2.1.1 GSK-3β and BACE-1 inhibition

Based on the above-mentioned findings, GSK-3β and BACE-1 have emerged as ideal candidates for the multitarget approach based on the tau and amyloid cascade, respectively. Therefore, simultaneous GSK-3β and BACE-1 inhibition represents an innovative breakthrough for AD treatment.

Prati et al.⁹⁴ reported novel hybrid compounds combining a guanidine moiety and a cyclic amide group as both BACE-1 and GSK-3β inhibitors. The fluorinated derivative (1, Figure 2) showed an IC₅₀ of 18.03 µM to inhibit BACE-1 activity and an IC₅₀ of 14.67 µM to inhibit GSK-3β activity in vitro. This compound exhibited an anti-inflammatory profile able to reduce iNOS expression induced by lipopolysaccharide (LPS) in primary rat glial cells, without any toxicity in glial or neuronal cells. Interestingly, it showed neurogenic properties. Regarding its pharmacokinetic profile, it had good oral bioavailability (66%) and a good blood–brain barrier (BBB) permeability. In an effort to optimize 1 properties, the triazinone core was modified expanding its aromatic substitution pattern, the 3,4-dihydro-triazinone carbonyl oxygen was replaced by sulfur to access 3,4-dihydro-triazinthiones and the inclusion of different N-alkyl or N-aryl groups at the exocyclic C6-amino group.⁹⁵ The N-propyl derivative (2, Figure 2) was the most potent GSK-3β inhibitor (IC₅₀ = 4.34 µM), while it exhibited moderate BACE-1 inhibitory potency (IC₅₀ = 36.83 µM). Compound 2 showed anti-inflammatory properties reducing nitrite production and iNOS expression in primary rat glial cells. Interestingly, compound 2 kept its neurogenic properties at 10 µM, being able to differentiate neuronal stem cells into mature neurons.

Taking a different approach to identifying dual BACE-1/GSK-3β inhibitors, Di Martino et al.⁹⁶ reported a novel MTDL based on curcumin.⁹⁷ Regarding BACE-1 inhibition, novel derivatives showed improved inhibition compared with curcumin-β-keto-enol tautomer with IC₅₀ values ranging from nanomolar to low micromolar (IC₅₀ between 40 and 2.69 µM) and increased GSK-3β inhibitory potency (IC₅₀ values from 0.53 to 16.99 µM). Derivatives 3 and 4 (Figure 2) were the most potent BACE-1 and GSK-3β inhibitors, with an IC₅₀ to inhibit BACE-1 of 1.65 and 0.04 µM, and IC₅₀ toward GSK-3β of 0.53 and 2.49 µM, respectively. Although compounds 3 and 4 showed uncertain capability to cross the BBB, they did not show any cytotoxic effects in the T67 cell line. Moreover, compound 4 increased the expression of NAD(P)H dehydrogenase quinone 1 (NQO1), a phase II antioxidant enzyme that catalyzes the reduction of quinones to hydroquinones by scavenging superoxide molecules.⁹⁸

Paudel et al.⁹⁹ described a triple target, with compound 5 being able to inhibit cholinesterases, BACE-1, and GSK-3β. Arylbenzofuran (5, Figure 2) was able to inhibit AChE, BuChE, BACE-1, and GSK-3β with IC₅₀ values of 4.61, 1.51, 0.73, and 6.36 µM, respectively. Derivative 5 also inhibited Aβ_1-42 self-induced and Aβ_1-40 AChE-induced aggregation in a concentration-dependent manner being more potent than reference drugs curcumin and donepezil. Similarly, in 2021, Wang et al. described a notopterol derivative with furacoumarin as a scaffold also with triple AChE/BACE-1/GSK3β inhibitory activity, having an IC₅₀ of 1.0, 20.0, and 15.0 μM, respectively. This compound (6) also showed good BBB penetrability, suitable bioavailability, and oral safety. Interestingly, it improved memory and learning in an AD in vivo model induced by Aβ.¹⁰⁰

Similarly, Nozal et al.¹⁰¹ reported a novel 1,2,3-triazole derivative designed to inhibit BACE-1, GSK-3β, and LRRK2 (leucine-rich repeat kinase 2). LRRK2 plays a potential role in AD as it resides within a region on chromosome 12, being linked to LOAD cases,¹⁰² and it is a key modulator of neuroinflammation, as LRRK2 inhibition attenuates this process and cytotoxicity in in vivo AD models.¹⁰³ Derivative 7 (Figure 2) showed an IC₅₀ value of 3.31 µM to inhibit BACE-1 activity and IC₅₀ values of 2.36 and 0.82 µM to inhibit both GSK-3β and LRRK2, respectively. Derivative 7 also exhibited an interesting neuroprotective profile against the tau phosphorylation cellular toxicity model at 10 µM and the reduction of Aβ-induced neurotoxicity, showing a significant decrease in Aβ_1-42 levels in extracellular fluids at the same concentration. Further studies may be performed to determine the BBB penetration profile of this molecule.

2.1.2 GSK-3β and tau aggregation inhibition

Tau cascade is recognized as a promising target for drug development for AD treatment.¹⁰⁴ In this context, recent studies described novel MTDLs able to inhibit both tau aggregation and GSK-3β, a mixture of activities that could, potentially, reduce hyperphosphorylated tau inside neurons. Following this hypothesis, Gandini et al.¹⁰⁵ reported the design, synthesis, and biological evaluation of novel 2,4-thiazolidinedione derivatives. The N-heterocyclic derivatives, with the indole (8, Figure 3) and the 5-methoxy-N-methylindol derivative (9, Figure 3), inhibited GSK-3β at micromolar range (IC₅₀ values of 4.93 and 0.89 µM, respectively). Interestingly, both compounds were able to inhibit 60% and 80% tau aggregation at 10 µM, respectively, measured by Thioflavin T (ThT) fluorescence assay using the tau-derived hexapeptide Ac-VQIYK-NH₂ (AcPHF6). Moreover, they were able to cross the BBB by passive permeation. Both derivatives were able to improve cell viability in the okadaic acid-induced tau hyperphosphorylation cellular toxicity model at 10 µM. Additionally, derivative 9 also displayed inhibition of both the truncated ¹⁸K-tau fragment and full-length tau aggregation, as demonstrated by a heparin-induced tau assembly assay in which the fibrillization of the truncated ¹⁸K-tau fragment or the full-length 2N4R tau was monitored by ThT fluorescence.

Thereafter, Ali et al.¹⁰⁶ reported novel dual GSK-3β/tau aggregation inhibitors based on 3-arylidene-indolin-2-one structure. Compounds 10 and 11 (Figure 3) were identified as potent GSK-3β inhibitors (IC₅₀ values of 1.7 and 1.9 µM, respectively) with a dose-dependent tau ^301LP anti-aggregation effect in HeLa cells level (45% tau aggregation reduction for compound 10 and 40% for compound 11, both at a 10 µM concentration). Additionally, the cytotoxic effect of the compounds was assessed in K562, U251, HCT-116, A375, and PBMC cell lines for 72 h with the lowest LD₅₀ values of 12.8 µM (HCT-116) and 4.7 µM (K562) for compounds 10 and 11, respectively.

2.1.3 GSK-3β and Aβ aggregation inhibition

SAR502250 (12, Figure 4) is a potent GSK-3β inhibitor (IC₅₀ = 12 nM)¹⁰⁷ evaluated in neuroprotection and neuropsychiatric AD models by Griebel et al.¹⁰⁸ This compound was able to attenuate tau hyperphosphorylation in the spinal cord and the cortex of P^301L human tau transgenic mice. Moreover, it prevented neuronal death of primary hippocampal neurons cultures treated with Aβ neurotoxic fragment Aβ_25–35. Considering behavioral performance, SAR502250 improved cognitive deficit of aged transgenic APP^(SW)/Tau^(VLW) mice or adult mice after infusion of Aβ_25–35.

In 2020, Shi et al.¹⁰⁹ described novel 1H-pyrrolo[2,3-b]pyridine family of compounds designed as MTDLs being GSK-3β inhibitors and metal chelators. Lead compound 6-(5-(4-((pyridine-4-ylamino)methyl)phenyl)−1H-pyrrolo[2,3-b]pyridine-3-yl)quinoline-8-ol (13, Figure 4) showed high potency to inhibit GSK-3β (IC₅₀ = 66 nM). GSK-3β also plays a central role in the regulation of the Wnt/β-catenin signaling pathway.¹¹⁰ In the resting state, GSK-3β phosphorylates β-catenin, triggering its proteasomal degradation maintaining it at a very low concentration in the cytosol.¹¹¹ Therefore, GSK-3β inhibition can lead to stabilized β-catenin, which translocates to the nucleus and activates the transcription of its target genes.¹¹² Compound 13 increased β-catenin abundance in a dose-dependent manner (β-catenin/glyceraldehydes 3-phosphate dehydrogenase (GAPDH) ratio increased from 0.30 of the control to 0.46, 0.57, 0.83 at 5, 10, and 20 µM, respectively) and the inactive form of GSK-3β phosphorylated at Ser⁹ (^pGSK-3β/GAPDH ratio increased from 0.35 of the control to 0.53 at 20 µM) in SH-SY5Y cells. Additionally, compound 13 decreased Ser³⁹⁶ phosphorylated tau levels in a concentration-dependent manner when cells were treated with 20 µM of Aβ_25-35 (^pTau/GAPDH ratio decreased from 0.83 of the control to 0.33 at 20 µM). Interestingly, compound 13 demonstrated Fe²⁺, Zn²⁺, Cu^2+, and Al³⁺ chelation capacity at 20 µM. Finally, 13 inhibited Aβ_1-42 aggregation (55.1% inhibition at 40 µM) and reduced the Aβ_1-42 aggregation (61.8% reduction at 40 µM) being more potent than reference drug curcumin. Considering its chelation capacity, 13 inhibited the Cu²⁺-induced Aβ_1-42 aggregation by 66.5% at 40 µM, being more potent than the reference compound clioquinol. Interestingly, 13 increased the mRNA expression of neurogenesis markers (growth-associated protein 43 (GAP43), N-myc proto-oncogene protein (N-myc), and microtubule-associated protein 2 (MAP-2)) and promoted neurite outgrowth at 10 µM in SH-SY5Y cells.

2.1.4 GSK-3β and CDK5 inhibition

CDK5 is a proline-directed serine-threonine kinase, which participates in the regulation of neuronal movements, such as neuronal migration and differentiation, regulation of synaptic plasticity, neurotransmitter release, and memory consolidation. However, under pathological conditions, CDK5 is activated¹¹³ promoting aberrant APP, tau protein hyperphosphorylation, and neurofilaments, thus accelerating AD development. CDK5 deregulation contributes to Aβ production, formation of NFTs, synaptic damage, mitochondrial dysfunction, and neuronal apoptosis.¹¹⁴ CDK5 and GSK-3β deregulation has been implicated in abnormal tau hyperphosphorylation in AD.¹¹⁵ The ATP binding site of CDK5 shares a high homology with the corresponding ATP site of GSK-3β, thus a lack of selectivity over this kinase may be desirable; in this context, compounds that inhibit both enzymes may provide increased efficacy.¹¹⁶

Holzer et al.¹¹⁷ reported novel dual CDK5/GSK-3β inhibitors based on the benzofuropyridine scaffold. Previous findings showed that the benzofuropyridine moiety was able to inhibit cyclin-dependent kinase 1 (CDK1).¹¹⁸ Successfully, the 3-ethoxy-4-phenylbenzofuro[2,3-b]pyridine-6-ol derivative (14, Figure 5) exhibited interesting triple inhibitory capacity with IC₅₀ values of 0.17, 0.08, and 0.46 µM for CDK1, GSK-3β and CDK5, respectively. Moreover, this compound reduced tau phosphorylation by 61% at 8 µM and inhibited tau aggregation by 65% at 10 µM.

Reinhardt et al.¹¹⁹ showed that CDK5 knockdown was not sufficient to afford protection, indicating that the combination of GSK-3β and CDK5 inhibition is required to protect human iPSC-derived motoneurons against inflammatory-like conditions. Using a phenotypic screening, Reinhardt et al.¹¹⁹ identified a pyrazolo[1,5-a][1,3,5]triazine derivative (15, Figure 5) as an active hit. Compound 15 inhibited both GSK-3β (IC₅₀ = 1.4 µM) and moderately inhibited CDK5 (20% at 1 µM) and protected neurons in vivo in zebrafish models of motoneuron degeneration and AD being a preliminary hit compound for optimization.

Another line of research has been directed to indirubins, extensively studied as potential dual CDK5 and GSK-3β inhibitors. Leclerc et al.¹²⁰ reported the synthesis and pharmacological evaluation of a series of indole derivatives and dimers. Novel derivatives showed potent CDKs (IC₅₀ = 50–100 nM) and GSK-3β (IC₅₀ = 5–50 nM) inhibitory capacity. Authors demonstrated that indirubins inhibit GSK-3β by binding at the ATP binding site, similar to their binding to CDKs. From this chemical family, indirubin-3'-monoxime derivative 16 (Figure 5) inhibited human recombinant tau hyperphosphorylation in a dose-dependent manner with an IC₅₀ value of 100 nM. Moreover, 16 also reduced dopamine and 3’,5’-cyclic adenosine monophosphate-regulated neuronal phosphoprotein 32 (DARPP-32) phosphorylation by CDK5 in striatum area brain slices at 10 µM.

To find a novel chemical core with this combination of targets, Xie et al.¹²¹ performed a virtual screening toward GSK-3β and selected compounds with high binding scores that were thereafter docked toward CDK5. Compounds 17 and 18 (Figure 5) showed the highest energy binding capacities being selected as potential dual GSK-3β and CDK5 inhibitors, although they were not experimentally tested. In vitro evaluation of hit compounds demonstrated their neuroprotective capacity against Aβ_25-35-induced toxicity in human neuroblastoma cells, reducing cell death by 14% and 17%, respectively, at 10 µM.

2.1.5 GSK-3β inhibition and NRF2 induction

As described, OS is one of the most prominent biochemical alterations in AD.¹⁷ There is evidence of neurotoxic effects of hyperphosphorylated tau, Aβ aggregates, and OS, and more importantly, there is an intense crosstalk between these factors creating a pathological feedback loop, intensifying neuronal damage and accelerating cognitive decline.⁴²

Cells exposed to increased ROS activate multiple signaling pathways to counteract or mitigate this aggression. Among them, the NRF2/electrophile response element (EpRE) transcriptional pathway is considered the most important antioxidant response, promoting the expression of numerous antioxidant and anti-inflammatory enzymes.¹²² As mentioned above, GSK-3β constitutes an NRF2 negative regulation mechanism, which phosphorylates NRF2 creating a recognition site for β-Transducing repeat-containing protein (β-TrCP). β-TrCP leads to Cullin-1/Rbx1-mediated NRF2 ubiquitination and its subsequent degradation.⁵⁹ Thus, GSK-3β promotes NRF2 cytosolic localization, inhibiting its transcriptional activity and blocking its antioxidant and cytoprotective functions. Considering these insights, Gameiro et al.¹²³ reported the first compound targeting GSK-3β and NRF2 simultaneously as a novel approach for AD therapy. The chosen 6-amino-4-(3-methoxyphenyl)−3-methyl-2,4-dihydropyrano[2,3-c]pyrazole-5-carbonitrile (19, Figure 6) was able to inhibit GSK-3β with an IC₅₀ value of 7.2 µM and to induce NRF2 with a CD (concentration required to double the specific luciferase reporter activity) value of 11.3 µM. Additionally, 19 showed neuroprotective properties at 1 µM against the okadaic acid-induced tau hyperphosphorylation and the rotenone/oligomycin A-induced OS model¹²⁴ with protection percentages of 56.2% and 55.3%, respectively. Interestingly, compound 19 showed anti-inflammatory activity being able to reduce nitrite and ROS production in primary glial cell cultures challenged with LPS (IC₅₀ nitrite reduction = 12.9 µM; IC₅₀ ROS reduction = 18.1 µM). ROS reduction upon LPS treatment was related to a decrease of NADPH oxidase (NOX) activity in a concentration-dependent manner. Moreover, further studies demonstrated that compound 19 was able to reduce TNF-α release induced by LPS by 33% at 10 µM, by 50% at 30 µM and iNOS expression at 30 µM to almost basal levels in primary glial cultures.

2.1.6 GSK-3β and HDAC inhibition

Recently, epigenetic modifications have been considered a promising strategy for AD treatment.¹²⁵ The role of histone acetylation in rescuing learning and memory impairment has promoted its regulation through histone deacetylases (HDACs) as a novel AD treatment strategy.^{124, 126} Histones are building blocks of the nucleosome, the chromatin fundamental unit. Histone acetylation plays an important role in regulating chromatin condensation and gene transcription. Thus, histone acetylation activates transcription while histone deacetylation suppresses gene transcription. These acetylation and deacetylation processes are catalyzed by histone transferases (HATs) and HDACs, respectively, in a balanced state; however, in AD, this homeostasis is disturbed.¹²⁷ Therefore, HDAC inhibitors have gained increasing interest as potential AD drugs.¹²⁸

GSK-3β contributes to the regulation of histone modifications involved in epigenetics. GSK-3β directly phosphorylates histone 1.5, histone 3,¹²⁹ HDAC3 ¹³⁰ and HDAC4.¹³¹ In this line, in hippocampal neurons, GSK-3β phosphorylates HDAC6 enhancing its activity.^{132, 133} GSK-3β also modulates post-translational modifications of histones by regulating several HATs.¹³⁴ Furthermore, combined GSK-3β and HDACs inhibition induces synergistic neuroprotective effects compared with single target inhibition, with potentially improved therapeutic selectivity.^{135, 136}

De Simone et al.¹³⁷ reported the first GSK-3β and HDAC dual inhibitor for AD treatment. Authors described a new hybrid family that combines hydroxamic acid, an HDAC inhibitor that chelates the Zn²⁺ ion located at HDACs active site, and the maleimide-like scaffold moiety, which is known to competitively bind to the GSK-3β ATP binding site. The 6-(3-(1,3-dioxoisoindolin-5-yl)thioureido-N-hydroxyhexanamide derivative 20 (Figure 6) showed IC₅₀ values of 12.78 and 3.19 µM toward HDAC1 and HDAC6, respectively, and it was able to inhibit GSK-3β with an IC₅₀ of 2.69 µM. Interestingly, compound 20 induced H3 acetylation and blocked Cu²⁺-induced tau hyperphosphorylation. Derivative 20 also showed neuroprotective properties toward OS-induced neuronal death in SH-SY5Y cells treated with H₂O₂ at 0.1 µM by restoring cell viability to almost control values and restored p53 levels to control levels at 10 µM. Considering its neuroprotective character, compound 20 completely reversed neurotoxicity induced by 6-hydroxydopamine (6-OHDA) in primary differentiated cerebellar granule neurons at 5 µM and promoted neurogenesis at 10 µM. Moreover, the immunomodulatory capacity of compound 20 was demonstrated in primary microglial cultures being able to reduce neuroinflammation by decreasing the expression of nitric oxidase synthase 2 (NOS2) and mannose Receptor C-type 1 (MRC1) in cells treated with LPS and compound 20 at 5 µM. Finally, compound 20 showed no toxicity in in vivo wildtype zebrafish model. Compound 20 also demonstrated effectiveness in vivo in a model of cyclin-dependent kinase-like 5 gene (CDKl5) deficiency,¹³⁸ being able to exert a neuroprotective effect following 15-day intraperitoneal treatment (50 mg/kg), evidenced by an increase in hippocampus neuronal survival, a restoration of neuronal dendritic spines number and mature state and a decrease in hippocampus microglia enlarged body size, related to a state of activation. Thus, although not evaluated in an AD model, compound 20 presents an adequate profile to exert a therapeutic effect in the brain in vivo.

2.1.7 GSK-3β and DYRK1A inhibition

The dual-specificity tyrosine phosphorylation-regulated kinase-1A (DYRK1A) is a protein kinase that belongs to the Dyrk family. DYRK1A can phosphorylate tau on several Ser and Thr residues. Indeed, it is responsible for tau phosphorylation at epitopes that prime tau for further phosphorylation by GSK-3β.¹³⁹ Additionally, DYRK1A activity has been related to an enhancement of the APP amyloidogenic processing via phosphorylation increasing Aβ production.¹⁴⁰ In good agreement, DYRK1A activity is increased in AD brains and its inhibition led to cognitive improvement in the APP/PS1/Tau 3xTg-AD mice model by reducing tau hyperphosphorylation and Aβ levels.¹⁴⁰ Thus, DYRK1A inhibition in combination with GSK-3β can reduce tau hyperphosphorylation more effectively, preventing NFT formation while simultaneously reducing Aβ pathology.¹⁴⁰

In this line, Liu et al.¹⁴¹ reported a new series of GSK-3β and DYRK1A inhibitors based on the β−carboline structure of the alkaloid harmine. This work led to the identification of compound ZDWX-25 (21, Figure 7) as a potent dual GSK-3β and DYRK1A inhibitor (IC₅₀ = 71 nM, IC₅₀ = 103 nM, respectively) with good BBB permeability (PAMPA assay). Additionally, ZDWX-25 treatment (1 µM) reduced tau and DYRK1A hyperphosphorylation (^Ptau-Ser³⁹⁶, ^PDYRK1A-Tyr^321/273) and NFT formation while inhibiting GSK-3β and DYRK1A in okadaic acid-injured SH-SY5Y cells. Interestingly, ZDWX-25 also showed a protective effect in vivo in the APP/PS1/Tau 3xTg-AD mice model.¹⁴² Treatment with ZDWX-25 for 2 months at 15 mg/kg reduced tau hyperphosphorylation and aggregation in the hippocampus and temporal cortex, reducing the cognitive impairment associated with the model.

In a drug repurposing strategy, Grygier et al.¹⁴³ reported that CX-4945 (22, Figure 7) (also known as silmitasertib), a clinically approved casein kinase 2 (CK2) inhibitor for cancer treatment,¹⁴⁴ is also a dual GSK-3β (IC₅₀ = 0.19 µM) and DYRK1A (IC₅₀ = 0.16 µM) inhibitor. CX-4945 reduced tau hyperphosphorylation exerting a protective effect in in vitro and in in vivo models based on DYRK1A overexpression.¹⁴⁵ Additionally, CX-4945 showed an anti-inflammatory effect in the astrocytoma cell line U373 cells and human primary astrocytes stimulated with IL-1β or TNF-α.¹⁴⁶ In IL-1β stimulated astrocytes, CX-4945 treatment (10 µM) reduced IL-6 and MCP-1 production (55% and 75% reduction, respectively). Similarly, in TNF-α-stimulated astrocytes, a reduction in IL-6 and MCP-1 (50% and 60%, respectively) levels was observed following the same treatment. Interestingly, these anti-inflammatory properties were not reproduced by the CK2 inhibitor 4,5,6,7-tetrabromo-1H-benzotriazole, suggesting that its GSK-3β and/or DYRK1A inhibitory capacity might be related to this activity.

Other compounds based on 3,6-diamino-1H-pyrazolo[3,4-b]pyridine structure (23, Figure 7)¹⁴⁷ or tetrahydropyridine isoindolone (valmerin) structure (24, Figure 7)^{148, 149} have been reported as multitarget GSK-3β, DYRK1A, and CDK5 inhibitors. Among valmerin derivatives, several compounds showed a strong inhibitory effect on the three kinases, with IC₅₀ values in the submicromolar range for GSK-3β and CDK5 and in the low micromolar range for DYRK1A. Among them, compound 24 exhibited most potent multitarget inhibition (IC₅₀ GSK-3β = 7 nM; IC₅₀ CDK5 = 35 nM; IC₅₀ DYRK1A = 2.2 µM). However, the potential therapeutic effect of these compounds in AD models has not been assessed yet. Indeed, a cytotoxic effect in the submicromolar range was found for several of the valmerin derivatives in a panel of 6 cell lines: Huh7 (liver), Caco2 (colon), MDA-MB231 (breast), HCT-116 (colon), PC3 (prostate), and NCIH727 (lung).

2.1.8 Dual GSK3β and AChE inhibitors

Additionally to previously described involvement of AChE and GSK-3β in AD pathology, both enzymes are cross-linked under pathological conditions. It was initially described an increased expression of AChE synaptic isoform during apoptosis activation.¹⁵⁰ Later it was demonstrated that GSK-3β mediates the increased expression of synaptic AChE during apoptosis.⁵⁸ As many cells undergo apoptosis due to the presence of senile plaques and NFTs of AD brains, the combination of AChE and GSK-3β as therapeutic targets for the development of MTDLs for AD treatment has been recently investigated.

Another example are thee tacrine-pyridothiazole derivatives described by Jiang et al.¹⁵¹ Their tacrine moiety was included to bind at the AChE CAS, and the pyridoxathiazole fragment, was inclued to interact at the GSK-3β ATP-binding site. Among these hybrids, the 2-(2-(cyclopropanecarboxamido)pyridine-4-yl)−4-methoxy-N-(3-((1,2,3,4-tetrahydroacridin-9-yl)amino)propyl)thiazole-5-carboxamide derivative (25, Figure 8) showed a nanomolar inhibition of both hAChE (IC₅₀ = 6.4 nM) and GSK-3β (IC₅₀ = 66 nM). Authors proposed that the pyridoxathiazole moiety is able to bind at the AChE PAS to generate a dual binding ligand with potential implications in Aβ aggregation. In this regard, compound 25 inhibited 46% Aβ self-aggregation at 20 µM. Moreover, compound 25 significantly reduced tau hyperphosphorylation in mouse neuroblastoma N2A cells. In vivo studies confirmed that compound 25 significantly ameliorated cognitive decline induced by scopolamine in ICR mice and showed reduced hepatotoxicity compared with tacrine.

In the same line, Jiang et al.¹⁵² have recently reported a novel hit compound (26, Figure 8) obtained by fusing the pharmacophores of pyridothiazole with benzylpiperidine scaffold. It showed IC₅₀ values of 0.3 μM for hAChE and 3 nM for hGSK-3β. Furthermore, it has a good iinhibitory activity of Aβ1-42 self-aggregation with a ~25% of reduction at 25 μM, which leads to anti-inflammatory and neuroprotective properties in an in vitro model of OS neuronal death. It also exhibited good BBB permeability (PAMPA assay).

In 2019, Oukoloff et al. designed a new series of tacrine-valmerin hybrids as potential AChE/GSK-3β inhibitors, including a 1,2,3-triazole moiety as linker.¹⁵³ As a result, hit compound 27 (Figure 8) showed the most promising in vitro activities, inhibiting both human AChE and GSK3β at the nanomolar range (9.5 and 7 nM, respectively). Moreover, in the neuroblastoma cell line SH-SY5Y it exhibited low cytotoxicity and a good ability to penetrate the BBB in tne MDCK-MDR1 cell line without interacting with efflux pumps such as P-gp.

Yao et al.¹⁵⁴ reported novel tacrine-pyrimidone hybrids as dual AChE/GSK-3β inhibitors for AD treatment. The 2-((7-((3-chloro-5,6,7,8-tetrahydroacridin-1-yl)amino)heptyl)amino)−6-(3-fluoropyridin-4-yl)−3-methylpyrimidin-4(3H)-one derivative (28, Figure 8) was selected as lead candidate showing high potency to inhibit both AChE and GSK-3β in the nanomolar range (IC₅₀ = 51.1 and 89.3 nM for AChE and GSK-3 β, respectively). Derivative 28 confirmed its ability to inhibit GSK-3β in vitro by reducing ^ptau levels at both Ser⁸⁶ and Ser³⁹⁶ sites at 0.5 and 5 µM, respectively, and protecting neuronal cells from cytotoxic glyceraldehyde in SH-SY5Y derived neuron models at the same concentrations. Moreover, derivative 28 was able to maintain the axon length of neurons in the presence of this toxic, preventing neurite shortening in differentiated SH-SY5Y cells. Behavioral studies in the AD scopolamine-related mouse model showed the ability of 28 to ameliorate the memory impairment caused by scopolamine, achieving the therapeutic effects of tacrine at the same administered dosage (15 mg/kg).

2.2.1 GSK-3β inhibition and calcium channel blockade

Based on these evidence, Bisi et al.¹⁶⁰ described the first dual VGCCs and GSK-3β modulators for AD treatment. Authors designed a new maleimide-anthracene-based derivative with potential therapeutic activity in AD. Among described derivatives, the N-heptyl-12,14-dioxo-9,10-[3,4]epipyrroloanthracene-9(10H)-carboxamide derivative (29, Figure 9) exhibited interesting GSK-3β inhibition capacity (IC₅₀ = 3.12 µM) and it was the most potent VGCC blocker reducing Ca²⁺ entry elicited by K⁺ 70 mM-evoked depolarization in SH-SY5Y cells with an IC₅₀ value of 9 µM. Following these results, the authors reported a hit-to-lead optimization program focused on reducing molecular weight by simplifying the complex polycyclic structure.¹⁶¹ However, although some of the new derivatives maintained a multitarget profile, they experienced a reduction in potency as exemplified by compound 30 (Figure 9), (GSK-3β IC₅₀ = 52.5 µM; Ca²⁺ entry blockade IC₅₀ = 12.0 µM).

Recently, Michalska et al.¹⁶² reported a novel series of 1,4-dihydropyridine derivatives as dual VGCC antagonists and GSK-3β inhibitors. Among these compounds, the 6-amino-4-(4-bromophenyl)-3-methyl-4,7-dihydro-2H-pyrazolo[3,4-b]pyridine-5-carbonitrile derivative (31, Figure 9) exhibited interesting GSK-3β inhibition capacity (IC₅₀ = 2.35 µM) and it was able to reduce the VGCC response to 62% at 10 µM. Derivative 31 also demonstrated an interesting antioxidant activity (ORAC = 0.96 Trolox equivalents, T.e.), potent anti-inflammatory capacity able to reduce nitrite production in BV2 microglial cells (EC₅₀ = 4.50 µM), good neuroprotective profile in in vitro models of tau hyperphosphorylation, OS, and [Ca²⁺]_c overload at 1 µM with survival percentages of 86.7%, 59.7%, and 69.4%, respectively. Interestingly, derivative 31 reduced cell death induced by tau hyperphosphorylation in acute treatment of hippocampal slices treated with okadaic acid by reducing ROS production.

2.2.2 Calcium channel blockade and H3R antagonism

Histamine is an endogenous amine that acts as a neurotransmitter in the CNS. The histaminergic system is composed of histamine and its receptors, playing an important role in maintaining brain homeostasis function.¹⁶³ There are four histamine receptors (H1R-H4R) involved in inflammation in the periphery, gastric acid secretion, dilatation of capillaries, and muscle contraction. H3R activation inhibits neurotransmitter release; thus, the development of H3R antagonists was proposed as a therapeutic strategy to increase, correct, and regulate histamine levels and other neurotransmitters, such as ACh, serotonin, dopamine, and norepinephrine.¹⁶⁴ Additionally, the H3R-mediated histaminergic autoinhibitory effect depends on Ca²⁺ release from intracellular stores, however, it is impaired in depolarized cells.¹⁶⁵ Depolarization of histaminergic neurons increases [Ca²⁺]_c through VGCCs.¹⁶⁶

Based on these considerations, Malek et al.¹⁶⁷ reported the design, synthesis, and pharmacological evaluation of novel diethyl 2,6-dimethyl-4-(4-(2-(amine-1-yl)alkoxy)phenyl-1,4-dihydropyridine-3,5-dicarboxilates as MTDLs for AD treatment. Its design considered the connection of the selective l-VGCC blocker 1,4-dihydropyridine (DHP) scaffold and the cycloalkylamine moiety present in the H3R antagonist Pitolisant. After pharmacological evaluation of all designed derivatives, the diethyl 2,6-dimethyl-4-(4-((5-(piperidin-1-yl)pentyl)oxy)phenyl)-1,4-dihydropyridine-3,5-dicarboxylate derivative 32 (Figure 9) was identified as a moderate VGCC blocker (IC₅₀ = 21 µM), with potent affinity toward H3R (K_i = 565 nM), selective hBuChE inhibition (IC₅₀ = 7.83 µM) and antioxidant properties (ORAC = 3.6 Te). The multitarget activity of compound 32 resulted in an interesting neuroprotective profile. 32 increased cell viability in the rotenone/oligomycin A-induced OS cell model at 1 and 3 µM (64.8% and 49.9% protection, respectively). It also protected cells in the okadaic acid-induced tau hyperphosphorylation cell model (41.8% and 20.9% protection, respectively). Furthermore, compound 32 restored cognitive impairment induced by LPS in vivo, demonstrating its potential as a multitarget drug for AD treatment.

2.2.3 Calcium channel blockade and AChE inhibition

Zhang et al.¹⁶⁸ described a new 1,4-dihydro-naphthyridine derivative SCR-1693 (33 × 52, Figure 18), ethyl 5-amino-4-(2-chlorophenyl)-2,7,7-trimethyl-1,4,6,7,8,9-hexahydrobenzo[b][1,8]-naphthyridine-3-carboxylate compound 33 × 52 as a selective, reversible, and noncompetitive AChE inhibitor (IC₅₀ = 0.68 μM) based on previously developed MTDLs reported.^{169, 170} It also inhibited in a dose-dependent manner KCl-evoked intracellular Ca²⁺ increase in SH-SY5Y cells (IC₅₀ = 28.6 μM) and possessed high efficiency in decreasing Aβ_25–35-induced SH-SY5Y cell death at a dose of 0.02–2.5 μM. Finally, in a mouse model of impairment in learning and memory induced by icv-Aβ2_5–35 injection, 33 ×52 at 1 mg/kg improved this impairment in several tests (Morris water maze and Y-maze).

A particularly relevant approach was undertaken by Arce et al.¹⁷¹ to develop a novel MTDL family based on donepezil scaffold and l-glutamic acid as linkers for different substituents. Lead compound 34 showed interesting and selective AChE inhibitory capacity with an IC₅₀ of 0.53 μM and the capacity to reduce OS. However, its VDCC blockade capacity was not initially designed and this activity was subsequently described. Interestingly, compound 34 blocked Ca²⁺ entry to the cytosol of chromaffin cells after 70 mM K⁺ stimuli with an IC₅₀ of 5.3 μM.¹⁷² Further evaluation demonstrated a specific l-VDCC blockade by this compound. To complement its multitarget profile, compound 34 elicits antioxidant effect and this combination resulted in a potent neuroprotective profile against OS with good BBB penetration. This profile prompted its in vivo evaluation in a cerebral stroke model in which compound 34 reduced the volume of the infarct by reducing OS and pro-inflammatory markers¹⁷³ (Figure 10).

In 2021, Ismaili et al.¹⁷⁴ reported 35 (Figure 18) as an MTDL with hAChE (IC₅₀ = 0.06 μM), hBuChE (IC₅₀ = 17.3 μM), MAO-B (IC₅₀ = 12.3 μM), GSK-3β (IC₅₀ = 24.3 μM), and tau (75.3% at 50 μM) and Aβ_1-42 aggregation (66% at 50 μM) inhibitory activities and as calcium channel antagonist (IC₅₀ = 1.4 μM). This compound, named ethyl 1-(2-(1-benzylpiperidin-4-yl)ethyl)-6-methyl-2-oxo-4-(4-oxo-4H-chromen-3-yl)-1,2,3,4-tetrahydropyrimidine-5-carboxylate, possesses a selected benzylpiperidine motif of donepezil and chromone and has also proved to be BBB permeable and effective in reducing scopolamine-induced cognitive deficits in mice (10 mg/kg intraperitoneally).

2.3.1 PDEs and AChE inhibition

Considering the positive results obtained with PDE inhibitors, novel MTDLs were designed based on tadalafil (36, Figure 11) (a PDE5A inhibitor approved for the treatment of pulmonary hypertension) aiming to integrate dual inhibitory activity toward AChE and PDE5.¹⁷⁸ The first-generation of AChE-PDE5 dual inhibitors were synthesized by replacing the substituent group at the N-atom of piperazine-2,5-dione with morpholine-4-ethyl, 1-benzylpyrrolidine-3-yl, N,N-dimethylaminoethyl, N-methyl-N-benzylaminoethyl, and substituted or unsubstituted benzylpiperidine moieties. These compounds showed improved BBB permeability and AChE inhibitory activities compared with tadalafil, maintaining their PDE5 inhibitory capacity. Representative derivative 37 (Figure 11) exhibited significant and selective inhibitory capacity against both targets (AChE IC₅₀ = 0.032 µM and PDE5A IC₅₀ = 1.53 µM). 37 citrate derivative relieve scopolamine-induced learning and memory deficits in in vivo, demonstrating that 37 treated animals exhibited shorter escape latency and less frequent errors than scopolamine group at a dosage of 10 mg/kg, being comparable with donepezil.

Unfortunately, 37 was nearly insoluble in water (water solubility of citrate = 250 µg/mL), which may limit its further development due to poor oral bioavailability. Herein, Ni et al.¹⁷⁹ developed alternative tadalafil derivatives to obtain more potent dual-target PDE5A/AChE inhibitors with improved water solubility. Among new compounds, (6 R,12aS)-2-(2-(1-benzylpiperidin-4-yl)ethyl)-6-(4-methoxyphenyl)-2,3,6,7,12,12a-hexahydropyrazino[1',2':1,6]pyrido[3,4-b]indole-1,4-dione (38, Figure 11) exhibited good selective dual-target AChE/PDE5 inhibition and good BBB permeability. Furthermore, its citrate form possessed improved water solubility to 420 µg/mL, and a good capacity to improve cognitive impairment induced by scopolamine in the in vivo model, inhibiting cortical AChE activities and enhancing CREB phosphorylation ex vivo.

Recent examples of MTDLs directed toward these targets are the novel pyrazolopyrimidone-donepezil hybrids reported by Hu et al.¹⁸⁰ Their design fused the pyrazolopyrimidinone subunit of previously reported PDE9A inhibitor C33 and benzyl piperidine subunit of donepezil using different linkers. Among them, compound 6-((2-(1-benzylpiperidin-4-yl)ethoxy)methyl)1-cyclopentyl-1,5-dihydro-4H-pyrazolo[2,4-d]pyrimidin-4-one (39, Figure 11) exhibited excellent and balanced dual-target AChE and PDE9A inhibitory activities (IC₅₀ = 0.285 and 0.223 µM, respectively), low neurotoxicity, and the ability to cross the BBB. The in vivo studies of 39 revealed that it ameliorates learning deficits induced by scopolamine and improves cognitive and spatial memory in the Aβ₂₅₋₃₅-induced cognitive deficit mice in the Morris water maze test.

Liu et al.¹⁸¹ study focused on the development of potential AD treatments by exploring dual inhibitors of PDE4 and AChE using dihydro-1H-inden-1-ones. Among the array of compounds evaluated, compound 40 (Figure 11), marked by the introduction of a 2-(piperidin-1-yl)ethoxy group at the 6-position of the indanone ring, demonstrated remarkable inhibitory activities and selectivity against both AChE and PDE4D. Compound 40 exhibited strong AChE inhibition with an IC₅₀ value of 0.28 μM and PDE4 inhibition (IC₅₀ = 1.88 μM). Furthermore, it displayed notable anti-neuroinflammatory properties, outperforming both donepezil and combination therapy (donepezil + rolipram). Compound 40 revealed low neurotoxicity and exhibited a significant neuroprotective effect against Aβ_25-35-induced cell death.

2.3.2 PDEs and HDAC inhibition

Simultaneous PDEs and HDACs inhibition has recently been validated as a potentially novel therapeutic approach for AD treatment.¹⁸² Rabal et al.¹⁸³ reported the first dual PDE5 and HDAC inhibitors designed as a combination of sildenafil, a PDE5 inhibitor, and the pharmacophoric features of HDAC inhibitors. PDE5 is distributed in the hippocampus, cortex, and cerebellum and is upregulated in the brains of AD patients compared with age-matched healthy control subjects.¹⁸⁴ Consequently, cGMP levels are significantly decreased in the cerebrospinal fluid of AD patients.¹⁸⁵ Thus, PDE5 inhibition could increase cGMP levels,¹⁸⁶ and GSK-3β inhibition could decrease phosphorylated tau levels,^{186, 187} enhancing synaptic plasticity and cognitive function, as previously demonstrated in different AD animal models.¹⁸⁸ After a deep lead optimization strategy, authors selected the 3-[[4-ethoxy-3-(1-methyl-7-oxo-3-propyl-6H-pyrazolo[4,3-d]-pyrimidin-5-yl)phenyl]methyl]cyclobutenecarbhydroxamic acid 41 (Figure 12) as lead compound. Derivative 41 showed a highly potent and balanced PDE5 (IC₅₀ = 60 nM) and HDAC6 (IC₅₀ = 91 nM) inhibition demonstrating, also, good selectivity as it showed lower inhibitory activity against class I HDACs (HDAC1, HDAC2, and HDAC3, IC₅₀ = 310, 490, and 322 nM, respectively). Selectivity toward HDAC6 is associated with a safer profile as strong inhibition of class I HDACs is associated with toxic effects. Compound 41 showed low cytotoxicity in the THLE-2 hepatic cell line and in primary neuronal cultures (LC₅₀ = 7.2 µM and LC₅₀ = 17.7 µM, respectively) and an improved ability to cross the BBB. This multitarget profile prompted an in vivo evaluation in the Tg2576 AD model. Chronic treatment of Tg2576 mice during 4 weeks with compound 23 (40 mg/kg, intraperitoneal, daily) decreased Aβ₄₂ levels, APP, and hyperphosphorylated tau to almost 50% compared with vehicle-treated animals. More importantly, it elicited a positive effect on spatial memory and reversed cognitive impairment. Tau hyperphosphorylation reduction was related to a significant increase of GSK-3β inactive form levels. Additionally, compound 41 could restore memory through the cGMP/CREB pathway activation that plays an important role in synaptic plasticity.¹⁸⁹ Additionally, it rescued the impaired long-term potentiation evident in hippocampal slices from APP/PS1 mice.

To improve the biochemical and ADMET properties of compound 41, a new series of derivatives were designed by replacing the sildenafil moiety with different PDE5 inhibitors, such as vardenafil and tadalafil,¹⁹⁰ or by including different HDAC6-selective inhibitors, such as tubastatin A and nexturastat A.¹⁹¹ Vardenafil-derivative 42 (Figure 12) displayed a potent HDAC6 affinity (IC₅₀ = 110 nM), moderate class I HDACs inhibitory profile (IC₅₀ against HDAC1 and HDAC2 of 286 and 1260 nM, respectively), and good PDE5 inhibitory activity (IC₅₀ = 5 nM).¹⁹⁰ On the other hand, the sildenafil-based analog 43 (Figure 12) also showed a potent HDAC6 affinity (IC₅₀ = 15 nM), moderate class I HDACs inhibitory profile (IC₅₀ against HDAC1, HDAC3, and HDAC2 of 345, 2090, and 3720 nM, respectively), good PDE5 inhibitory activity (IC₅₀ = 11 nM), and reduced the levels of the AD-related markers hAPP and tau hyperphosphorylation in Tg2576 neurons.¹⁹¹

Continuing with the optimization process, chemical series of compound 41 were developed and evaluated in in vivo models of AD.¹⁹² The sildenafil- derivative 44 (Figure 12) showed potent PDE5A inhibition (IC₅₀ = 114 nM) and the highest HDAC1 and HDAC2 inhibition in the mid-nanomolar range (IC₅₀ = 180 and 668 nM, respectively). Compound 44 also showed an adequate ADMET profile and in vivo target engagement (histone acetylation and cAMP/cGMP response element-binding phosphorylation) in the CNS. Thus, 44 was assayed in a mouse model of AD (Tg2576); however, it failed to produce significant improvement in memory after 2 weeks of treatment (20 mg/kg).

Recently, a novel series of dual PDE9 and HDAC inhibitors have been reported.¹⁹³ These compounds were designed using selective PDE9 inhibitors (PF-04447943 and PF-04449613) as well as privileged chemical structures, bearing zinc binding groups (hydroxamic acids and ortho-amino anilides) to target HDAC proteins. The PF-04447943-based hydroxamic acid analog 45 (Figure 12), bearing an N¹-cyclopentyl group, was the most potent HDAC6 inhibitor, (IC₅₀ = 40 nM), and it exhibited potent PDE9 inhibitory capacity (IC₅₀ = 107 nM). Moreover, it exhibited a good in vitro ADMET profile and BBB permeability to advance through in vivo efficacy models. Chronic treatment of aged Tg2576 mice with 45 reversed the AD phenotype in an effective manner.¹⁹⁴ In fact, Tg2576 mice treatment with 45 during 4 weeks (40 mg/kg, intraperitoneally) promoted a significantly higher response in the fear conditioning test (60.9% freezing) compared with vehicle-treated animals (36.9%). In the Morris water maze test, the latency to find the platform was prolonged in Tg2576 mice relative to the WT mice, whereas the escape latency was shorter when these Tg2576 mice received 45, although globally no significant differences were found between groups. Aβ₄₂ levels in parieto-temporal cortical extracts showed a significant decrease in Tg2576 mice treated with 45 (7.70 pg Aβ/µg of protein) relative to those that received the vehicle alone (17.8 pg Aβ/µg of protein). These results showed that 45 ameliorates memory impairment and diminished AD pathological markers although its effect was lost after a wash-out period of 4 weeks.

2.3.3 Metal ions chelation and PDE inhibition

Unbalanced metal ion levels are a well-established pathological AD feature.^195-197 Thus, metal−Aβ interactions modulation and metal distribution in the brain could represent an interesting therapeutic strategy to both inhibit Aβ aggregation and reduce the OS status.¹⁹⁸ Aβ can complex transition metal ions such as Fe³⁺ and Cu²⁺ through histidine residues.¹⁹⁹ Binding to these metals favors peptide aggregation and its pro-oxidant effects, generating ROS via the Fenton reaction.⁵⁵ This interaction causes the reduction of the ions from Fe³⁺ to Fe²⁺ producing H₂O₂ as a by-product. Then, the Fenton reaction takes place where Fe²⁺ reduces H₂O₂ recovering Fe³⁺ and releasing OH^- and OH^·, highly reactive species with a very short half-life.²⁰⁰ Abnormally high concentrations of both metals Fe³⁺ and Cu²⁺ are accumulated in Aβ plaques.²⁰¹ Additionally, it has been suggested that OS promotes the amyloidogenic pathway of APP processing, thus, initiating a positive feedback loop in which oxidative damage promotes Aβ misfolding and misfolded Aβ promotes oxidative damage.²⁰²

Su et al.²⁰³ reported a new series of multifunctional agents that combine PF-04447943 (PDE9 inhibitor) and clioquinol as bio-metal chelator. Among them, derivative 46 (Figure 13) exhibited an excellent and selective inhibitory activity toward PDE9 (IC₅₀ = 34 nM), the ability to chelate bio-metals at 20 µM and the subsequent formation of 46-metal ion (II) complex, together with antioxidant properties (ORAC = 4.47 Te). Compound 46 demonstrated a good capacity to inhibit Cu²⁺-induced Aβ_1-42 aggregation (64.7% reduction) and disaggregation of Cu²⁺-induced Aβ_1-42 fibrils (64.6% reduction at 50 µM).

In this line, Wang et al.²⁰⁴ used clioquinol and moracin M as scaffolds to design new multi-target hybrids for AD. Clioquinol moiety was included aiming to chelate bio-metals whereas moracin M-related group was included to integrate PDE4D inhibitory capacity. Compound WBQ5187 (47, Figure 13) was selected as lead compound showing good PDE4D inhibitory potency (IC₅₀ = 0.32 µM), appropriate bio-metal chelating capacity, which demonstrated the interaction of 47 with Cu²⁺, Zn²⁺, Fe²⁺, and Fe³⁺ at 50 µM, antioxidant capacity (ORAC = 3.6 Te) and potent anti-inflammatory properties (EC₅₀ = 1.50 µM to reduce NO production in LPS-stimulated BV2 cells). Considering the interconnection of metal ions and Aβ, compound 47 inhibited self-induced Aβ_1-42 aggregation (67.5% and 86.3% inhibition at 5 and 25 µM, respectively), disassembled self-induced Aβ aggregates (83.9% disaggregation of Aβ_1-42 fibrils at 50 µM), and modulated Cu²⁺-induced Aβ aggregation (80.4% inhibition at 50 µM) being more efficient than clioquinol and moracin M. Interestingly, compound 47 showed significant interactions with Aβ_1-42 peptide through Arg⁵, His⁶, His¹⁴, Gln¹⁵, and Phe²⁰ residues as demonstrated by NMR. Finally, compound 47 showed a potent neuroprotective profile in vivo; it protected hippocampal neurons from necrosis and demonstrated improvements in cognitive abilities and conventional reference spatial memory compared with those animals treated with donepezil and clioquinol, or untreated rats.

2.4.1 Metal ions chelation and MAO inhibition

Mono-amine oxidase enzymes (MAO-A and MAO-B) catalyze oxidative deamination of mono-amines producing H₂O₂ and ROS as secondary products, thus, increasing oxidative injury and promoting the toxic environment characteristic of neurodegeneration.^{205, 206} Monoamine neurotransmitters regulated by MAOs, such as dopamine, norepinephrine, and serotonin (5-HT), are crucial for memory and cognition.²⁰⁷ MAO inhibitors were proposed to improve monoaminergic neurotransmission and decrease ROS formation and OS, thus exerting antioxidant and neuroprotective properties and, consequently, to induce cognitive improvement.²⁰⁸ A prominent example is rasagiline (48, Figure 13), a potent and selective MAO-B irreversible inhibitor approved in 2005 for Parkinson's disease (PD) treatment, which has entered Phase II clinical trials for AD treatment.²⁰⁹

Based on these premises, Mi et al.²¹⁰ developed a novel class of multitarget hybrids by linking coumarin (MAO-B inhibitor) and hydroxypyridinone (Fe³⁺ chelator) using a cleaver approach based on click chemistry. Obtained derivatives displayed excellent Fe³⁺ chelating activity and moderate to good MAO-B inhibitory activity. Compound 49 (Figure 14) exhibited the most potent MAO-B inhibitory capacity with an IC₅₀ value of 0.68 µM and a Fe³⁺ affinity of pFe³⁺ = 19.8.

Using a similar approach, Xie et al.²¹¹ reported a family of selegiline-clioquinol hybrids. Selegiline (MAO-B inhibitor) was also approved for PD treatment.²¹² The (R)-2-(((4-(2-(methyl(prop-2-yn-1-yl)amino)propyl)phenyl)amino)methyl)phenol (50, Figure 14) was selected as the lead compound being a potent MAO-B inhibitor (IC₅₀ = 0.21 µM) and effective bio-metal chelator of Cu²⁺, Fe²⁺, and Zn²⁺. Compound 50 final structure also integrates good antioxidant activity (ORAC = 4.20 Te) and the ability to inhibit Cu²⁺-induced Aβ_1-42 aggregation (69.1% at 50 µM).

Complementing this family, benzylideneindanone derivatives showed multifunctional activity as potential anti-AD drugs due to their ability to inhibit MAO-B activity, to block self-induced Aβ aggregation, their antioxidant, and bio-metal chelator capacity.²¹³ Compound 51 (Figure 14) had the greatest ability to inhibit Aβ_1-42 aggregation (80.1% at 20 µM) and to inhibit MAO-B (IC₅₀ = 7.5 µM). Moreover, it chelated bio-metals such as Cu²⁺ and Fe²⁺, with a stoichiometry of the 51-Cu²⁺ complex of 1:1, inhibit Cu²⁺-induced Aβ_1-42 aggregation noticeably through chelating Cu²⁺, and disassembling the well-structured Aβ fibrils at 50 µM.

2.4.2 Metal ions chelation, Aβ aggregation inhibition, and H3R antagonism

The AD pathological network is highly complex and the crosslinks among its components lead to the proposal of novel trifunctional multitarget hybrids by Sheng et al.²¹⁴ Their rational design included an aminopropoxyphenyl moiety as H3R antagonist and the 3-hydroxy-4-pyridinone subunit due to its chelating properties with high affinity for Cu²⁺, Zn²⁺ and Fe²⁺,²¹⁵ as well as its potential radical scavenging. The new 1-phenyl-3-hydroxy-4-pyridinone derivatives share similar structural features with SFK-64346, a well-known Aβ aggregation inhibitor that contains an aminopropyloxy-phenyl-nezofuran moiety. Among them, compound 3-hydroxy-2-methyl-1-(4-(3-(pyrrolidine-1-yl)propoxy)phenyl)-pyridin-4(1H)-one (52, Figure 15) showed excellent selective H3R antagonism (IC₅₀ = 0.32 nM), free radical scavenging capacity (ORAC = 1.54 T.e.), potent metal chelating activities of Cu²⁺ and Fe²⁺, with a stoichiometry of the 52-Cu²⁺ complex of 2:1, confirming that the 3-hydroxy-4-pyridinone moiety is essential for the metal chelating activity. 52 demonstrated self-induced Aβ₁₋₄₂ anti-aggregation capacity (80.1% inhibition at 20 µM and IC₅₀ value for Aβ₁₋₄₂ aggregation of 2.85 µM) and Cu²⁺-induced Aβ₁₋₄₂ anti-aggregation ability (57.9% disaggregation at 50 µM).

2.4.3 Metal ions chelation and tau aggregation inhibition

Kim et al.²¹⁶ reported the effects of metal ions on Aβ and tau aggregation. Zn²⁺, Cu²⁺, Fe³⁺, Mg²⁺, Mn²⁺, Pb²⁺, Cd²⁺, Hg²⁺, and Al³⁺ promote tau hyperphosphorylation and induce tau aggregation, whereas Fe²⁺ and Li⁺ reduce tau hyperphosphorylation.

Focusing on the tau hypothesis, Silva et al.²¹⁷ reported the evaluation of natural polyphenols and synthetic derivatives as potential tau aggregation inhibitors with metal chelating properties. The structure–activity relationship (SAR) study demonstrated that the nitrocatechol scaffold is required for a significant tau anti-aggregating activity. α-cyanocarboxamide derivative 53 (Figure 15) was one of the most potent inhibitors of tau aggregation, measured as an aggregation of the tau-derived peptide AcPHF6 (90.5% inhibition at 50 µM). Its potent anti-aggregating activity was explained by molecular docking studies, suggesting that the cyano group formed two hydrogen bond interactions with NH₃⁺ of two Lys side chains (distance <2.8 Å) and the amide carbonyl was involved in a hydrogen bonding interaction with Lys (distance = 1.85 Å). The N-benzyl ring established a π-cation interaction with NH₃⁺ of Lys (distance = 4.82 Å). These additional interactions support the increased tau-aggregation inhibitory activity of 53. The presence of a less flexible amide bond increases conformational stability in the steric zipper β-sheet model of tau hexapeptide ³⁰⁶VQIVYK³¹¹. Compound 53 also exhibited significant Cu²⁺ chelation capacity.

2.5.1 AChE and BACE-1 inhibition

As mentioned above, under pathological conditions, BACE-1 cleaves APP, followed by the γ-secretase complex, causing the generation of Aβ peptides that tend to oligomerize and accumulate as Aβ deposits, contributing to Aβ fibril formation. Aβ assembly progresses oligomers and senile plaques, causing neurotoxicity and inducing neuronal cell death.¹⁴ Furthermore, AChE is able to catalyze these conformational changes accelerating the aggregation process. Both enzymes, BACE-1 and AChE are linked by the production and aggregation process. Thus, the development of new molecules targeting both enzymes represents an interesting approach to designing new agents for AD treatment.

Zha et al.²²⁰ reported a new family of benzofuran hybrids to obtain activity against these selected targets. The N¹-(benzofuran-2-ylmethyl)-N⁷-(1,2,3,4-tetrahydroacridin-9-yl)-heptane-1,7-diamine derivative (54, Figure 16) revealed a nanomolar potency to inhibit cholinesterase activity (IC₅₀ = 0.86 nM and 2.18 nM toward hAChE and hBuChE, respectively) and micromolar capacity to inhibit BACE-1 (IC₅₀ = 1.35 µM). Moreover, it inhibited self- and AChE-induced Aβ aggregation and presented lower hepatotoxicity than tacrine. In vivo studies demonstrated better behavioral test performance and cognitive improvement in scopolamine-treated ICR mice compared with vehicle-treated animals.

Compelling research evidence has demonstrated that donepezil analogs have a double bond on the indanone moiety that can act as BACE-1 inhibitors due to their structural rigidity.²²¹ Therefore, numerous donepezil analogs were synthesized by alkylation or benzylation of the indanone central core,^{221, 222} or by the connection of this scaffold with other subunits, such as phthalimide or pyrimidine.^{223, 224} Regarding the structural modifications, compounds reported by Green et al.²²² were able to inhibit AChE and BuChE. Interestingly, the p-bromobenzyl analog (55, Figure 16) also inhibited BACE-1 with an IC₅₀ value of 3.4 µM.

Costanzo et al.²²¹ obtained donepezil-like compounds using a greener version of the synthetic procedure previously reported for methoxy-substituted donepezil precursors by aldol condensation. The bis-methoxy-derivative (56, Figure 16) showed selective and potent AChE inhibitory capacity at the nanomolar range (IC₅₀ AChE  58 nM and IC₅₀ BuChE = 4.74 µM) and inhibited BACE-1 with an IC₅₀value lower than 1 µM, IC₅₀ = 0.697 µM).

In this line, donepezil analogs developed by Gabr et al.²²⁵ represented by lead compound 57 (Figure 16) showed potent AChE and BACE-1 inhibitory capacity (AChE IC₅₀ = 4.11 nM and BACE-1 IC₅₀ = 18.3 nM), metal chelating properties, low toxicity on SH-SY5Y neuroblastoma cell line, and good BBB permeability.

The connection of donepezil pharmacophore benzylpiperidine with the pyrimidine-2,4-diamine scaffold yielded a new MTDL core possessing multifunctional activity against cholinesterase enzymes, BACE-1 and Aβ aggregation inhibitory capacity.²²⁴ Biological assays revealed that the nature of the substituent on the C4 benzylamine group affected the biological profile. The N²-(1-benzylpiperidin-4-yl)-N⁴-(3,4-dimethoxybenzyl)pyrimidine-2,4-diamine candidate 58 (Figure 16) inhibited AChE (IC₅₀ = 9.90 µM), BuChE (IC₅₀ = 11.40 µM), and BACE-1 (34% inhibition at 10 µM). Target combination led to the ability to inhibit self- and AChE-induced Aβ aggregation (17.4% and 59.3% at 100 µM, respectively). It also demonstrated low toxicity in the SH-SY5Y neuroblastoma cell line (81% of cell viability at 40 µM).

Belluti et al.²²⁶ used the fluorinated benzophenone scaffold as a central core of a new small library of 3-fluoro-4-hydroxy analogs for AD. The benzophenone is a privileged scaffold, a versatile pharmacophore nucleus, which could be structurally modified to provide analogs to reach different therapeutic targets.²²⁷ The in vitro screening allowed the authors to identify fluoro-4-hydroxyphenyl-4-[4-(2-hydroxyethyl)piperidin-1-ylmethyl]phenylmethanone (59, Figure 16) as the lead candidate with dual AChE (IC₅₀ = 2.52 µM) and BACE-1 (IC₅₀ = 2.32 µM) inhibitory profile along with antioxidant properties (compound 59 was able to inhibit 30% of ROS formation elicited by the organic peroxide t-BuOOH at 100 µM and Aβ_25-35 peptides at 5 µM in the OS cell model at concentrations of 3 and 5 µM of 59). The presence of the hydroxy group at the aromatic ring could explain the ability of the compound to counteract ROS formation at the neuronal level.

Viayna et al.²²⁸ reported the synthesis of rhein-huprine hybrids as disease-modifying AD potential drugs. The in vitro screening of this new family of compounds showed potent inhibitory activities against hAChE, hBuChE, and BACE-1, along with the ability to act as dual Aβ₄₂ and tau anti-aggregating agents. Furthermore, they showed good BBB permeability. (60, Figure 16), selected as a lead candidate, (IC₅₀ values of 2.39, 513, and 80 nM toward hAChE, hBuChE, and BACE-1, respectively), revealed an interesting protective profile against the Aβ-induced synaptic dysfunction, preventing the loss of synaptic proteins. Consequently, compound 60 demonstrated a positive effect to improve the long-term potentiation in an ex vivo study using brain slices of C57bl6 mice. In vivo studies in APP-PS1 transgenic mice treated by intraperitoneal injection for 4 weeks with compound 60 (2 mg/kg) reduced soluble Aβ and increased mature APP levels.

Kumar-Waiker et al.²²⁹ have recently reported a series of N-benzylpiperidine hybrids of 5-phenyl-1,3,4-oxadiazol-2-thiol, being 61 (Figure 16) the lead compound showing high hAChE (IC₅₀ = 0.076 μM), hBuChE (IC₅₀ = 1.204 μM), and hBACE-1 (IC₅₀ = 0.230 μM) inhibitory capacity. Moreover, it showed a significant Aβ aggregates reduction. Its interest prompted its in vivo evaluation in which it exhibited significant learning and memory behavior restoration in the scopolamine- and Aβ-induced mouse AD models, respectively. In these in vivo models, authors analyzed different parameters in ex vivo hippocampal slices in which compound 61 showed reduced levels of AChE, malondialdehyde, and NO, and decreased levels of TNF-α and IL-6 mRNA, among other properties.

Fares et al.²³⁰ reported new tacrine-chromene derivatives with a MTDL profile. Derivative 62 (Figure 16) showed hAChE (IC₅₀ = 0.44 μM), hBuChE (IC₅₀ = 0.08 μM), BACE-1 (IC₅₀ = 0.38 μM), and MAO-B (IC₅₀ = 5.15 μM) inhibitory capacity, a tetra-model combination not reported before. 62 showed low cytotoxicity in SH-SY5Y and HepG2 cell lines and good predicted BBB permeability. Besides, they had the ability to bind to PAS, which allowed them to decrease Aβ aggregation (59% at 10 μM), antioxidant activity, and metal-chelating properties toward Fe²⁺, Zn²⁺, and Cu²⁺.

Finally, Digiacomo et al.²³¹ described a novel series of tacrine derivates by combining caffeic, ferulic, and lipoic acids with tacrine, which exhibited relevant antioxidant and both AChE and BACE1 inhibitor capacities. Among these compounds, 63 showed an IC₅₀ of 0.04 μM and 1.03 μM for AChE and BACE1, respectively. Furthermore, it also had BACE1 inhibitory activity (47% at 10 μM) and displayed an appreciable neuroprotective effect at 3 μM against glutamate-induced neuronal death in HT22 neuronal cells.

2.5.2 AChE inhibition and serotonin receptor antagonism

The serotonergic system is implicated with the cholinergic system in the cognitive function of neurodegenerative diseases.²³² Serotonin (5-HT) receptors 5-HT_1A, 5-HT₄, 5-HT₆, and 5-HT₇ are associated with learning and memory. Serotonergic neurons also accumulate NFTs and a decreased number of these neurons are reported in AD brains.²³³ In addition, Aβ deposits in the projection sites of these neurons lead to axonal degeneration and, finally, to neuronal loss.²³¹ Moreover, the 5-HT4 and 5-HT6 receptors can modulate the activity of several secretases, thereby affecting APP processing.²³⁴ Thus, partial 5-HT₄ receptor agonists such as donecopride could be used to treat the cognitive symptoms of AD acting as a functional agonist when low receptor numbers are present, minimizing the potential receptor desensitization.^{235, 236} The 5-HT₆ receptor is expressed primarily in the brain cortex and the hippocampus, and it modulates primarily γ-aminobutyric acid (GABA) and glutamate levels, facilitating the secondary release of other neurotransmitters like ACh.²³⁷ Thus, 5-HT₆ receptor antagonists would be beneficial to ameliorate AD symptoms. 5-HT₆ receptor and cholinesterase have been combined as targets of novel MTDLs recently reported.^{236, 238, 239} Representative examples of these series of compounds are depicted in Figure 17.

In 2020, Rodríguez-Lavado et al.²⁴⁰ reported a new series of indole-benzamide derivatives as MTDL for the serotonin transporter (SERT) and AChE, due to their affinity on the serotonin systems.²⁴¹ Novel derivatives showed IC₅₀ values in the micromolar range toward human AChE (hAChE) (from 3.4 µM to higher than 50 µM). Interestingly, the 5-fluorine substitution at the indole moiety resulted in nanomolar affinity in SERT, derivatives 64 and 65 (Figure 17) the most potent with IC₅₀ values of 9.2 nM and 1.9 nM, respectively. Additionally, the 1-(4-(3-(1H-indol-3-yl)propyl)piperazin-1-yl)-2-(4-(3-bromobenzoyl)piperazin-1-yl)ethan-1-one derivative (66, Figure 17) exhibited significant anti-Aβ activity, inhibiting 50% of Aβ₄₀ levels in N2A mouse neuroblastoma cells at 10 µM.

Li et al.²⁴² in 2021 reported compound 67 (Figure 17) as a hybrid derivative of the antidepressant vilazodone (5-HT1A receptor partial agonist and serotonin transporter inhibitor) and donepezil aimed to treat comorbidity of AD and depression. This compound exhibited strong triple-target bioactivities in vitro of AChE, 5-HT_1A, and SERT with IC₅₀ values of 2.29, 58.6, and 29.22 nM, respectively. Compound 67 also showed acceptable brain distribution, ameliorating depressive symptoms and cognitive dysfunction (Y maze assay and tail suspension test) in the scopolamine mouse model.

Wichur et al.²⁴³ have also recently reported two novel compounds (68 and 69, Figure 17) with both 5-HT6R antagonism (K_i = 13 nM and K_i = 15 nM respectively) and cholinesterase inhibition. Compound 68 is a tacrine-based ligand with balanced activities toward eeAChE (IC₅₀ = 8 nM) and eqBuChE (IC₅₀ = 24 nM). Interestingly, it reduced Aβ aggregation by 75% at 10 µM. At the same concentration, it did not show toxicity on HepG2 cells; however, it showed low BBB permeability according to the PAMPA assay. On the other hand, 69 contains rivastigmine-derived phenyl N-ethyl-N-methylcarbamate fragment, being able to inhibit eqBuChE (IC₅₀ = 455 nM) and both tau and Aβ aggregation properties (79% and 68% at 10 µM, respectively). Furthermore, 69 did not influence CYP isoenzymes 3A4 and 2D6 activity at those concentrations and it is BBB membrane-permeable in PAMPA assay.

Finally, Asproni et al.²⁴⁴ reported a novel compound (70, Figure 17) that featured a tricyclic scaffold, namely thieno[3,2-h]cinnolinone, thienocyclopentapyridazinone and thienocycloheptapyridazinone, connected through alkyl chains with hAChE inhibitory capacity (IC₅₀ = 6.89 µM) and serotonin 5-HT6 receptor subtype interaction (K_i = 8.4 nM).

2.5.3 AChE inhibition and NRF2 induction

Exacerbated OS is considered one of the main pathological pathways related to NDD development including AD. OS is closely connected to the apparition of senile plaques, as previously discussed, besides, aberrant proteins increase the formation of ROS accelerating the advance of the disease. The NRF2/EpRE transcriptional pathway is considered the most important intrinsic cell defense mechanism under neuropathological conditions. It reduces OS and inflammation by promoting the expression of numerous antioxidant and anti-inflammatory enzymes.²⁴⁵

Based on these considerations, Benchekroun et al.²⁴⁶ designed a trifunctional hybrid based on ferulic acid, a hydroxycinnamic acid with antioxidant and neuroprotective properties, melatonin, an endogenous neurohormone with potent antioxidant and neuroprotective capacity, and tacrine as AChE inhibitor. The novel ferulic acid-tacrine-melatonin hybrid 71 (Figure 18) was selected as the lead compound with the best pharmacological profile. Compound 71 inhibits cholinesterase activity with IC₅₀ values of 1.29 µM toward hAChE and 234 nM toward hBuChE; it significantly activates the NRF2 transcriptional pathway in AREc32 cells with a CD value of 6.31 µM. This combination of targets resulted in an interesting neuroprotective profile against Aβ_1-40 and Aβ_1-42 toxicity, H₂O₂, and rotenone/oligomycin A-induced OS cell models (70.6%, 57.4%, 37.6%, and 33.7% protection, respectively, at 1 µM). To complete its evaluation, 71 showed a potent antioxidant effect in the ORAC test (ORAC = 9.11 Te), BBB permeability, and low toxicity in HepG2 cells.

Fusing the benzylpiperidine motif from the AChE inhibitor donepezil and the 1,2,4-oxadiazole core a from an NRF2 activator previously reported,²⁴⁷ Wang et al.²⁴⁸ have recently reported the hybrid compound 72 (Figure 18) with great AChE inhibitory (eeAChE IC₅₀ = 0.07 μM; hAChE IC₅₀ = 0.38 μM) and a significant NRF2 induction, since it upregulates HO-1, NQO1, and GCLm (downstream proteins) and favors its translocation to the nuclei. 72 also exhibited neuroprotective functions in vitro against the H₂O₂ insult (above 50% of cell survival at 10 μM) and it also reduced Aβ_1-42 aggregation. Interestingly, it also reduced ROS levels and pro-inflammatory cytokines liberated after LPS insult in BV2 cells. In the scopolamine AD mouse model, this compound was also effective (10 mg/kg) in increasing the arrival times and simplifying the tracks in the Morris water maze, improving mice cognition and alleviating inflammation via oral administration (10 mg/kg).

An interesting approach was reported by Cores et al.²⁴⁹ describing a new class of MTDL based on bisaventhramide, a natural product with excellent antioxidant capacity. From all derivatives, compound 73 (Figure 18) showed AChE inhibitory capacity (IC₅₀ = 12 μM) and Nrf2 induction activity (CD = 36 μM) being able to activate the phase II antioxidant response-related enzyme expression. Moreover, it showed a lack of cytotoxicity in neuronal and hepatic cellular and antioxidant capacity. The target combination included in derivatives 73 resulted in an interesting neuroprotective profile in two different AD models, OS, and tau hyperphosphorylation. Although further in vivo evaluation has not been performed yet, its profile positions compound 73 as a candidate for future lead optimization and in vivo evaluation.

In this line, a new class of MTDL hybrids was reported as curcumin-piperlongumine derivatives. This design was proposed to combine the NRF2 induction capacity of curcumin and the antioxidant and anti-inflammatory properties of the natural alkaloid peperlongumine. Compound 74 (Figure 18) was selected as the lead candidate due to its balanced combination of activities, although in this case the AChE activity was not evaluated. Compound 74 structural modifications compared with 73 resulted in an increased NRF2 induction capacity (CD = 4.9 μM) and improved antioxidant capacity (ORAC = 1.6 T. e. and DPPH IC₅₀ = 14.2 μM). Moreover, compound 74 showed low toxicity in cellular models and the ability to Tau protein aggregation in the PHF6 aggregation model. Regarding its anti-inflammatory properties, it showed a concentration-dependent reduction of nitrite production induced by LPS in the BV2 microglial cell line showing an IC₅₀ of 550 nM. It also showed potent neuroprotective activity in the OS and tau hyperphosphorylation toxicity models of AD complementing its pharmacological profile.

Guo et al.²⁵⁰ described compound 75 as a dimethyl fumarate-tranilast modified dithiocarbamate hybrid with improved AChE inhibitory activity (IC₅₀ = 0.053 μM) and potent Nrf2 activation at 7 μM confirmed with Western blot. The total Nrf2 amount was upregulated in a time-dependent manner, being at 4 h the maximum level of nuclear Nrf2 in BV2-2 microglia cells. This property also led to an increase in Nrf2-dependent antioxidant enzymes, including HO-1, NQO1, and GPX4, that showed the biggest upregulation observed at 10 μM. Moreover, compound 75 showed an interesting neuroprotective profile against H₂O₂-induced damage. In the LPS inflammatory model, 75 decreased ROS accumulation and NO production, and this activity was associated with a reduction in iNOS and COX-2 levels. It also significantly blocked TNF-α and IL-6 production. Besides, 75 was also predicted to cross the BBB in the PAMPA assay (Pe 17.95 × 10⁻⁶ cm/s). Finally, 75 was tested (30 mg/kg) in the scopolamine mice model, which reversed the step-down latency and number of errors.

Recently, Herrera-Arozamena et al.²⁵¹ described a novel series of indole- and naphthalene-oxadiazolones, among which N-propargyl-oxadiazolone 76 outstood as a hit compound. It displayed remarkable selectivity to quinone reductase 2 (QR2) ligand (K_i = 3.2 nM), a potent NRF2 induction in the AREc32 cell line (CD = 1.76 μM) and the ability to promote cell differentiation into neurons, as it greatly increased the expression of both TuJ-1 and MAP-2 (two well-known neuronal markers). It is also predicted to be BBB permeable as demonstrated by the PAMPA assay and had antioxidant properties in ORAC assay (1.8 T.e.)

2.5.4 AChE and MAO inhibition

The two isoforms of MAO (MAO-A and MAO-B) produce peroxidases, which cause OS and the depletion of neurotransmitters; inhibiting these enzymes allows for the accumulation of depleted neurotransmitters and decreases the formation of oxidative-free radicals, which eventually confers neuroprotection.²⁵²

One of the first examples of this MTDL combination, Ladostigil (77, Figure 19), was designed considering the connection of MAO enzymes and OS.²⁵² It combined the carbamate moiety of rivastigmine and the indole-amino moiety of the MAO-B inhibitor rasagiline to target both MAO-B and cholinesterases in the brain. This compound was intensively evaluated as being the most representative MTDL, which has completed a Phase II clinical trial for the treatment of mild cognitive impairment and AD-type dementia (NCT01354691). Preclinical studies showed that Ladostigil 77 improves memory in both scopolamine-induced and streptozocin-induced cognitive impairment rat models. Further experimental data revealed that Ladostigil is able to regulate APP processing, exert neuroprotective properties, and reduce microglial activation in addition to its MAO and cholinesterase inhibitory properties (IC₅₀ > 100, 37.1, 31.80 and 1.98 µM toward MAO-A, MAO-B, AChE, and BuChE, respectively).

Another example of a dual AChE and MAO-B inhibitor is the pyrimidinylthiourea family. Lead derivative 1-ethyl-3-(6-(4-((methyl(prop-2-yn-1-yl)amino)methyl)-1H-imidazol-1-yl)pyrimidin-4-yl)thiourea (78, Figure 19) developed by Li et al.²⁵³ displayed good and balanced inhibitory activities against AChE (IC₅₀ = 0.324 μM) and MAO-B (IC₅₀ = 1.427 μM). Molecular docking studies showed that the pyrimidinylthiourea moiety binds AChE CAS and the propargylamine moiety directly interacts with MAO-B flavin adenine dinucleotide (FAD). Moreover, compound 78 showed antioxidant ability (ORAC = 1.126 T.e.), and specific Cu²⁺-chelating ability producing selective 78-Cu²⁺ complexes with the thiourea fragment as responsible for the metal chelating action. Additionally, derivative 78 showed significant inhibitory activity against Cu²⁺-induced Aβ_1-42 aggregation, which is related to its neuroprotective properties in rat primary cortical neurons challenged with Cu²⁺-induced Aβ neurotoxicity. Furthermore, 78 showed low cytotoxicity and appropriate BBB permeability in vitro. Furthermore, the 78-hydrochloride derivative (30 mg/kg) significantly improved memory and cognitive function in the scopolamine-induced cognitive impairment in a mouse model, showing shorter escape latency and less frequent errors in the channel water maze.

Reis et al.²⁵⁴ reported a novel multitarget derivative as a dual AChE-MAO-B inhibitor. Its design included a chromone core functionalized with a phenylcarboxamide moiety and an acrylate subunit containing a tertiary amine to target MAO-B. From the small chemical library obtained, derivative 2-(dimethylamino)ethyl (E)-3-(4-oxo-3-(phenylcarbamoyl)-4H-chromen-6-yl)acrylate (79, Figure 19) was identified as the most promising multitarget lead. Compound 79 is a selective hMAO-B inhibitor (IC₅₀ = 0.63 µM) with cholinesterase inhibitory activity in the low micromolar range. Compound 79 demonstrated a wider safety window, with no significant decrease in cellular viability in SH-SY5Y and HepG2 cell lines.

Nadeem et al.²⁵⁵ evaluated the indole core containing 2-arylidine derivatives of thiazolopyrimidine as dual AChE/MAO-B inhibitors. Successfully, these new derivatives (80 and 81, Figure 18) exhibited interesting double inhibitory capacity with IC₅₀ values of 10.36, 0.042, and 0.069 µM for AChE, and 0.13, 0.33, and 0.14 µM for MAO-B, respectively. Moreover, all the compounds showed BBB penetration and were nonneurotoxic in the neuroblastoma SH-SY5Y cell line at 40 µM.

Li et al.²⁵⁶ reported compound 82 (Figure 19) as a dual AChE-MAO-B inhibitor (IC₅₀ = 0.58 and 0.41 μM, respectively). This compound is a novel chromanone–1-benzyl-1,2,3,6-tetrahydropyridin hybrid that exhibited low neurotoxicity, showed interesting neuroprotection against mitochondrial dysfunction and oxidation (DCF-DA fluorometric assay, JC-1 fluorescent assay), and displayed improvements in cognitive behaviors and spatial memory (Morris water maze) in a scopolamine-induced AD mouse model (intraperitoneal 10 mg/kg).

Rullo et al.²⁵⁷ have also recently reported 84 (Figure 19) as a dual AChE-MAO-B inhibitor, a coumarin-based compound with a CH₂OH/CHF₂ bioisosteric replacement. This compound displayed outstanding in vitro activities (IC₅₀ = 550 nM and 8.2 nM toward AChE and MAO-B, respectively). Moreover, it showed high solubility and brain-permeant capacity, good oral bioavailability, and negligible cytotoxic effects in SH-SY5Y and HepG2 cell lines.

Finally, Lu et al.²⁵⁸ described in 2013 a novel series of resveratrol derivates, with compound 85 as the lead compound of the study. Compound 85 inhibited AChE, MAO-A, and MAO-B (IC₅₀ = 6.27, 7.08, and 14.1 μM, respectively). This compound exhibited a potent inhibitory activity of both self-induced Aβ (82% disaggregation at 25 µM, (IC₅₀ = 6.51 µM) and Cu²⁺-induced Aβ_1-42 aggregation (94.23% inhibition at 50 µM) and a potent scavenger capacity (ORAC = 4.72 T.e.). It also has the ability to chelate biometals, such as Cu²⁺, Fe²⁺, Fe³⁺, and Zn²⁺. Besides, it significantly reduced ROS levels in SH-SY5Y at the DCFH-DA assay at 2.5 μM, and based on measured permeability in the PAMPA assay, it was also able to cross the BBB.

2.5.5 AChE inhibition and H3R antagonism

An MTDL strategy involving AChE and H3R may exert a synergistic effect on upregulating synaptic levels of ACh, which is a potential approach for the treatment of AD. Previous in vivo studies showed that H3R activation reduced ACh release in the rat frontoparietal cortex and hippocampus.²⁵⁹

One of the first examples of this type of combination was a novel series of quinoxaline derivatives developed by Huang et al.²⁶⁰ Compound 86 (Figure 20) showed potent inhibitory activities against AChE, H3R, and BACE-1. Its structure was designed to combine the 2-amino-3,4-dihydroquinazoline core of a previously reported BACE-1 inhibitor as the core ring moiety for H3R antagonism and a benzyl pyrrolidine fragment of the AChE inhibitor BYYT-25 to function as the basic center for H3R. After molecular docking optimization, new quinoxaline derivatives were synthesized and biologically evaluated. N-(3-aminoquinoxalin-2-yl)-4-(piperidin-1-ylmethyl)benzamide 86 showed the most promising profile with potent activity toward H3R, AChE, and BACE-1 targets (H3R antagonism, IC₅₀ = 280.0 nM; H3R inverse agonism, IC₅₀ = 189 nM; AChE, IC₅₀ = 483 nM; BACE 1, 46.64% inhibitory rate at 20 µg/mL) and high selectivity for H3R over H1R, H2R, or H4R.

In this line, Bajda et al.²⁶¹ described a novel multifunctional compound from a nonimidazole H3R ligand library applying docking-based virtual screening studies to select the hits that were able to bind both CAS and PAS of AChE and BuChE. The most promising derivatives combined the flavone moiety via a six-carbon atom linker with a heterocyclic moiety, including asazepane, piperidine, or 3-methylpiperidine. These compounds showed the highest inhibitory activities toward cholinesterases and well-balanced potencies against H3R. Two derivatives were chosen as good lead structures for further optimization and development of novel multitarget anti-AD agents. Derivative 2-(4-(6-(azepan-1-yl)hexyloxy)phenyl)-6-methyl-4H-chromen-4-one hydrogen oxalate (87, Figure 20) (IC₅₀ = 0.46 µM for AChE; IC₅₀ = 0.44 µM for BuChE; Ki = 159.8 nM for H3R) and its analog 6-methyl-2-(4-(6-(piperidin-1-yl)hexyloxy)phenyl)-4H-chromen-4-one hydrogen oxalate (88, Figure 20) (IC₅₀ = 0.36 µM for AChE; IC₅₀ = 0.76 µM for BuChE; Ki = 228.2 nM for H3R).

Combining AChEIs with H3R antagonism in one structure, Łażewska et al.²⁶² have reported novel acetyl- and propionyl-phenoxy-pentyl(-hexyl) derivatives 89 (Figure 20) as promising compounds to dually inhibit cholinesterases and antagonize H3Rs. Compound 89 (1-(4-((5-(azepan-1-yl)pentyl)oxy)phenyl)propan-1-one) has a high affinity for human H3Rs with a K_i of 30 nM with high capacity to inhibit hAChE and hBuChE (IC₅₀ = 3.60 and 0.55 µM, respectively). Furthermore, it also exhibited low toxicity in HepG2 and SH-SY5Y cells at 50 µM.

2.5.6 AChE inhibition and metal ions chelation

Chaves et al.²⁶³ designed novel MTDLs combining the metal chelating group hydroxyphenylbenzimidazole (BIM) to the benzylpiperidine or benzylpiperazine subunit of donepezil intending to combine AChE inhibition, Aβ aggregation, and bio-metal dysregulation modulation. Successfully, obtained derivatives were able to chelate Cu²⁺ and Zn²⁺ through their bidentate BIM moiety. Compound 90 (Figure 21), selected as the lead compound, exhibited an IC₅₀ value of 1.87 µM toward AChE, displayed the highest reduction of Aβ-induced cell toxicity, and reduced 43.7% and 66.0% self- and Cu-induced Aβ_1-42, respectively.

A series of hybrid compounds, including the bis-hydroxy-triazole tridentate, nontoxic, and highly safe Fe²⁺ chelator, deferasirox, and tacrine, were reported by Wang et al.²⁶⁴ as MTDLs against AD. Novel hybrids combined the metal chelation ability of deferasirox and the AChE inhibitory properties of tacrine. Successfully, in vitro studies showed that novel compounds exhibited good multifunctional activities, inhibiting AChE and chelating metal ions. In particular, compound 4-(3,5-bis(2-hydroxyphenyl)-1H-1,2,4-triazol-1-yl)-N-(8-((1,2,3,4-tetrahydroacridin-9-yl)amino)octyl)benzamide 91 (Figure 21) was able to chelate both Fe³⁺ and Cu²⁺ ions and displayed moderate inhibitory activity against AChE (IC₅₀ = 13.7 μM). Further evaluation demonstrated an interesting radical scavenging capacity (56% RNS reduction measured by DPPH assay), low cytotoxicity, and good neuroprotective activity against the toxicity induced by H₂O₂ in PC12 cells.

A novel series of tacrine–hydroxyphenyl benzimidazole hybrid molecules were synthesized and evaluated by Hiremathad et al.,²⁶⁵ for their multitarget activity. Due to their structure, the tacrine moiety was expected to provide AChE inhibition, while the hydroxyphenyl benzimidazole moiety should exhibit metal chelation. Among the series, several molecules were proposed as lead compounds; 92 (Figure 21) showed the highest AChE inhibition (IC₅₀  =  6.3 nM) and the most effective against the neurotoxicity induced by the Fe/Asc (500 μM/5 mM, respectively) in SH-SY5Y; 93 (Figure 21) with the highest self- and Cu-induced Aβ aggregation inhibition (74.6% and 77.3%) thanks to its metal chelating capacity; and 94 (Figure 21), which exhibited the best antioxidant (radical scavenging) capacity (EC₅₀ = 564 μM in the DPPH assay).

Fancellu et al.²⁶⁶ developed a strategy for repositioning the well-known AChE inhibitor tacrine by coupling it to benzofuran derivatives. Out of this series, compound 95 (Figure 21) outstood for its good activity for AChE inhibition (IC₅₀ = 0.13 μM), Aβ self-aggregation inhibition (57.8%), metal (Cu²⁺ and Fe²⁺) chelating ability, and neuroprotective effects (at 20 µM) in Aβ₄₂- and L-ascorbic Acid (AscH(-))/ferrous sulfate-induced toxicity assay in SH-SY5Y cells.

Finally, Jalili-Baleh et al.²⁶⁷ synthesized and evaluated a series of coumarin-lipoic acid conjugate as anti-AD agents, being 96 (Figure 21), the lead compound of the series with potent AChE (IC₅₀ = 16.4 μM) and Aβ self- and AChE-induced aggregation (51.2% and 47.4%, at 100 μM, respectively) inhibitory activities. This compound had also inhibitory effects on intracellular ROS formation, selective bio-metal chelation (Cu²⁺ and Fe²⁺), and significantly protected PC12 cells against H₂O₂-induced death (64% at 10 μM) and Aβ₄₂-induced cytotoxicity (64.7% at 10 μM).

2.6.1 UPS and autophagy activation

As previously described, the UPS system is affected in AD contributing to aberrant protein aggregates accumulation including Aβ and hyperphosphorylated tau.²⁶⁸ Thus, UPS activation has been proposed as a potential target for AD treatment facilitating protein aggregates clearance. In this line, several attempts to find small molecules with the ability to activate and/or accelerate the UPS system have reported different types of molecules able to induce its activity.²⁶⁹ In 2017, a concise search of UPS activators revealed different types of active small compounds, including p38 MAPK inhibitors (PD169316, 97, Figure 22) and MK2 inhibitors (98, Figure 22).²⁶⁹

In this line, Santoro²⁶⁹ et al., recently reported a novel class of UPS activators from pyrazolone core. The authors tested a small chemical library from which aminopyrine (99, Figure 22) demonstrated UPS activation capacity in vitro and in neuronal cells. It was further demonstrated that aminopyrine stimulates the ChT-L activity of the 20 s proteasome by binding to the CPα rings. Interestingly, this compound showed a neuroprotective effect in vitro against Aβ₄₂-induced cytotoxicity.

In this sense, there are a few groups whose multitarget strategies have been also focused on autophagy, reporting several compounds with promising results. This is the case of compound 100, described in 2019 by Sestito et al.,²⁷⁰ as a novel NMDAR antagonist (Ki = 458 nM) equivalent of memantine with hydrogen sulfide-releasing abilities (a gaseous transmitter with dual roles as a neuromodulator and smooth muscle relaxant). AD neuropathology is marked by a sharp decrease in H₂S generation in the CNS. Thus, an exogenous substance that produces H₂S also promotes neuroprotection against oxidative damage, preventing inflammation, encouraging cell proliferation, and maintaining the mitochondrial function, eventually providing defense against damage caused by Aβ to cells. On functional assays, 100 exhibits neuroprotective and ROS scavenging properties in vitro at 10 μM in a model of inflammation induced by LPS and TNF-α in H9-derived human neural stem cells. Besides, in another experiment with H9 and rat microglia cells, 10 μM of 100 reduced amyloid oligomer formation when incubated with Aβ_1-42. Autophagy induction capacity was also demonstrated in U-87MG cells, a human glioblastoma cell line, which has autophagy impairment due to elevated mTOR. In this sense, treatment with 10 μM of 100 significantly altered the expression of common autophagy markers, including LC3II, p62, mTOR, and Akt, according to western blot studies. Finally, in the same cell line, 24 h treatment with 10 μM of 100 did not exert cytotoxicity as demonstrated in the MTT assay. In this line, a similar pro-drug conjugate approach was reported by García-Viñuales et al.²⁷¹ in which silybin B was conjugated to a trehalose moiety to generate an interesting anti-Aβ aggregation derivative with neuroprotective properties.

In 2021, Lin et al. described compound 101 as a tacrine/genipin derivate.²⁷² Based on the hypothesis that autophagy has been shown to directly impact the formation of Aβ plaque by influencing the secretion of Aβ into the extracellular space, they decided to combine the AChE inhibitor activity of tacrine with the autophagy inductor genipin, more specifically with one of its derivates: CHR21, resulting in 101. In this study, a functional assay was performed on SH-SY5Y exposed to cytotoxic 35 mM glutamate, which led to cell death through different mechanisms: programmed cell apoptosis, upregulation of protein levels, such as AChE, i-NOS, and formation of A oligomers of high molecular weight, while downregulated the ratios of LC3-II/LC3-I and phospo-ULK/ULK (autophagy-related markers). In this sense, treatment with 1.25 µM 101 corrected all of these glutamate-induced dysfunctions: it inhibited glutamate, increased AChE (IC₅₀ = 5.8 nM) and BuChE (IC₅₀ = 3.7 nM) levels, decreased NO and iNOS levels,  reduced proapoptotic proteins p53 and Bax2 expression, and upregulated autophagy-related markers. 101 was also able to cross the BBB in the PAMPA assay.

More recently, Varma et al. reported in 2023 a novel 4-methoxybenzaldehyde thiosemicarbazone derivate, 102, assed for cholinesterase inhibitory activity (IC₅₀ of 30.23 µM), NMDA antagonism (IC₅₀ of 89.4 µM) Compound 102 showed, also, anti-inflammatory properties, antioxidant potential and autophagy induction capacity. In the functional assays, BV-2 cells served as a model of nitrite production when activated by LPS, 102 significantly decreased NO levels (IC₅₀ = 89.9 µM). For the assessment of autophagy, SH-SY5Y cells were transfected to express the transgene mCherry-GFP-LC3 (a protein that monitors the formation of autophagosomes and autolysosomes), reporting that treatment with 5 µM of 102 significantly increased puncta fluorescence (an indicator of upregulated autophagy flux). Finally, for the study of the antioxidant activity, treatment with 25 µM of 102 significantly reduced ascorbate consumption.

2.6.2 Aβ and tau aggregation inhibitors

An alternative approach to facilitate aberrant protein clearance is the inhibition of aberrant protein aggregation. In this line, a plethora of compounds has been designed with the ability to interact with Aβ or p-Tau oligomers avoiding or imitating their aggregation capacity. Considering the critical role of Aβ aggregates in the pathogenesis of AD,²⁷³ their pharmacological targeting has been considered a promising approach for the treatment of the disease. Indeed, the recent satisfactory results obtained with anti-Aβ monoclonal antibodies in phase III clinical trials for AD, able to reduce the Aβ burden leading to cognition improvement, confirm their therapeutic potential. In this line, compounds able to reduce Aβ aggregation could provide similar protective effects. In a similar fashion, tau aggregates are also deeply involved in AD pathogenesis. Thus, compounds able to simultaneously inhibit the aggregation of Aβ and tau, therefore acting on two of the main pathogenic pathways, could provide a more efficient treatment for AD.

Taniguchi et al.²⁷⁴ evaluated the anti-aggregation properties of several classes of compounds and found various phenothiazines and polyphenols, including the well-known methylene blue, with strong anti-Aβ and anti-tau aggregation activities. Based on this work, Fuse and coworkers²⁷⁵ developed a new series of thiophene-based compounds with a dual anti-aggregation effect. Among them, derivative 103 (Figure 23) was the most potent, exhibiting submicromolar IC₅₀ values for inhibition of the aggregation of both pathogenic proteins (IC₅₀ = 0.29 μM for Aβ_1-42 aggregation and IC₅₀ = 0.30 μM for tau aggregation). A series of AD-oriented multitarget 1-benzylamino-2-hydroxyalkyl derivatives was reported in 2018 by Panek et al.²⁷⁶ and then further expanded.²⁷⁷ Novel compounds showed different pharmacological profiles based on their activities as eeAChE, hBuChE, hBACE1, Aβ_1-42 aggregation, and tau aggregation inhibitors. Derivative 104 showed an overall balanced activity as an inhibitor of Aβ_1-42 aggregation (IC₅₀ = 3.09 μM), tau aggregation (69% at 10 μM), hBuChE (IC₅₀ = 5.74 μM), and hBACE1 (IC₅₀ = 41.60 μM). Additionally, the authors showed through kinetic studies with Aβ_1-40 that compound 104 could act at several steps of Aβ aggregation, as it could avoid aggregation via interaction with soluble Aβ_1-40 while also avoiding fibers elongation and exhibiting disaggregating properties.

In 2020, Wichur et al.²⁷⁸ developed a series of 1-benzylpyrrolidine-3-amine multitarget compounds that combined Aβ_1-42 and tau anti-aggregation properties with additional activities. Among them, compound 105 was selected as the hit compound based on its overall pharmacological profile. It showed potent inhibition of Aβ_1-42 (45% at 10 μM) and tau (54% at 10 μM), eqBuChE inhibition (IC₅₀ = 2.39 μM), free radical scavenging properties, and metal chelating capacity.

Pérez-Areales et al.²⁷⁹ developed a novel series of derivatives by merging 7-methoxy-2,2-dimethylchroman-6-ol and 6-chlorotacrine scaffolds, which showed an interesting AD-oriented pharmacological profile. Novel compounds were Aβ_1-42 and tau aggregation inhibitors as well as AChE, BuChE, and BACE1 inhibitors while showing good BBB permeability according to the PAMPA assay. The authors evaluated compound 106, which showed an overall balanced pharmacological profile (Figure 23), in the double-transgenic APP/PS1 mice. Dairy intraperitoneal 10 mg/kg doses of the compounds for 4 weeks led to a reduction of Bace1 gene expression, an increase in GPX1 and NRF2 protein levels, and Hmox1 gene expression. Additionally, compound 106 reduced sAPPβ protein levels and increased Adam10 gene expression. 106 treatment led to positive tendencies in improving cognition and reduction of amyloid burden and OS, although most of the observed effects did not reach statistical significance.

Similarly, Szalaj et al.²⁸⁰ reported a series of novel multitarget compounds that included inhibition of Aβ and tau aggregation in their multifaceted profile. Compound 107 showed interesting inhibition of Aβ_1-42 and tau aggregation in cells (59% and 56% inhibition at 10 μM, respectively) with additional activities as inhibition of hBACE1 (IC₅₀ = 4 μM), hAChE (IC₅₀ = 26 nM), and eqBuChE (IC₅₀ = 5 nM) and antagonism of the human 5-HT₆ receptor (K_i = 94 nM).

Curcumin is one of the better-known compounds with dual Aβ and tau anti-aggregation effects (IC₅₀ Aβ_1-42 aggregation = 5.4 μM and IC₅₀ tau aggregation = 3.0 μM).²⁸¹ Based on its structure, Narlawar and coworkers²⁸² generated a library of curcumin-derived isoxazoles and pyrazoles, which showed tau aggregation inhibition properties and high affinity for Aβ_1-42 aggregates. Later, Okuda et al.²⁸¹ developed novel curcumin derivatives with improved potency against both targets. Among them, compound 108 (PE859) was selected as a hit compound. 108 exhibited a slightly superior inhibition of Aβ_1-42 and tau aggregation (IC₅₀ = 1.2 and 0.66 μM, respectively) than curcumin. Additionally, compound 108 showed a good pharmacokinetic profile, being able to reach the CNS following oral administration. 108 therapeutic effect was evaluated in JNPL3 human tau P301L transgenic mice.²⁸³ Treatment with 108 at a 40 mg/kg/day dose for 6 months was able to reduce sarkosyl-insoluble tau levels in the spinal cord of the mice, leading to an improvement of the characteristic motor dysfunction of the mice, measured via tail hanging and rotarod tests. Encouraged by these results, the authors also evaluated compound 108 in the senescence-accelerated mouse prone 8 (SAMP8) mice model.²⁸⁴ 108 was orally administered at 1 and 3 mg/kg/day doses to 2-month SAMP8 mice for 9 weeks. Treatment with 108 was able to improve mice's cognitive function in the Y-maze test at both doses, although no statistical differences with the control group were observed in the Morris water maze and rotarod tests. Nevertheless, 108 was able to reduce insoluble Aβ_1-40 and insoluble tau levels, with no effect on the levels of their soluble forms.

Pérez-Areales and coworkers²⁸⁵ also reported another family of compounds based on huprine structure, the shogaol-huprine hybrids, which showed Aβ_1-42 and tau anti-aggregating activities. The most potent compound with this dual activity among the reported derivatives was compound 109, which was able to reduce Aβ_1-42 and tau aggregation at 10 μM in cells (71% and 51% inhibition, respectively). 109 also displayed inhibition of hAChE (IC₅₀ = 18.3 nM), hBuChE (IC₅₀ = 742 nM), free radical scavenging effects (10.5 and 8.8 Trolox equivalents in ABTS and DPPH assays), and BBB permeability in the PAMPA assay.

Sola et al.²⁸⁶ also reported a family of multitarget compounds based on huprine and levetiracetam structures with moderate dual Aβ_1-42 and tau anti-aggregation properties. Compound 110, which inhibited Aβ_1-42 (36% inhibition at 20 μM) and tau (54% inhibition at 20 μM) aggregation, hAChE (IC₅₀ = 4.2 nM), and hBuChE (IC₅₀ = 232 nM), showed target-engagement in C57BL6J mice following intraperitoneal 5 mg/kg administration, showing a 43% inhibition of brain AChE activity. Additionally, the potential therapeutic effect of 110 was evaluated in the APP/PS1 mice AD model. Daily intraperitoneal administration of 110 at 5 mg/kg dose for 4 weeks to 5 months in APP/PS1 mice was able to reduce the spontaneous epileptic seizures and, also, to improve their cognition in the two-object recognition test. At the molecular level, 110 treatment reduced the cortical Aβ burden of the APP/PS1 mice, while no statistically significant reduction was observed for the soluble Aβ_1-42 and Aβ_1-40 levels. Additionally, 110 treatment led to a reduction of the GFAP-positive astrocytes and Iba1-positive microglial cells surrounding Aβ plaques.