cAMP-dependent PKA was the first kinase to be discovered in the mid-1960s, with its structure determined by X-ray diffraction. PKA is particularly important in the phosphorylation of serine and threonine residues in several substrates [13, 89]. It consists of two subunits, one called as regulatory (R), where the bond with the cAMP occurs, and another one called as catalytic (C), responsible for bonding with the substrate. When R subunit is free, a tetrameric complex is formed in subunit C, leading to enzyme inactivation. Thus, when cAMP binds to PKA in R position, the release of this site occurs, and a structural change in the molecule allows the interaction of the protein to be phosphorylated in C [90]. Into the cell, PKA is responsible for phosphorylating and activating a cAMP-responsive binding protein (CREB), which modulates synaptic plasticity and is indispensable for memory maintenance (Figure 3A). However, in AD, secretases cleave the APP and are responsible for the formation of Aβ plaques, decreasing the amount of available cAMP, making it impossible to undo the tetrameric complex of PKA, and therefore, not occurring enzyme activation. As a result, there is a drastic reduction in CREB production (Figure 3B), causing the characteristic memory impairment observed in AD [11].

The analysis of the PKA-CREB complex activity revealed a negative control of tau protein by reducing the gene transcription expressed in mRNA. This occurs by binding to cAMP response elements, which are linked to the expression of DNA genes, that suppress expression after binding to the complex. Thus, there is a reduction in the production of tau protein when phosphorylation of PKA-CREB occurs at its binding site (CRE). However, when a reduction in the kinase complex activity is observed, there is an exacerbation in tau levels, being more available to the action of phosphorylating enzymes, resulting in the hyperphosphorylation observed in AD patients. This reveals a mutual relationship between PKA-CREB and tau, and its maintenance is of paramount importance to ensure the stabilization of microtubules and to prevent the formation of NFTs [91, 92].

An in vitro study was carried out by Liu and co-workers [93] suggested that pre-phosphorylation of tau protein at specific serine sites by PKA (Ser214, Ser262, and Ser409) would enable subsequent tau phosphorylation by other PKs. Particularly, the activity of GSK-3β may be induced in Thr181 residues, Ser199, Ser202, Thr205, Thr217, Thr231, Ser396, and Ser422 of tau protein. In addition, experiments with monkey brains have shown that with increasing primate age, there is an increase in the rate of PKA phosphorylation of tau in the prefrontal cortex. In addition, in the communication of pyramidal neurons in dendritic spines, it was observed a decline of phosphodiesterase 4A (PDE4A), the enzyme responsible for controlling PKA signaling, and promoting its increased activity [94]. Thus, it is suggested that the maintenance of homeostasis in PKA activity is crucial, since its deregulation can promote decompensation of neuronal plasticity and its excess can provide the neurodegeneration, formation of NFTs, and Aβ plaques observed in AD.

Danshensu (5, Figure 4) is a phenolic acid found in Salvia miltiorrhiza and Prunella vulgaris var. described in the literature as a compound capable of acting in the CNS with anxiolytic and neuroprotective activities. Thus, Bae and co-workers [95] carried out an in vivo evaluation of this compound against cognitive impairment, especially the memory loss in mice. Western blotting analysis of the brains of animals submitted to classical behavioral assays, such as passive avoidance test, Y-maze, and open field tests, showed that compound 5 could promote a significant improvement in memory by the activation of dopamine receptors (D1) in the cerebral cortex. These receptors promote the phosphorylation of PKA on CREB, forming the active complex PKA-CREB, which is essential for the maintenance of cognitive ability, suggesting a possible mechanism of action to be explored in the treatment of cognitive dysfunctions related to AD [95].

In another study, Cui and co-workers [96] identified compound BPN14770 (6, Figure 4) as an allosteric PDE4 inhibitor, an enzyme responsible for maintaining the cytoplasmic levels of cAMP. Based on a set of experimental results, BPN14770 was highlighted as a new drug prototype candidate for AD, capable of promoting neuronal plasticity and memory preservation [96], possibly operating by a mechanism of action different from the other known approved drugs. Additionally, studies with animals with neuronal damage induced by Aβ1-42 deposition, revealed that BPN14770 promotes neuroprotective effects by activation of the PKA-CREB pathway, which in turn, induces structural and functional maintenance of important synaptic proteins, such as synaptophysin, and stimulated memory preservation in mice. The involvement of the PKA-CREB pathway was also evidenced by a comparative study of a control group treated with H-89, a potent PKA inhibitor, in which the beneficial effects of BPN14770 were not observed, corroborating the involvement of PKA and CREB as a probable mechanism of neuroprotection [96].

CDKs belong to a family of PK in which cyclin is the regulating molecule of phosphorylation activity. The vast majority of CDKs participate in the regulation of different stages of the cell cycle [13], and particularly CDK5 is not part of this group, being an enzyme found in post-mitotic cells, especially in the CNS [11]. The activation of CDK5 depends on two proteins located in the neuronal membranes, p35 and p39. The complex formed by this kinase, its activating proteins, and cyclin has a short half-life of around 25 minutes for p35, being very well-controlled. CDK5 has an important physiological role in the CNS, both during development and in the adult life of neurons. It is responsible for phosphorylating protein units involved in the synaptic plasticity, as well as action on the tropomyosin kinase B receptor (TrkB), and brain-derived neurotrophic factor (BDNF), the constitution of the cerebral cortex, migration of nerve cells during growth, and memory maintenance by regulating PKA signaling (Figure 5A) [11, 97].

Under pathological conditions of oxidative stress, overexcitation, or apoptotic signals, an excessive amount of Ca²⁺ enters the neurons, leading to the activation of calpains, which are calcium-dependent proteases that cleave p35 into p25 and p39 into p29 (Figure 5B). However, the CDK5-p25 complex is highly stable, which contributes to the hyperactivation of CDK5 and, consequently, to its dysregulation in the body [97]. In addition, there is an increase in the half-life of p25, which is up to 10-fold higher than p35 [98]. This factor promotes abnormal and uncontrolled hyperphosphorylation of proteins, which contributes to neurotoxicity and neurodegeneration, inducing the appearance or maintenance of diseases, such as AD, Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), and encephalopathies related to prions [98, 99].

In AD, CDK5 contributes to the formation of Aβ plaques by involvement in the cleavage of APP by the phosphorylation of a threonine subunit in the APP structure and by acting on the transcription factors of genes of β-secretase, which favors the synthesis of Aβ peptides. In addition, the increase of Aβ plaques in the brain leads to an increased Ca²⁺ inflow to the cytoplasm, which promotes the formation of the CDK5-p25 complex, causing enzymatic hyperactivation [97]. CDK5 is also responsible for the hyperphosphorylation of numerous serine and threonine residues in tau, besides regulating other enzymes involved in the formation of NFTs, such as GSK-3β [97].

In a recent review, Maitra and Vincent [100] highlighted four possible mechanisms of action of potential CDK5 inhibitors, acting on the conversion of p35 to p25, in the stabilization of the CDK5-p25 complex, inhibition of CDK5-p25 in the catalytic site, and reduction of CDK5 and p25 expression. An in vivo study aimed at evaluating new CDK5 inhibitor candidates led to the identification of amidopyrazole derivative 7 (Figure 6), which showed a very significant effect on the reduction of enzyme activity. This derivative was able to inhibit the activity of CDK5 in a non-selective manner, acting on the CDK5-p25 complex with a nanomolar potency (IC₅₀ = 178 nM), with good BBB permeability. Complementary studies showed a dose-dependent effect and promising druggability properties related to ADME [101].

Dendrobium nobile is an Asian orchid species constituted by a high concentration of alkaloids, awakening interest in the search for anti-Alzheimer agents. A phytochemical study led to the obtaining of a mixture containing six sesquiterpene alkaloids that were identified and quantified as dendrobine (8; 92.6%), dendrobine-N-oxide (9; 3.3%), nabilonine (10; 2.0%), dendroxine (11; 0.9%), 6-hydroxy-nobilonine (12, 0.32%) and 13-hydroxy-14-oxodendrobine (13; 0.07% Figure 7). This alkaloidal extract was administered orally in mice (40 mg/kg), and submitted to an evaluation of behavioral and biological brain profiles (Western Blot), revealing a significant decrease in stress in the endoplasmic reticulum, which is an organelle responsible for protein synthesis, as well as reduction in calpain 1, CDK5 and GSK-3β levels, and a consequent decrease in the rate of tau hyperphosphorylation. Data from the behavioral evaluation were also consistent with biochemical changes, evidenced by the improvement of cognitive functions and corroborating their neuroprotective potential against AD pathogenesis [102, 103].

The derivative 4,5,6,7-tetrachloro-2’,4’,5’,7’-tetraiodofluorescein (14, Figure 7), a dye also known as rose bengal, is a synthetic compound capable of promoting the stabilization of the cell cytoskeleton and preventing microtubule disorganization. Recently, in vivo studies have shown that this compound attenuates the cognitive deficit caused by the deposition of NFTs in Drosophila [104]. Based on this evidence, Mou’s group [105] evaluated a possible mechanism underlying this neuroprotective effect, which could be related to the inhibition of kinases, such as CDK5 and GSK-3β. In vivo assays revealed a decrease in CDK5 activity in the hippocampus, which was corroborated by in silico results, suggesting that compound 14 interacts with CDK5 through the ATP binding site. Thus, it was evidenced that the neuroprotective effect observed for compound 14 could probably be due to a reduced rate of tau hyperphosphorylation by CDK5 inhibition [105].

To study the neuroprotective effects of pyrimido[4,5-b]indole derivatives, Loidreau and co-workers [106] synthesized a series of eighteen 4-pyrimidoindole derivatives, which were evaluated for their inhibition profile over the CDK5-p25 complex. In vitro, results revealed derivative 15 (Figure 7) as the most promising CDK5 inhibitor (IC₅₀ = 6 µM), being also capable to inhibit casein kinases CK1δ and CK1ε (IC₅₀ = 0.7 µM for both), two important kinases involved in cardiac signaling [107], and 1A kinase, that is regulated by double specificity tyrosine phosphorylation (DYRK1A, IC₅₀ = 3.1 µM) [106], mainly associated with diabetes pathogenesis [108].

The protein glycogen synthase 3 was discovered in the mid-1980s [109] and is one of the most studied PKs, capable of phosphorylating serine and threonine groups in their specific substrates. Its importance was soon recognized, since its substrates are involved in various cellular signaling processes, including the regulation of cell life, cytoskeleton organization, gene transcription, and energetic metabolization, promoting the consolidation of memory, neuron formation processing, and synaptic plasticity [110]. In addition, recent studies have shown the participation of GSK-3β in the Wingless pathways (Wnt), with a capital role during embryonic development [111]. Its name derives from its ability to phosphorylate the enzyme responsible for converting glucose into glycogen, causing its inactivation [112]. GSK can occur in two isoforms, GSK-3α originated from the gene gsk3α, and GSK-3β, transcribed from the gene GSK-3β [113, 114]. Both forms have a 98% structural homology, and their differentiation occurs in the catalytic site. GSK-3β is found in higher a concentration in the CNS [110], and is capable of autophosphorylation in the tyrosine 216 residue, which promotes its activation. The inactivation process is mediated by other PKs such as PKA, PKB, and ribosomal S6 kinases of 90 kDa (p90RSK), in the serine 9 residue and phosphorylation of p38 from MAPK in the serine 389 residue [110, 114].

Recent studies have shown that the relationship between kinase activity and AD can be associated with PSEN1, a gene responsible for the action of γ-secretase enzymes and, consequently, in the APP cleavage and formation of Aβ plaques. Another relevant fact is that GSK-3β can regulate the expression of the BACE1 gene, an enzyme responsible for the synthesis of β-secretase by the NF-κβ signaling pathway, whose function is related to the triggering of inflammatory processes [115]. As a result, GSK-3β influences the increase of the abnormal cleavage of APP by β-secretase, contributing to the formation of Aβ plaques and memory impairment observed in AD (Figure 8) [116]. Finally, the association between GSK-3β and PSEN1 also promotes tau hyperphosphorylation, contributing to the formation of NFTs [113]. Studies conducted by Ly and co-workers [116] revealed that the specific GSK-3β inhibition can induce a significant reduction in the production of β-secretase, resulting in a decreased Aβ aggregation and, therefore, being a potential target for AD treatment.

One of the most cited GSK-3β inhibitors is lithium. The mechanism of action associated with this effect is not fully elucidated, but it is believed that Li⁺ can compete with Mg²⁺, preventing its binding to GSK-3β, causing a destabilization of ATP conformation in the enzyme and, consequently, inhibiting its catalytic activity [117]. Some marine-origin bioactive compounds have shown promising results as allosteric inhibitors of GSK-3β, such as palinurine (16, IC₅₀ = 1.9 μM, Figure 9) and the quinone derivative VPO7 (17, IC₅₀ = 2.8 µM, Figure 9) [117]. Based on experimental results, it has been well accepted that selective inhibition of GSK-3β could be a promising strategy for AD treatment, aiming to reduce tau hyperphosphorylation and other pathogenic factors related to neurotoxicity, protein misfolding, cognitive impairment, and motor functions, contributing to the discovery of innovative and more effective drugs.

Other studies with marine natural products led to the identification of meridianin (18, Figure 9), an indole-pyrimidine metabolite obtained from the tunicate Aplidium Savigny, and lignarenone (19, Figure 9), obtained from mollusks of the Scaphanderlignarius genus. Both compounds showed inhibitory activities of GSK-3β, promoting the increase in the phosphorylation of the Ser9 residue, which leads to a reduction in the enzymatic activity [118].

Regorafenib (20, Figure 9) is a multi-kinase inhibitor, which targets tumor cells and tumor microenvironment, approved for the treatment of bowel cancer. Further studies aimed at expanding its applications and drug repositioning, led to the evaluation of its potential neuro-anti-inflammatory properties in AD models. In vivo studies with regorafenib being administered orally to mice for two weeks, followed by immunohistochemical analysis, showed a significant decrease in Aβ plaque deposition and tau hyperphosphorylation as a result of a marked effect on the inhibition of GSK-3β in the cortex and hippocampus. In addition, regorafenib reduced cytokine-induced neuroinflammation in in vitro experiments with microglial cells [119].

A study carried out with thieno[3,2-c]pyrazole-3-amine derivatives revealed a new chemical architecture with promising inhibitory activity of GSK-3β. Particularly, derivative 21 (Figure 10) showed a surprising in vitro nanomolar inhibitory potency (IC₅₀ = 3.1 nM), reducing tau phosphorylation in Ser396 residue, without significant toxicity at concentrations lower than 50 μM in SH-SY5Y neuroblastoma cells type [120].

A group of Chinese researchers studied a new series of α-carboline derivatives, that was previously proven to act as antimicrobial, anti-inflammatory, and anxiolytic, as well as showing CNS excitatory properties. A series of eighteen new compounds were synthesized and evaluated for their effects on the GSK-3β inhibition, revealing derivative (22, IC₅₀ of 1.71 µM, Figure 10), as the most promising inhibitor, with low cytotoxicity. Additional western Blot experiments indicated that compound 22 was capable of reducing okadaic acid-induced tau hyperphosphorylation in SHSY5Y neuronal cells, which was corroborated by in silico studies. Computational data suggested that this compound could interact by H bonding involving its different nitrogen functional groups, as well as hydrophobic interactions between the carbonyl group and the catalytic site of GSK-3β [121].

The molecular framework of harmine (23, Figure 10), a known GSK-3β inhibitor, was used as a structural model in the design of a novel series of β-carboline analogues with potential neuroprotective effects. Pharmacological evaluation stood out compound 24 with the best inhibitory profile for GSK-3β (IC₅₀ = 71 nM), being 450-fold more potent than the prototype harmine (IC₅₀ = 32.1 μM). In addition, this compound strongly inhibited DYRK1A (IC₅₀ = 103 nM), apparently competing with ATP, and a Western Blot analysis confirmed the inhibition of tau hyperphosphorylation in SH-SY5Y cells treated with okadaic acid. Further animal studies with APP/PS1/tau transgenic mice in the classic Morris aquatic labyrinth (MWM) behavioral model showed that administration of compound 24 led to a significant improvement in cognitive and memory capacity, with good BBB permeability. In silico studies evidenced the interaction of compound 24 at the ATP-binding site of GSK-3β, revealing some key structural subunits for structural recognition and affinity, such as the carbonyl ester function that acts as an H-binding acceptor group. Despite its moderate cytotoxicity, compound 24 was considered an interesting drug candidate prototype and a useful model for the development of new drug candidates against AD [122].

Quinoxalines have also been explored for their potential as GSK-3β inhibitors, leading to the design of different architectures of structural analogues, preserving the quinoxaline nucleus as a central pharmacophore. In a recent example, Kumar Jain and co-workers [123], searching for novel quinoxaline-based GSK-3β inhibitors, synthesized a series of compounds, in which the quinoxaline nucleus is connected to an aromatic system through a hydrazone spacer subunit. Diverse substituents on the phenyl moiety were used to furnish structure-activity relationships and their pharmacophoric insights. As a result, derivatives 25 and 26 (Figure 11) were highlighted for their best results in the selective inhibition of GSK-3β, with IC₅₀ values of 0.27 and 0.39 μM, respectively. Molecular docking studies suggested their good affinity for the enzyme due to a H-bond interaction with a Val135 residue, and hydrophobic interactions with Arg141 and Thr138 residues, which would be the main recognition sites in GSK-3β [123].

Bis(propyl)-cognitin (B3C, 27, Figure 11) is a dimer derived from tacrine (28), with the presence of three methylene groups intercalating two 4-aminoquinoline fragments, with improved AChE inhibitory activity. To evaluate its potential therapeutic effects against AD, Hu and co-workers [124] investigated possible mechanisms of action of B3C in the attenuation of Aβ-induced neurotoxicity. In an in vitro model for neuroprotection, using PC12 cells treated with Aβ1-42 fibrils, it was observed a significant reduction of 81% in the GSK-3β activity. This result was then confirmed in vivo, in which a western blot analysis confirmed that the treatment of mice with compound 27 resulted in a significant reduction of phosphorylation of Ser9 residue due to GSK-3β inactivation [124].

Liang and co-workers [125] proposed a molecular simplification of iso-orientin 6-C-glycosyl compound 29 (Figure 11), a flavonoid with neuroprotective properties against NFTs and selective and reversible GSK-3β inhibition, in the structural design of a new series of glucosyl-flavonoid analogues. Thirty new compounds were synthesized and biologically evaluated, revealing compound 30 as the most potent analogue (IC₅₀ = 0.59 μM) in the inhibition of GSK-3β, being 300-fold more potent than the original prototype 29 (IC₅₀ = 184.9 μM). In addition, molecular modeling studies revealed important hydrophobic interactions that may explain its selectivity for the GSK-3β in detriment to the GSK-3α isoform [125].

Searching in the literature for small molecules capable of inhibiting GSK-3β, Zhang’s group [126] carried out a virtual screening approach, leading to the selection of twenty new potential bioactive ligands. Further biological studies, highlighted compounds 31 and 32 (Figure 12) for their strong in vitro inhibitory activity of GSK-3β (> 70%) at a concentration of 100 µM, and IC₅₀ values of 8.48 and 2.19 µM, respectively. In addition, the amino acid residues Ile62, Val70, Tyr134, and Leu18 were identified as the main interaction sites of both compounds, being described as promising prototypes for the development of novel GSK-3β inhibitors with therapeutic potential against AD, especially for compound 32 which showed a 4-fold higher potency than 31 [126].

One of the most used and versatile strategies for the rational design of new bioactive molecules is molecular hybridization (MH). This approach is based on the combination of two or more pharmacophoric or auxophoric subunits from different bioactive compounds, originating a new single hybrid molecular architecture, that preserves or reproduces the pharmacological properties of the original prototypes [127–132]. By this approach, medicinal chemists have used MH for planning new bioactive chemical entities with innovative structural features, as well as improved potency and/or affinity, reduced potential drug-drug interactions and side effects, and especially for the design of multifunctional or multi-target directed ligands (MTDLs) [133]. Based on these principles and considering that AD is known for its multifactorial pathophysiology, with many concomitant changes in biochemical pathways, cellular dysfunctions, and imbalanced protein processing, the combination of different pharmacophores in a single molecule seems to be a promising approach to access bioligands capable to modulate multiple molecular targets concomitantly, including inhibition of PK, in addition to inhibition of AChE, monoamine oxidase B (MAO-B), BACE1, metal chelation, antioxidant, and anti-inflammatory properties [129, 130, 134–147].

Tacrine (28) and pyrimidone 33 (Figure 13) were used as prototypes to generate a new series of 4-aminoquinoline-piridyn-pirimidinone molecular hybrids. The biological evaluation addressed different targets related to AD pathogenesis led to the identification of compound 34 (Figure 13), showing a potent dual pharmacological profile by inhibiting both AChE (IC₅₀ = 51.1 nM) and GSK-3β (IC₅₀ = 89.3 nM). In addition, compound 34 reduced tau phosphorylation in Ser199 and Ser396 residues in SH-SY5Y cells, improving memory skills and spatial orientation in mice behavioral models, without significant cytotoxicity [148].

In another approach, Jiang and co-workers [149] also used tacrine as an AChE inhibitor prototype in the MH with G1 (35, Figure 13), an ATP competitive inhibitor with high affinity and ability to reduce in vivo tau hyperphosphorylation, to generate a new scaffold of thiazolopyridyl tetrahydroacridines, designed as potential dual cholinesterase-GSK-3 inhibitors. Pharmacological evaluation of 28 synthetic compounds revealed derivative 36, with a nanomolar inhibitory potency of both GSK-3β and AChE, with IC₅₀ values of 22.2 and 1.2 nM, respectively. Further studies confirmed that this compound was effective in the inhibition of tau phosphorylation, contributing to the decrease in neurofibril deposits in the intraneuronal environment. In vivo experiments based on a behavioral scopolamine-induced amnesia assay evidenced that compound 36 was also able to restore memory and cognition in mice, suggesting it could be a promising multitarget ligand for further development addressed to AD therapy [149].

Oukoloff’s group [150] explored the combination of the structures of tacrine and valmarin (37, Figure 14), connected by an alkyltriazole fragment, to generate a new hybrid molecular architecture represented by compound 38 (Figure 14). This compound was highlighted in a series of eight tacrine-valmarin hybrids, which was evaluated in vitro against AChE and GSK-3β, exhibiting a highly potent dual inhibitory effect on human AChE (IC₅₀ = 9.5 nM) and GSK-3β (IC₅₀ = 7 nM). Further in silico studies suggested that the triazole subunit in the spacer moiety plays an important role in the molecular recognition, acting as an auxophore in the inhibitory activity of both enzymes. Moreover, compound 38 showed low toxicity in both human neuroblastoma (SH-SY5Y) and HuH7 liver cell cultures [150], representing another important contribution to the development of novel multitarget-directed drug candidates against AD.

In 2015, Sivaprakasam and co-workers [151] reported pyrrolopiridinone derivative 39 (Figure 14) as a novel orally in vivo potent GSK-3β inhibitor. Based on these findings, the structure of compound 39 was used as a bioactive model for combination with the metal chelator 1,2-diamine fragment 40 (Figure 14), aimed to generate a new molecular pattern with multifunctional properties against AD. The rationale of this designing approach took into account that imbalanced concentration of biometals, such as Cu²⁺, Fe³⁺, and Zn²⁺, and consequent oxidative stress and exacerbation in the ROS production are intimately associated with neuronal dysfunction and AD progress. As a result, a new series of acylamidopyridine hybrids was synthesized and biologically evaluated, leading to the identification of the imine derivative 41 (Figure 14) as the most potent GSK-3β inhibitor (IC₅₀ = 49 nM), reducing Aβ aggregation and tau hyperphosphorylation, besides a metal chelating ability for Cu²⁺ and Al³⁺. In addition, compound 41 was able to reduce ROS formation, exhibiting a significant antioxidant activity (IC₅₀ = 7.3 μM) and neuroprotective effect in SH-SY5Y and PC12 cell cultures, without considerable cytotoxicity. Altogether, these results suggest derivative 41 as a new multifunctional ligand to be developed as a promising drug candidate prototype for AD therapy [152].

In another approach for searching new dual GSK-3β/AChE inhibitors, Jiang and co-investigators [153] used donepezil and pyridylthiazole derivatives 42 and 43 (Figure 15), which had been already described as GSK-3β inhibitors, as molecular prototypes in the design of a new series of N-benzyl-pyridylamide hybrids. As a result, 60 new compounds were synthesized and evaluated in vitro, resulting in the identification of the 4-fluorbenzyl derivative 44, which exhibited the best dual inhibitory profile on AChE (IC₅₀ = 0.1 nM) and GSK-3β (IC₅₀ = 31 nM), also inhibiting BuChE (IC₅₀ = 5.3 μM). Molecular docking studies indicated that the isonicotinamide fragment of 44 was able to interact at the ATP catalytic site through a hydrogen bond, reinforcing the affinity for GSK-3β. In addition, this compound showed adequate BBB permeability and neuroprotective properties, reducing glutamate toxicity in HT-22 cells and tau hyperphosphorylation in a dose-dependent manner in neuroblastoma N2a cells. Thus, hybrid compound 44 was featured as a promising drug candidate prototype for AD, with an innovative neuroprotective mechanism of action [153].

Based on the previous results obtained for compound 44 and still exploring the pyridylthiazole core as a pharmacophore for selective kinase inhibition, Jiang’s group [154] proposed a second generation of dual cholinesterase/GSK-3β inhibitors, based on the MH of the selective GSK-3β inhibitor 45 (Figure 16) and the N-benzyl-piperidine subunit form donepezil. The biological evaluation of 29 synthetic N-benzyl-piperidine-pyridylthiazolediamide hybrids, highlighted compound 46 as the most promising multifunctional ligand, exhibiting an improved dual inhibitory action of GSK-3β (IC₅₀ = 0.3 nM) and AChE (IC₅₀ = 0.3 μM). Moreover, in vivo studies evidenced its anti-inflammatory properties and good bioavailability in brain tissue after oral administration, as well as its ability to reduce LPS-induced ROS production and neuroprotective activity in neuronal cells [154]. Overall, these results not only revealed compound 46 as a promising innovative disease-modifying drug candidate prototype, but also highlighted the advantage of exploring the pyridylthiazole core in the search for multifunctional ligands suitable for further drug development against AD.

Donepezil (1) was also used as a structural model in the design of a novel molecular architecture of a series of potential multitarget-directed ligand candidates, in combination with chromone subunit 47, which has been described as a pharmacophore related to MAO inhibition and antioxidant activity, and with the dihydropyrimidine subunit present in SQ 32926 (48) and its triazine bioisoster (49) (Figure 17), described as acting on Ca²⁺ receptors and inhibiting GSK-3β, respectively. As a result, fourteen new hybrids were synthesized, and their biological evaluation revealed the racemic derivative 50, as the best multifunctional ligand, inhibiting both GSK-3β (IC₅₀ = 29.1 μM) and AChE (IC₅₀ = 0.06 µM). Further evaluation of the isolated enantiomers of 50 revealed (+)-50 as the eutomer, exhibiting multiple and selective inhibitory properties of GSK-3β (IC₅₀ = 10.6 μM), MAO-B (IC₅₀ = 11.2 μM) and AChE (IC₅₀ = 0.09 µM), whereas the distomer (–)-50 also exhibited multifunctional effects, but with a 2-fold lower potency on GSK-3β (IC₅₀ = 24.3 μM), being equipotent against MAO-B (IC₅₀ = 12.3 μM) and AChE (IC₅₀ = 0.07 μM). In addition, (+)-50 exhibited significant antioxidant and neuroprotective activities, affecting Ca²⁺ channel functionality, reducing Aβ and tau aggregation, and the consequent senile plaque formation and NFTs deposits, respectively, as well as good BBB permeability. Altogether, these biological data highlight compound 50, and particularly the (+)-50 isomer, for its potential innovative contribution to the development of new drugs with improved pharmacological mechanisms of action against AD [155].

MARKs are a group of calcium/calmodulin-dependent proteins, consisting of four enzymes involved in several cellular processes, and associated with the regulation of the microtubule dynamics by phosphorylation of tau protein in the cell cycle of CNS cells, transduction of intracellular nerve signals, maturation of axons, and in conservation of synaptic plasticity [11, 156]. MARK1 is involved in the microtubule stabilization and signaling processes by the Wnt pathway. The MARK2, in turn, is present in several cellular processes, involved in the regulation of the cell cycle by Cdc25 phosphorylation, neuronal migration, signaling processes of the Wnt pathway, microtubule stabilization, determination of cell polarity and gene transcription by HDAC phosphorylation (histone deacetylases, responsible for inhibiting gene transcription in the nucleus). Similarly, MARK3 is also present in the Wnt signaling pathways, being involved in microtubule stabilization, cellular signaling pathways linked to Ras/Raf, and gene transcription by HDAC phosphorylation. In addition, MARK3 was identified as being associated with the phosphorylation of PTPH1, an important protein in cell signaling, and PKP2, the protein involved in the cellular junction. Finally, like the other three MARK isoforms, MARK4 plays a crucial role in the microtubule dynamics. MARK4 is also involved in programmed cell death, since it is upregulated in the early stages of ischemic events. However, experimental data from different studies have also suggested that the upregulation of MARK4 increases the likelihood of neurons to die, exerting a central role in neurodegeneration [156, 157]. In AD, it was evidenced that MARK2 and MARK4 have an increased enzymatic capacity, being able to hyperphosphorylate tau in the Ser262 residue. Furthermore, the genes encoding MARK4, which is more active than MARK2, are highly expressed in AD [11, 156]. Thus, after tau phosphorylation, MARK4 seems to be more likely to suffer hyperphosphorylation by other kinases, such as GSK-3β, CDK5, MAPK, and even by MARK itself, which leads to exacerbation in the microtubule destabilization and the consequent formation of NFTs, a well-known hallmark in the AD’s pathophysiology [156].

Studies on the development of selective MARK4 inhibitors as drug candidates have gained growing attention in recent years. In 2020, Turab Naqvi and co-workers [12] published a very interesting review highlighting the most recent insights related to the state-of-art of MARKs functions and their relevance as targets to address innovative and improved disease-modifying drug candidates for AD.

Particularly interested in the modulation of MARK4 activity related to AD pathogenesis, Voura and co-workers [158] explored a series of acridone derivatives as potential MARK4 inhibitors. Biological studies led to the identification of derivatives 51, 52, and 53 (Figure 18), as the most potent and selective MARK4 inhibitors, being capable of inhibiting this isoform in a dose-dependent manner with IC₅₀ values of 1.80, 2.20, and 4.5 μM, respectively. Further in vitro experiments evidenced that these three compounds were capable of significantly reducing tau hyperphosphorylation in SH-SY5Y cells (at 10 μM concentration), but all promoted an increase in ROS production [158].

Shamsi and co-workers [159] carried out in silico studies to evaluate the activity of donepezil and rivastigmine, two well-known selective AChE inhibitor drugs used in AD clinics, regarding their inhibitory properties on MARK4. The results showed that both compounds were able to significantly inhibit this kinase in a very similar potency, expressed by their IC₅₀ values of 5.3 and 6.74 μM, respectively. Apparently, the MARK4 inhibition profile is due to the possible interaction of both compounds with the ATP binding site, blocking its action in MARK4. Thus, these results were considered a relevant starting point in the design of drug candidates, with an innovative dual-target-based mechanism of action, targeting the concomitant inhibition of AChE/MARK4, and expanding the therapeutic application of the parent-approved drugs [159].

In 2023, Alrouji and co-workers [160] reported that huperzine A (54, Figure 18), a known natural product with a remarkable AChE inhibitory activity, was also able to inhibit MARK. In vitro studies showed that huperzine A significantly inhibits MARK4 (IC₅₀ = 4.91 μM), and subsequent computational studies suggested that its affinity for this kinase could be a result of H-bond interactions in the active site. Given these experimental data, huperzine A, which has been widely reported in the literature as a prototype of AChE inhibitors, has its importance expanded as a structural model for the design of new multifunctional inhibitors for the development of new drugs against diseases related to MARK4 [160].

In another computational approach, a 1,280 structurally diverse synthetic bioactive library was studied, leading to the identification of 6 new potential MARK4 inhibitors. Further in vitro studies revealed compounds PD-173952 (55) and PD-166285 (56, Figure 19) as the most prominent nanomolar inhibitors with IC₅₀ values of 3.3 and 3.5 nM, followed by PF-431396 (57, IC₅₀ = 11 nM), sunitinib malate (58, IC₅₀ = 38 nM), DMH-4 (59, IC₅₀ = 0.27 µM), and PHA 767491 (60, IC₅₀ = 0.8 µM). Interestingly, the most potent compound 55 showed a nanomolar inhibitory range on the other MARKs (IC_{50 MARK1} = 10 nM, IC_{50 MARK2} = 52 nM, and IC_{50 MARK3} = 100 nM), showing a 3.3, 15.7, and 30.3-fold higher selectivity for MARK4, respectively. On the other hand, compound 56, which showed to be equipotent to 55 in the inhibition of MARK4, was not significantly selectively related to other MARK isoforms. Despite their different inhibitory potencies, compounds 57 and 59 showed similar selectivities for MARK4, while sunitinib malate exhibited a non-selective nanomolar inhibition of MARK2 (IC₅₀ = 13 nM) and MARK3 (IC₅₀ = 2.7 nM). Conversely, compound 60, identified as the weaker MARK4 inhibitor, showed a 2.4 and 5-fold higher selectivity for MARK1 (IC₅₀ = 0.33 μM) and MARK2 (IC₅₀ = 0.16 μM), respectively. Importantly, compound 60 has been already described as a nanomolar GSK-3β inhibitor (IC₅₀ = 22 nM), reinforcing its multi-target pharmacological profile and its importance for further development as a drug candidate with an innovative mechanism of action against AD [161].