compound (E)-2f showed a significant anti-aggregation effect and even had less fibrillization than the negative control, while at 3 µM concentration it has moderate efficacy, similar to standard compound 3a, and demonstrated no efficacy at 1 (positive control = AD aggregation levels). In other words, seed-and + indicate the ab-sence and presence of fibrils of tau protein with P301L mutation as seed, respectively. At 10 µM concentration, compound (E)-2f showed a significant anti-aggregation effect and even had less fibrillization than the negative control, while at 3 µM concentration it has moderate efficacy, similar to standard compound 3a, and demonstrated no efficacy at 1 µM ( Figure 9). Surprisingly, at 30 µM concentrations, both (E)-2f and 3a completely protected Hela cells from tau aggregates ( Figure S41). . ImageJ counted more than 100 spots in the same area in each sample. *: denotes significant difference from control (seed+) at p < 0.05, **: denotes significant difference from control (seed+) at p <0.01.
Western blot analysis was then used to confirm the amount of the total tau in Hela cells used for microscopic observation. The lysate of Hela cells, after incubation with compounds (E)-2f and 3a at 10 and 30 µM concentrations, were analyzed by immunoblotting using the same anti-Tau antibody ( Figure S42). ImageJ was used to quantify the total tau at each concentration and compared to the total tau level of Hela cells with positive seed (seed+) alone as a control. As depicted in Figure 10, the total tau level at each concentration of our compounds demonstrated little or no difference compared to untreated Hela cells. . ImageJ counted more than 100 spots in the same area in each sample. *: denotes significant difference from control (seed+) at p < 0.05, **: denotes significant difference from control (seed+) at p < 0.01.
Western blot analysis was then used to confirm the amount of the total tau in Hela cells used for microscopic observation. The lysate of Hela cells, after incubation with compounds (E)-2f and 3a at 10 and 30 µM concentrations, were analyzed by immunoblotting using the same anti-Tau antibody ( Figure S42). ImageJ was used to quantify the total tau at each concentration and compared to the total tau level of Hela cells with positive seed (seed+) alone as a control. As depicted in Figure 10, the total tau level at each concentration of our compounds demonstrated little or no difference compared to untreated Hela cells.

Discussion
The structural cores of our synthesized compounds are as follows: (a) the first o an oxindole ring or its bioisostere 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one ring a methylene linker, and (c) an aryl substitution which is attached to the methylene lin Following the literature-known condensation reaction "Knoevenagel condensation" applied a strategy to optimize the main parts of our targeted compounds. Interestin all target compounds produced by 24 h refluxing of different azaindole or substit oxindoles with the corresponding aldehyde, except compound 3c, were not produ with the reflux method. Instead, stirring in an ice bath for two hours followed by stir

Discussion
The structural cores of our synthesized compounds are as follows: (a) the first one is an oxindole ring or its bioisostere 1,3-dihydro-2H-pyrrolo [2,3-b] pyridine-2-one ring, (b) a methylene linker, and (c) an aryl substitution which is attached to the methylene linker. Following the literature-known condensation reaction "Knoevenagel condensation", we applied a strategy to optimize the main parts of our targeted compounds. Interestingly, all target compounds produced by 24 h refluxing of different azaindole or substituted oxindoles with the corresponding aldehyde, except compound 3c, were not produced with the reflux method. Instead, stirring in an ice bath for two hours followed by stirring the reaction mixture at room temperature for 12 h yielded our target compound 3c.
When pyridine-2-carbaldehyde was used in condensation, only (E) isomer was produced; however, when pyridine-4-carbaldehyde or amine substituted benzaldehyde was used, a combination of E/Z isomers was produced. The unusual behavior of 2-pyridine derivatives compared to 4-pyridyl amine substituted benzaldehyde equivalents can be attributed to the secondary nonbonded contact between the nitrogen and the 4-H oxindole proton, which stabilizes the coplanar arrangement of the aromatic rings in 2-pyridine derivatives [27].
In the GSK3β inhibitory assay, compound (E)-2f had the best inhibitory activity on GSK3β with an IC 50 of 1.7 µM. Preliminary SAR of our compounds revealed that chloro substitution at positions 5 and 6 on the oxindole ring enhanced activity compared to the non-substituted oxindole. However, non-substituted oxindole showed more efficacy than 5 or 6 fluorooxindole. Surprisingly, the efficacy was significantly increased when the oxindole was replaced with its N bioisostere nucleus. Furthermore, N bioisostere demonstrated the greatest potency.
Moreover, 2-pyridyl is more effective than 4 pyridyl substituent and amine substituted phenyl. The amine substitution on the phenyl ring is more potent to be dimethylamine than diethylamine. In brief, the superior activity of (E)-2f is mainly due to azaindolin-2-one chemotype, while 2-pyridyl substitution enhances this GSK3β inhibitory activity.
Molecular simulation studies of compounds 2a-f and 3a-d were performed using Molecular Operating Environment to explain the potency difference between indolin-2-one and azaindolin-2-one scaffolds on GSK3 inhibition activity in vitro (MOE). The docking interactions and orientations of compounds 2a-e demonstrates their weak inhibitory activity against GSK3β in vitro. Briefly, indolin-2-one based compounds (2a-e) interact with the two essential amino acids (Asp 133 and Val 135) through relatively weak H-bonds compared to the native ligand. On the contrary, azaindolin-2-one, containing compounds (2f and 3a-d), made two strong H-bonds with the two essential amino acids (Asp 133 and Val 135) with their amidic NH group and N atom of the azaindolin-2-one moiety.
Compounds 2f and 3c are the only azaindolin-2-one-based compounds in the E configuration. Other azaindolin-2-one compounds (3a, 3b, and 3d) are a mixture of E and Z configurations. It is evident that the E configuration is responsible for the actual inhibitory activity of azaindolin-2-one compounds. Furthermore, the docking poses of the Z configuration of azaindolin-2-one compounds are very similar to those of the docked indolin-2-one compounds (2a-e). Therefore, compounds 2f and 3c are expected to have a better binding affinity to GSK3β protein than other azaindolin-2-one compounds (3a, 3b, and 3d). Compound 2f demonstrated the most significant inhibitory activity against GSK3 in vitro; this superior activity could be attributed to its ability to form additional hydrophobic interactions and its two strong H-bonds with the two critical amino acids.
Compound 3a has previously been demonstrated to have potent binding affinities in vitro on alpha-synuclein (α-syn), beta-amyloid (Aβ), and tau fibrils [9]. Therefore, we decided to validate the anti-aggregating characteristics of (E)-2f compared to 3a as a standard. Indeed, recombinant shortened tau isoforms have been more popular in recent years for in vitro tau aggregation. Adding anionic cofactors (e.g., heparin, arachidonic acid) results in the development of fibrils that mirror those found in the brains of AD patients. Consequently, we used Tau Aggregation Assay Kits to monitor the fibrillization of tau-transfected HeLa cells using fluorescence assay techniques. In this assay, the numbers of aggregates were assessed. Some aggregates were detected in the cells without fibrils as seeds, which may be formed by something naturally present in the cells. The number of aggregates increases when the expression vector of tau 2N4R isoform carrying P301L mutation (fibrils of tau protein with P301L mutation) is transduced into Hela cells, and our compounds suppress the aggregation. Both compounds, (E)-2f and 3a, have a concentrationdependent inhibitory activity against tau aggregation. Consequently, azaindolin-2-one scaffold in both compounds seems to be essential for anti-tau aggregation, as well as the previous GSK3β inhibition assay. One possible explanation of this anti-tau aggregation is that these compounds decrease the amount of tau expressed in Hela cells. Therefore, we performed a western blot analysis to confirm this hypothesis. The results revealed that both compounds, (E)-2f and 3a, at two different concentrations have minimal effect on total tau levels compared to untreated Hela cells. Consequently, the anti-tau aggregation effect of our compounds is due to their chaperon effect (direct binding to Tau) rather than lowering the tau expression.
Based on the findings, it can be concluded that our compounds have anti-tau aggregation activity, regardless of aggregate shapes. This tau aggregates formation is considered one of the models of tau fibrils formation.
The active site of GSK3β is highly similar to the active sites of cyclin-dependent kinase (CDK) family members, which might lead promiscuous inhibitors to induce undesirable effects [27]. CDK inhibition has been linked to peripheral blood mononuclear cells (PBMC) toxicity. To assess the cytotoxicity of our compounds on normal cells, we tested (E)-2f, which has been identified as a potent GSK3β inhibitor against PBMC viability and some selected cancerous cell lines. Moreover, compound 3a was used as a standard. (E)-2f is harmless to PBMC at doses much beyond their IC 50 values in a functional GSK3β test. In addition, (E)-2f moderately suppressed selected cancer cell lines such as K562, U251, and HCT116, while it needs a concentration over 100 µM to suppress A375.

General Procedure for the Synthesis of Compounds 2a-f
A solution of 1a-e (1 mmol) in absolute ethanol (30 mL) was added to picolinaldehyde (1 mmol) followed by a catalytic amount of piperidine (2 drops), and the mixture was refluxed for 8-12 h. After cooling to room temperature, a colored precipitate was collected by filtration, washed with absolute cold ethanol then diethyl ether. The resulting product was recrystallized from methanol to afford 2a-e.

GSK-3β Inhibition Assay
According to the manufacturer's protocol, the assay was conducted using GSK3β Kinase Enzyme System and ADP-Glo Kinase Assay (Promega Corporation, Madison, WI, USA) in 384-well plates. The kinase reaction was performed at room temperature using 2 ng/µL of human recombinant GSK3β, 0.2 µg/µL of GSK3 substrate (YRRAAVPPSP-SLSRHSSPHQ (pS) EDEEE; amino acids 636-661), and 25 µM ATP in 4X kinase buffer containing 50 µM DTT. The compounds were added to enzyme solution and incubated for 2 h at room temperature. Then, kinase activity was quantified using the ADP-Glo Kinase Assay as previously described [38].

Molecular Docking Study
All the docking algorithms, ranking, and visualization processes were performed using Molecular Operating Environment (MOE) 2019.01 (Chemical Computing Group, Montreal, QC, Canada) software within the active site of GSK-3β (PDB ID: 4IQ6), downloaded from the Protein Data Bank [29]. The protein structure (PDB ID: 4IQ6) was prepared through the QuickPrep suite executed in MOE. Then, all chemical structures of the docked compounds were built, and their energies were minimized using the default force field, AMBER10. Before docking, the native co-crystallized ligand was re-docked into the binding site using the default parameters to validate the docking study at the binding site. The top-ranked re-docked pose has an RMSD value of 1.4 Å, and its energy score was −7.4 kcal/mol. Nonessential ligands and water molecules were removed before docking. The compounds were docked in the active site using rigid receptor docking protocol, where ligand conformations were placed in the active site with the Triangle Matcher method using AMBER10 as a default force field, finally ranking the generated poses by the London dG scoring function. The generated docking poses were reviewed, and the poses of the best binding affinities were considered. All poses were generated and labeled using MOE software.

Tau Aggregation Inhibition in a Cell Model of Tauopathy and Western Blot Analysis
According to the manufacturer's protocol, the assay was performed using Tau Aggregation Assay Kit (Cosmo Bio, Tokyo, Japan [catalog number: TAU01]). Briefly, HeLa cells were transfected with pCMV-Tau(2N4R)-P301L (20 ng) using Lipofectamine 3000 (ThermoFisher Scientific, Waltham, MA, USA) and treated with F-Tau (RD)-P301L (fibrils of tau protein with P301L mutation as seed, included in the assay kit). Drugs were added after 6 h, and the cells were incubated for 1 day. Next, immunostaining was performed basically as previously described [39]. Tau proteins were stained with green fluorescence. In this staining, a TAU antibody (Proteintech, Rosemont, IL, USA [catalog number: 10274-1-AP]) was used as the first antibody, and goat anti-rabbit IgG (H + L) highly cross-adsorbed secondary antibody, Alexa Fluor 488 ThermoFisher Scientific) was used as the second antibody. An All-in-One Fluorescence Microscope BIOREVO BZ-9000 (Keyence, Osaka, Japan) was used as the microscope. ImageJ counted more than 100 spots in the same area in each sample [40,41].

Selective Cytotoxicity on Cancer Cells and Normal Blood Cells
The human cervical carcinoma cell line HeLa, obtained from the American Type Culture Collection (ATCC), was cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% heat-inactivated fetal bovine serum (FBS). The human glioblastoma cell line U251 (obtained from ATCC), colon carcinoma cell line HCT116 [provided by the RIKEN BRC through the National Bio-Resource Project of the MEXT/AMED, Japan (RCB2979)], and human melanoma cell line A375, obtained from ATCC, were cultured in DMEM supplemented with 10% heat-inactivated FBS. The human chronic myelogenous leukemia cell line K562 and human peripheral blood mononuclear cells (PBMC), both cells obtained from ATCC, were cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS described previously [43].

Conclusions
For years, we have been interested in AD disease and its multifactorial nature. In the current study, we succeeded in synthesizing a multitargeted chemotype that can deal with the multifactorial nature of AD. We designed a unique, second-in-class chemical series of dual inhibitors that inhibit GSK3β and tau aggregation based on structural resemblances and known SAR data for GSK3β and Tau aggregation inhibitors. We chose to begin with oxindole and its N bioisostere, which have been previously reported as binding motifs for inhibiting GSK3β and preventing tau aggregates' development, and destabilizing preexisting tau aggregates in a dose-dependent manner. Hybridizing these scaffolds with pyridyl group via methylene linker enhanced the GSK3β activity and ensured the antiaggregation effect. This design has resulted in (E)-2f, a low molecular weight fragment with specificity and potency at two structurally different tau targets, GSK3β and a fibrillar protein, and is safe on normal cells. Our research identified azaindolin-2-one as a promising scaffold that could be used to develop future multitargeted therapies for AD.  Figure S36. Cell viability Dose-response curve for inhibition of K562 cells; Figure S37. Cell viability Dose-response curve for inhibition of U251 cells; Figure S38. Cell viability Dose-response curve for inhibition of HTC116 cells; Figure S39. Cell viability Dose-response curve for inhibition of A375 cells; Figure S40. Cell viability Dose-response curve for inhibition of PBMC cells; Figure S41. Effects of different concentrations of seed positive (seed +), 2f (30 µM) and 3a (30 µM) on Hela cell morphology photographed under an inverted microscope 24 h after treatment compared to untreated Hela cells, seed negative (seed -); Figure S42. The whole gel of western blot analysis showing the effect of different concentrations of 3a (10 µM), 2f (10 µM) 3a (30 µM) and 2f (30 µM) on total tau proteins (A), using GAPDH as a reference (B) after 24 h treatment compared to Hela cells, seed positive (seed +).