Synthesis and biological evaluation of thieno[3,2-c]pyrazol-3-amine derivatives as potent glycogen synthase kinase 3β inhibitors for Alzheimer’s disease

Abstract Glycogen synthase kinase 3β (GSK-3β) catalyses the hyperphosphorylation of tau protein in the Alzheimer’s disease (AD) pathology. A series of novel thieno[3,2-c]pyrazol-3-amine derivatives were designed and synthesised and evaluated as potential GSK-3β inhibitors by structure-guided drug rational design approach. The thieno[3,2-c]pyrazol-3-amine derivative 16b was identified as a potent GSK-3β inhibitor with an IC50 of 3.1 nM in vitro and showed accepted kinase selectivity. In cell levels, 16b showed no toxicity on the viability of SH-SY5Y cells at the concentration up to 50 μM and targeted GSK-3β with the increased phosphorylated GSK-3β at Ser9. Western blot analysis indicated that 16b decreased the phosphorylated tau at Ser396 in a dose-dependent way. Moreover, 16b effectively increased expressions of β-catenin as well as the GAP43, N-myc, and MAP-2, and promoted the differentiated neuronal neurite outgrowth. Therefore, the thieno[3,2-c]pyrazol-3-amine derivative 16b could serve as a promising GSK-3β inhibitor for the treatment of AD.


Introduction
Alzheimer's disease (AD), characterised by memory loss and cognitive impairments, is a chronic neurodegenerative disorder that disturbs more than 50 million people's healthy life worldwide 1,2 . At present, only a few drugs, donepezil, galantamine, rivastigmine, tacrine, memantine, huperzine A (NMPA), GV-971 (NMPA) and aducanumab ( Figure 1(A)), are available for the treatment of this disease [3][4][5][6] ; however, there are no drugs that can effectively block or reverse the progression of AD, possibly due to the complicated aetiology of this disease. A number of hypotheses have been proposed for AD pathogenesis [7][8][9][10][11][12][13][14][15][16][17][18] , among which, the b-amyloid (Ab) deposit and tau protein hyperphosphorylation are the key concerns 19,20 . The coexistence of Ab plaques and tau intracellular neurofibrillary tangles (NFTs) in the neocortex is associated with the collapse of neural circuits and cognitive decline, and the interactions between Ab and tau exaggerate the pathology of AD 21,22 .
Glycogen synthase kinase-3 (GSK-3), a proline-directed serine/ threonine kinase, is closely associated with Ab deposits and tau hyperphosphorylation. GSK-3 has two subtypes, GSK-3a (51 kDa) and GSK-3b (47 kDa) in mammals. Most notably, in the brain, GSK-3b is the primary isoform and acts as the dominator for tau hyperphosphorylation 23,24 . The overactivation of GSK-3b was identified and co-localized with neurofibrillary tangles (NFTs) in postmortem AD brain 25,26 . Hyperphosphorylated tau lost the physiological ability to bind to tubulin, and therefore, detached from tubulin, resulting in the formation of paired helical filaments (PHFs) and subsequently aggregated to NFTs 27,28 . The abnormal deposition of NFTs led to extensive damage to the normal transport and signalling pathways, cell cytoskeleton, mitochondria, and neuronal cell death 29 . In addition to the tau pathway, GSK-3b could promote the Ab fibril generation and induce Ab aggregation 30 . In transgenic AD mice, the inhibition of GSK-3b could reduce the Abinduced toxicity and improve cognition performances 31 . Moreover, the overactivation of GSK-3b could cause neuroinflammation, neuronal death, and apoptosis 32,33 . In the light of the multifunctional roles of GSK-3b in AD pathology, GSK-3b becomes a potential target for the development of anti-AD drugs 34,35 .
Tideglusib (Figure 1(B)) is the small thiadiazolidinone GSK-3b inhibitor that entered the clinical trial for the treatment of AD 36 . Besides, the GSK-3b inhibitors, AR-A014418, AZD2858, AZD1080, as well as the GSK-3b and acetylcholinesterase (AchE) dual inhibitor 13, demonstrated anti-AD effects in AD animals 26,[37][38][39] (Figure 1(B)). Most of the reported GSK-3b inhibitors feature with the "double-sites occupation" pharmacophore model: a key skeleton interacted with the hinge region by forming two hydrogen bonds with Asp133 and/or Val135, a moiety connected to the key skeleton as hydrogen bond acceptor to interact with Lys85 side chain 40 . Following this model, we used the thieno [3,2-c]pyrazol-3amine as the key framework and designed a series of thieno [3,2c]pyrazol-3-amine derivatives as the potential GSK-3b inhibitors. The thieno [3,2-c]pyrazol-3-amine framework has the possibility to form triple hydrogen bonds with the hinge region so as to enhance its binding with the enzyme (Figure 2). The N atom of pyridine moiety connected with the thieno [3,2-c]pyrazol-3-amine may act as hydrogen bond acceptor to interact with Lys85 side chain. A variety of the substituents (R 1 , R 2 and R 3 ) were introduced to investigate their effects on the GSK-3b inhibitory activities.

GSK-3b inhibitory activity and kinase selectivity
All the targeted compounds were evaluated for their GSK-3b inhibitory activities in the calliper mobility shift assay in vitro. AR-A014418 (10, Figure 1(B)), a prototypical GSK-3b-specific inhibitor, was used as the positive control 37 .
At first, the effects of the acyl or sulphonyl groups at the thieno[3,2-c]pyrazol-3-amine on GSK-3b inhibitory activities were investigated. As shown in Scheme 1 and Table 1, the cyclopropanecarbonyl and the isobytyryl group showed similar effects on the GSK-3b inhibitory potency. Compounds 16a and 16b were very potent GSK-3b inhibitors with the IC 50 values of 4.4 nM and 3.1 nM, respectively. When the thieno[3,2-c]pyrazol-3-amine was substituted by the n-butyryl (16c) or benzoyl (16e), the resulting compound 16c or benzoyl 16e maintained high potency, but was about 10-fold less active than 16b. However, the sulphonamide 16d showed very weak GSK-3b inhibitory activity. The introduction of a phenyl group at the meta-position of the pyridine ring in 16a-16d showed subtle influences on the activity of their parent compounds. Compounds 17a and 17b were about 4-fold less active than 16a and 16b, but compound 17c was active as that of 16c. The sulphonamide 17d was not active. Therefore, the substitution of the thieno[3,2-c]pyrazol-3-amine with a sulphonyl group was not preferred.
As compound 16b showed very potent GSK-3b inhibitory activity, further structural modifications based on 16b were made at the para-position of the pyridine ring. In general, inducing a phenyl or substituted phenyl, a biphenyl, a naphthalenyl group at this position decreased the potency of 16b. The phenyl substituted analogue 18a showed modest GSK-3b inhibitory activity with the IC 50 of 84 nM. For the methyl or methoxy substituted phenyl derivatives, the activity was in the order of ortho-> para-> meta-. The ortho-methyl and the ortho-methoxy derivatives 18b and 18e were slightly more active than 18a, with IC 50 values of 64 and 64 nM, respectively. For the fluorine substituted phenyl analogues, the 4-F-phenyl derivative 18j was highly active, with an IC 50 of 15 nM, possibly due to the special features of fluorine atom with the smallest size and the largest electron-withdrawing property. The trifluoromethyl group (18k and 18l) also decreased the potency of 18a slightly. The biphenyl derivative 18m showed much weak activity with an IC 50 of 387 nM. The naphthalenyl derivative 18n was about 2-fold less active than 18a.
The potent GSK-3b inhibitor 16b was next subjected to kinase selectivity assay. A panel of kinases which were structurally related to GSK-3b was used for the GSK-3b selectivity studies 26,37,38 . Among a panel of 21 diverse kinases, at the concentration of 1.0 lM (>320-fold IC 50 value on GSK-3b), 16b showed in general good selectivity over most of the kinases except for the low selectivity over the GSK-3a and CDK5 and moderate selectivity over CK2 (Figure 3). GSK-3a and GSK-3b are known to share a  98% sequence homology at the catalytic site 42 . CDK5 and GSK-3 belong to the CMGC protein kinase family, which shared highly homology with each other 43,44 . CDK5 is known abnormally activated and also responsible for the tau hyperphosphorylation in AD 45 . It was known that Ab could increase CK2 activity, which in turn accelerated the tau phosphorylation in AD 46 . Therefore, the inhibition of CDK5 and CK2 by 16b may be beneficial for its anti-AD activities.

Molecular docking study
To investigate the possible binding mode of compound 16b with GSK-3b, molecular docking was performed on GSK-3b (PDB: 4ACG) 38 using Sybyl-X 2.0 softsuite. All docked conformations were ranked based on docking scores. As depicted in Figure 4(A and B), compound 16b fitted well into the ATP binding pocket of GSK-3b. The thieno [3,2-c]pyrazol-3-amine skeleton occupied the adenine pocket and formed triple hydrogen bonds with backbone atoms of Asp133 and Val135 in the hinge region, which is necessary for the ligand recognition. In addition to these hydrogen bonds, the thieno[3,2-c]pyrazol-3-amine portion also made hydrophobic interactions with the hydrophobic pocket formed by the residues of Ala83, Val110, Leu132, Asp133, Tyr134, Val135, and Leu188. The pyridine ring also participated in hydrophobic interactions with Phe67, Val70, and Cys199. Meanwhile, the N atom of pyridine served as a hydrogen bond acceptor to interact with Lys85 which located at the b-strand of N-terminal domain. Besides, the terminal isobytyryl group located at a position adjacent to the hinge region and produced hydrophobic interactions with Ile62 and Pro136. As expected, 16b followed the "doublesites occupation" pharmacophore model well that may contribute to its high inhibitory activity with GSK-3b.
GSK-3b is implicated in the Wnt/b-catenin signalling pathway, which plays an important role in neuronal development 29 . GSK-3b, together with adenomatous polyposis coli (APC), Axin, and casein kinase 1 (CK1), form a ploy-protein complex that regulates the hyperphosphorylation of b-catenin. Phospho-b-catenin is recognised by ubiquitin and degraded by proteasomes [52][53][54] . Pharmacological inhibition of GSK-3b leads to the activation and stabilisation of b-catenin, subsequently resulting in the accumulation of b-catenin in cytoplasm [55][56][57] . The activation of Wnt/b-catenin signalling pathway can promote synaptic growth, alleviate spatial memory impairment and neurodegeneration in Alzheimer's models [58][59][60] . Moreover, b-catenin also plays a pivotal role in cell adhesion complexes. The combination of b-catenin and N-cadherin elevates cell-to-cell interactions which is prerequisite for neuronal differentiation 61,62 . Therefore, we further evaluated the effect of 16b on b-catenin. In agreement with its GSK-3b inhibitory activity on SH-SY5Y cells, 16b increased b-catenin abundance in a dose-dependent manner. As shown in Figure 6(B), after treatment with 16b at the concentration of 5 lM, 10 lM and 20 lM, the b-catenin/GADPH ratio increased from 0.41 of the control to 0.54, 0.64, 0.76, respectively.

Inhibition of Ab-induced tau protein hyperphosphorylation
GSK-3b phosphorylates tau at sites (Ser199, Ser396, and Ser443), and the hyperphosphorylated tau aggregate into NFTs in AD 63 . Inhibition of GSK-3b could reduce tau hyperphosphorylation. Therefore, the cell-based assay examining Ab-induced tau phosphorylation at Ser396 represents a direct functional assay to measure the cellular activity of GSK-3b inhibitors 64 . To investigate the Figure 5. Cell viability of SH-SY5Y cells exposed to compound 16b at different concentrations (range from 3.125 À 50 lM) for 24 h. Vehicle treater cells were used as control. The results were expressed as the percentage of viable cells observed after treatment with compound 16b respect to vehicle-treated cells (100%) and shown as the mean ± SD from at least three separate experiments.

Effects of 16b on neuronal neurite outgrowth and GAP43, N-myc, and MAP-2 expressions in SH-SY5Y cells
Among kinds of pathological symptoms, neurogenesis impairment and neuronal loss play important roles in neurodegeneration in AD. Therefore, regulating neurogenesis is considered to be a promising therapeutic option for AD 65 .
Compelling evidence indicated that GSK-3b plays a large part in synaptic plasticity and neurogenesis 66,67 . Differentiated SH-SY5Y cells express neurogenesis-related markers, including growth-associated protein 43 (GAP43), the N-myc gene as well as neuronal polarity marker microtubule-associated protein 2 (MAP-2). GAP43 is an intrinsic determinant of neuronal development and plasticity which regulates axon growth and regeneration 68,69 . N-myc is indispensable to normal neurogenesis in the expansion of progenitor cell populations 70 . MAP-2 regulates neuronal development, structural stability and synaptic plasticity through the formation of axonal and dendritic processes 71,72 . Inhibition of GSK-3b could induce neurogenesis and promote the expressions of the neurogenesis-related markers 73 . Firstly, we used SH-SY5Y cells to study the effect of 16b on neurite outgrowth, using retinoic acid (RA) as a positive control 74 . After incubating with 16b (10 lM) or RA (10 lM) for 72 h, the morphology of differentiated neuronal neurite outgrowth were obtained. As depicted in Figure 8(A), 16b exhibited a substantial ability in inducing SH-SY5Y cells neurite outgrowth compared with RA.
We next quantified the neurogenesis-related markers to evaluate the effect of 16b on neurogenesis. Co-incubated 16b (10 lM) or RA (10 lM) with SH-SY5Y cells for 24 h, the mRNA expressions of GAP43, N-myc and MAP-2 were measured by the quantitative real-time reverse transcription-PCR (RT-PCR) analysis. As shown in Figure 8

Conclusion
In summary, a series of novel thieno[3,2-c]pyrazol-3-amine derivatives were designed, synthesised, and evaluated as potential GSK-3b inhibitors. Compound 16b exhibited potent GSK-3b inhibitory activity with IC 50 at single-digit nanomolar level. In a panel of 21 kinases, 16b showed in overall good selectivity over most of them except for the CDK5 and CK2 kinases. In cellular level, 16b showed no cytotoxicity against SH-SY5Y cells at the concentration up to 50 lM and inhibited GSK-3b through the up-regulation of the phosphorylation at Ser9. Meanwhile, 16b inhibited the Abinduced tau protein hyperphosphorylation at Ser396. In addition, b-catenin plays a crucial role in neurogenesis. Compound 16b inhibited GSK-3b, interfered the physiological degradation of b-catenin, resulting in the abundance of b-catenin. Moreover, 16b could increase the mRNA expressions of the recognised neurogenesis-related markers and promote the differentiated neuronal neurite outgrowth, which is very useful in face of progressive neurogenesis impairment and neuronal loss in AD. Compelling evidence have verified GSK-3b to process function link between Ab and tau. In view of the involvement in multiple pathways in AD progress, GSK-3b is being an interesting drug discovery target for the treatment of AD. As a potent GSK-3b inhibitor, 16b could serve as a promising lead for further investigation in facing the complicated pathogenesis of AD.

Chemistry
All solvents used were commercially available and were used without further purification unless otherwise noted. Starting materials used were either available from commercial sources or prepared according to literature procedures. For examined compounds, 1 H and 13 C nuclear magnetic resonance (NMR) spectra were recorded on a Bruker-400 or Bruker-600 NMR spectrometer, respectively. The following reference signals were used: TMS d 0.00, or the residual solvent signal of DMSO-d 6 d 2.50 ( 1 H), d 39.52 ( 13 C). MS spectra data were obtained using an API 4000 instrument. Highresolution mass spectrometry (HRMS) data were acquired by an AB-Triple TOF5600 or Agilent Q-TOF6540 instrument. Melting points (Mp) were measured on an X-6 micromelting point apparatus (Beijing Tech. Co., Ltd., Beijing, China). The reactions were followed by thin-layer chromatography (TLC) and visualised in an iodine chamber or with a UV lamp. Compounds were purified by column chromatography using silica gel (200 À 300 mesh). The purity (> 95%) of samples were determined using a Shimadzu LC-20AT series system (column, Shim-pack GWS C18, 4.6 Â 250 mm,

Kinase selectivity screen
Compound 16b was evaluated for kinase selectivity at Eurofins Cerep SA (Celle-L'Evescault, France) in enzymatic radioactive assays in a panel of 21 different kinases (including GSK-3b) from diverse families. Protein kinase were assayed in the presence of 1.0 lM compound 16b or vehicle (DMSO). The enzymatic activity was measured in the presence of K m ATP.

Molecular docking
Molecular docking was performed on the Sybyl-X 2.0 software (Tripos, St. Louis, MO) and the X-ray crystal structure of GSK-3b (PDB: 4ACG) was obtained from the RCSB Protein Data Bank. The protein was added with hydrogen atoms and charges. Waters were removed from the PDB file. Native ligand (6LQ) was extracted from the protein and used as a standard to generate the protomol. The binding pocket was defined as all residues within 5 Å of the original ligand. Finally, docking was performed by using the Surflex-Dock mode, and the conformations were used to analyse the interactions between ligand and GSK-3b. UCSF Chimaera 1.16 was used to visualise the result of docking 76 . Maestro 11.9 was used to show the 2D interactions diagram 77 .

Cytotoxicity on SH-SY5Y cell line
SH-SY5Y cells were cultured in DMEM/F12 (Dulbecco's modified Eagle medium and Ham's F-12, 1:1) with 10% FBS (fetal bovine serum), 1% penicillin and 1% streptomycin under 5% CO 2 atmosphere at 37 C. SH-SY5Y cells (1 Â 10 5 ) were seeded in 96-well plates and incubated for 24 h. Different concentrations of compound 16b were added into each well and incubated for another 24 h. The survival of cells was determined by MTT assay and the absorbance of each well were measured using a SpectraMax M5 multimode plate reader at 570 nm. Results were expressed as percentage of control and statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software Inc., San Diego, CA, USA).

Western blot analysis on p-GSK-3b and b-catenin
SH-SY5Y cells (1 Â 10 6 ) were seeded in 12-well plates (Corning, Los Altos, MA, USA) and incubated with compound 16b or LiCl at the indicated concentrations for 2.5 h at 37 C in 5% CO 2 . At the end of incubation, cells were lysed by addition of ice-cold RIPA buffer containing a protease inhibitor cocktail. The protein quantification was determined using a BCA protein assay kit (Jiangsu KeyGEN BioTECH Co., Ltd., Nanjing, China). Cellular lysates were mixed with an equal volume of SDS (sodium dodecyl sulfate) loading buffer (Jiangsu KeyGEN BioTECH Co., Ltd., Nanjing, China) and separated by electrophoresis (Bio-rad Power Supplies Basic, Shanghai, China) in polyacrylamide gel. Proteins were transferred from acrylamide gels to nitrocellulose membranes and blocked in a blocking buffer (PBS, 5% non-fat milk) for 1.5 to 2 h at 20 C. After overnight incubation at 4 C with primary p-GSK3b-Ser9 (Cell Signalling Technology, Danvers, MA, USA), or b-catenin (Cell Signalling Technology, Danvers, MA, USA), and GAPDH (Santa Cruz Biotechnology, Shanghai, China), the blots were washed in Tween 20-TBS (TBST, Jiangsu KeyGEN BioTECH Co., Ltd., Nanjing, China) for 20 min and then incubated with secondary antibody (IgG-HRP; Jiangsu KeyGEN BioTECH Co., Ltd., Nanjing, China) for 1 h at room temperature. The blots were washed by TBST for 20 min and detected with ECL chemiluminescent reagent (Jiangsu KeyGEN BioTECH Co., Ltd., Nanjing, China) for 3 min. Pixel intensity was quantitated using gel imaging system (SYNGENE G:BOX/iChemi XR5, ISS, San Diego, CA, US) and Gel-Pro32 software (Media Cybernetics, Bethesda, MD, USA). GAPDH was used as an internal control.

Inhibition of Ab-induced tau hyperphosphorylation
SH-SY5Y cells were seeded in 12-well plates until 80% confluence, serum-deprived for 12 h. Cells were pre-incubated with compound 16b or LiCl for 1 h, thereafter stimulated with Ab 25-35 (Sigma) for another 6 h. According to the previously reported method 64,75 , the phosphorylated tau was determined.

Neuronal neurite outgrowth assay and quantitative RT-PCR
SH-SY5Y cells (5 Â 10 3 ) were planted in 96-well plates and cultivated at 37 C for 24 h. Compound (RA or 16b, 10 lM) was then added and cultivated for 72 h. The morphology of neurite outgrowth was examined under an inverted microscope (2 Â 100; Olympus, Tokyo, Japan). After the SH-SY5Y cells were cultivated for 24 h, total RNA was extracted, and quantitative RT-PCR was performed according to the previously reported method 75 .