Novel Synthetic Coumarin-Chalcone Derivative (E)-3-(3-(4-(Dimethylamino)Phenyl)Acryloyl)-4-Hydroxy-2H-Chromen-2-One Activates CREB-Mediated Neuroprotection in Aβ and Tau Cell Models of Alzheimer's Disease

Abnormal accumulations of misfolded Aβ and tau proteins are major components of the hallmark plaques and neurofibrillary tangles in the brains of Alzheimer's disease (AD) patients. These abnormal protein deposits cause neurodegeneration through a number of proposed mechanisms, including downregulation of the cAMP-response-element (CRE) binding protein 1 (CREB) signaling pathway. Using CRE-GFP reporter cells, we investigated the effects of three coumarin-chalcone derivatives synthesized in our lab on CREB-mediated gene expression. Aβ-GFP- and ΔK280 tauRD-DsRed-expressing SH-SY5Y cells were used to evaluate these agents for possible antiaggregative, antioxidative, and neuroprotective effects. Blood-brain barrier (BBB) penetration was assessed by pharmacokinetic studies in mice. Of the three tested compounds, (E)-3-(3-(4-(dimethylamino)phenyl)acryloyl)-4-hydroxy-2H-chromen-2-one (LM-021) was observed to increase CREB-mediated gene expression through protein kinase A (PKA), Ca2+/calmodulin-dependent protein kinase II (CaMKII), and extracellular signal-regulated kinase (ERK) in CRE-GFP reporter cells. LM-021 exhibited antiaggregative, antioxidative, and neuroprotective effects mediated by the upregulation of CREB phosphorylation and its downstream brain-derived neurotrophic factor and BCL2 apoptosis regulator genes in Aβ-GFP- and ΔK280 tauRD-DsRed-expressing SH-SY5Y cells. Blockage of the PKA, CaMKII, or ERK pathway counteracted the beneficial effects of LM-021. LM-021 also exhibited good BBB penetration ability, with brain to plasma ratio of 5.3%, in in vivo pharmacokinetic assessment. Our results indicate that LM-021 works as a CREB enhancer to reduce Aβ and tau aggregation and provide neuroprotection. These findings suggest the therapeutic potential of LM-021 in treating AD.


Introduction
Alzheimer's disease (AD) is a progressive and irreversible neurodegenerative disease that gradually impairs memory and cognitive function [1]. The pathological hallmarks of AD include deposits of the extracellular senile plaques and intracellular neurofibrillary tangles, which eventually lead to extensive loss of neurons and synapses and brain atrophy. Senile plaques are caused by the aggregation and deposition of amyloid β peptides (Aβ) [2], fragments of the amyloid peptide precursor protein (APP) that result from its cleavage by βand γ-secretases [3]. Aβ tends to form oligomers and other high-order polymerized structures that are toxic to neurons [4]. Neurofibrillary tangles are composed of abnormally hyperphosphorylated microtubule-associated protein tau (MAPT) [5,6], which are also involved in the neurodegenerative process [7]. Accumulation of Aβ is considered to be an early event and could trigger or accelerate tau pathology in AD [8].
The molecular mechanisms underlying AD aetiology and pathogenesis remain elusive. Compelling evidence suggests that changes in cAMP-response-element (CRE) binding protein 1-(CREB-) mediated gene expression play an important role in AD pathogenesis. CREB is a critical nuclear transcription factor enhancing cognition, memory formation, and neuronal survival [9]. CREB activation occurs primarily through the phosphorylation of serine 133 by either of several kinases: protein kinase A (PKA) [10], Ca 2+ /calmodulin-dependent protein kinase II (CaMKII) [11], extracellular signal-regulated kinase (ERK) [12], and phosphatidylinositol 3-kinase (PI3K) [13]. Phosphorylated CREB binds to the coactivators CREB binding protein (CBP) and E1A binding protein p300 (EP300), resulting in the expression of target genes [14]. CREB expression is downregulated in hippocampal neurons of APP-transgenic mice, hippocampal tissues of AD patients, and Aβ-treated rat hippocampal neurons [15]. In cultured neurons, Aβ interferes with CREB activation and expression of its downstream genes [16]. The Aβ-impaired hippocampal synaptic plasticity and long-term potentiation can be reversed by activating the CREB signaling pathway [17]. These observations provide important insights into the role of CREB signaling in the development of AD.
At present, no treatment is available to slow the neurodegeneration of AD. The central role of the CREB signaling pathway in AD neurodegeneration suggests its potential as a therapeutic target for AD. The cognitive function of presenilin 1 (PS1)/APP transgenic mice is significantly improved by increasing CREB phosphorylation with phosphodiesterase inhibitors [18,19]. In our previous study, a synthetic coumarin-chalcone hybrid LM-031 upregulates CREB and demonstrates the neuroprotective potential to reduce Aβ, tau, expanded polyglutamine (polyQ) aggregation, and oxidative stress [20][21][22]. Here, we investigate three synthetic coumarin-chalcone derivatives made in our laboratory to determine their effects on CREB expression. The neuroprotective potential of these compounds for treating AD was further evaluated using Aβ-GFP-and ΔK280 tau RD -DsRed-expressing SH-SY5Y cells [23,24].

1,1-Diphenyl-2-Picrylhydrazyl
Assay. Stable free radical 1,1-diphenyl-2-picrylhydrazyl (DPPH) (Sigma-Aldrich) was applied to test the potential radical-scavenging activity of the coumarin-chalcone derivatives [27]. In brief, the compounds (10-160 μM) were added to an ethanol solution containing 100 μM DPPH. The mixture was vortexed for 15 sec, left to stand for 30 min at room temperature, and scavenging capacity was measured at a wavelength of 517 nm using a Multiskan GO microplate spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). The formula 1 − ðabsorbance of sample/absorbance of controlÞ × 100% was used to compute the radical-scavenging activity, and the interpolation method was used to calculate the concentration of 50% of maximal effect (EC 50 ).

Thioflavin T Binding Assay and Transmission Electron
Microscopy Examination. The inhibiting potential of the curcumin and coumarin-chalcone derivatives on Aβ amyloid aggregation was assessed using Aβ 42 peptide (AnaSpec, Fremont, CA, USA). In brief, the tested compound (5-20 μM) was added to a buffer solution (150 mM NaCl, 20 mM Tris-HCl (pH 8.0)) containing Aβ 42 peptide (5 μM) and incubated at 37°C for 48 h. Thioflavin T (10 μM) (Sigma-Aldrich) then was added, and the mixture was incubated for 5 min at room temperature. Subsequently, the fluorescence intensity was quantified at 420/485 nm of excitation/emission wavelengths using a FLx800 microplate reader (Bio-Tek, Winooski, VT, USA). EC 50 was computed as described.
The potential of Congo red and coumarin-chalcone derivatives in inhibiting tau misfolding was assessed using the E. coli-derived ΔK280 tau RD protein [21]. In brief, ΔK280 tau RD protein (20 μM) was incubated with the tested 2 Oxidative Medicine and Cellular Longevity compound (1−10 μM) in buffer (150 mM NaCl, 20 mM Tris-HCl (pH 8.0)) at 37°C for 48 h to form aggregates. Thioflavin T (5 μM) was then added, and the mixture was incubated at room temperature for 25 min. The fluorescence intensity of the samples was measured, and the EC 50 was computed as described.
To investigate Aβ and ΔK280 tau RD aggregates, samples of the fibrils formed with and without treatment of a tested compound were placed on a 200-mesh copper (holey-carbon) grid and visualized using a JEM-1230 transmission electron microscope (TEM) (JEOL, Tokyo, Japan) at an accelerating voltage of 100 kV.

CRE-GFP Reporter Construct and 293 Reporter Cells.
Using the pCRE-MetLuc2-Reporter plasmid (Clontech, Mountain View, CA, USA) as a template, the DNA fragment containing the cAMP response element (CRE) motifs and TATA-like promoter (P TAL ) flanked by MluI and AgeI restriction sites was amplified by PCR using sense (ACGC GTGCACCAGACAGTGACGTC, MluI site underlined) and antisense (ACCGGTCTGCTTCATCCCCGTGG, AgeI site underlined) primers. The amplified CRE-P TAL promoter fragment was cloned into pGEM-T Easy (Promega, Madison, WI, USA) and sequenced. The CRE-P TAL -containing MluI/AgeI fragment and GFP-containing AgeI/NotI fragment (from pEGFP-N1; Clontech) were subcloned into MluI-and NotI-digested pcDNA5/FRT/TO (Novagen, Madison, WI, USA). The resulting construct with CRE-P TALdriving GFP was used to establish 293 GFP reporter cells by targeting insertion into Flp-In-293 cells according to the supplier's instructions (Invitrogen). The cells were grown in a medium containing 5 μg/ml blasticidin and 100 μg/ml hygromycin.
2.10. Caspase 1, Acetylcholinesterase, and Real-Time PCR Assays. Aβ-GFP or ΔK280 tau RD -DsRed SH-SY5Y cells were seeded into 6-well plates (5 × 10 5 /well) and treated with retinoic acid, coumarin-chalcone derivatives, and doxycycline as previously described. On day 8, cells were collected and lysates were prepared with six freeze/thaw cycles. The supernatant was collected by centrifugation, followed by measurement of the caspase 1 activity in 50 μg aliquots of cell extract using the ICE fluorometric assay kit (BioVision, Milpitas, CA, USA). Following incubation of the mixture at 37°C for 2 h, fluorescence was measured with excitation/emission wavelengths at 400/505 nm (FLx800 fluorescence microplate reader). In addition, the collected cells were lysed by sonication. Acetylcholinesterase (AChE) activity in the supernatant was determined in 10 μg aliquots of cell extracts using the AChE activity assay kit (Sigma-Aldrich) according to the manufacture's protocol. Finally, the absorbance at 412 nm was measured using a Multiskan GO spectrophotometer (Thermo Fisher Scientific).
To measure the Aβ-GFP/tau RD -DsRed RNA expression on day 8, total RNA was extracted from collected cells and reverse-transcribed to cDNA (SuperScript III reverse transcriptase; Invitrogen, Waltham, MA, USA). Real-time quantitative PCR was performed using 100 ng of cDNA and the gene-specific TaqMan fluorogenic probe PN4331348 (EGFP) or customized Assays-by-Design probe for DsRed [24] and 4326321E (HPRT1) (StepOnePlus Real-time PCR system, Applied Biosystems, Foster City, CA, USA). The RNA expression level was calculated using the formula 2 ΔCt , ΔC T = C T ðcontrolÞ − C T (target), in which C T indicates the cycle threshold. In addition, Aβ-GFP or ΔK280 tau RD -DsRed SH-SY5Y cells were treated with retinoic acid (10 μM) on day 1, followed by the addition of LM-021 (5 or 10 μM) and doxycycline (5 or 2 μg/ml) on day 2, as described. Kinase inhibitors (10 μM) were added on day 6. On day 8, the cells were collected for protein expression analysis of BDNF, BCL2 (BCL2 apoptosis regulator), and BAX (BCL2 associated X, apoptosis regulator), and total/phosphorylated PKA, CaMKII, ERK, PI3K, and CREB. Also, cells were stained with DAPI and assessed for neurite outgrowth as described above.

Mouse Pharmacokinetics Study and Brain/Plasma
Ratio Determination. Male Crl:CD-1 (ICR) mice (8 weeks old) weighing 25-30 g were housed in cages with ad libitum access to food and drinking water on a 12 : 12 h light-dark cycle. All animal experiments were performed in accordance with the guidelines approved by the Rosetta Pharmamate Institutional Animal Care and Use Committee (IACUC Approval No: AF20015). Mice were randomly divided into 7 subgroups (6 time points and 1 vehicle control; n = 3 per subgroup). LM-021 (in 5% DMSO and 19% hydroxypropyl-βcyclodextrin) was administered intravenously (IV) at 5 mg/kg dose in volume of 10 ml/kg. Mice were euthanized in a CO 2 chamber at 0.25, 0.5, 1, 2, 4, or 8 h postdose. Whole blood was collected via facial veins in heparin-coated polypropylene tubes, and plasma was obtained by centrifugation (1000 × g for 15 min at 4°C) within 60 min of collection. After cerebral vasculature perfusion through the left ventricle with normal saline for 5 min, the whole brain was quickly excised and weighed. Plasma and brain samples were rapidly frozen and stored at −70°C until analysis (see supplementary materials).
Pharmacokinetic (PK) parameters, including plasma concentration following IV dose (C 0 ), elimination half-life (T 1/2 ), mean residence time (MRT), and area under the concentration-(AUC-) time curve from time 0 extrapolated to infinity (AUC (0-∞) ), were estimated using noncompartmental pharmacokinetic parameter analysis with Phoenix WinNonlin software (version 8.2, Pharsight Corporation, Mountain View, CA, USA). In addition, the ratio of brain/plasma was calculated from the area under the curve for brain and plasma concentrations.
2.14. Statistical Analysis. Data are expressed as the mean ± standard deviation of the results from three independent experiments. Comparisons between groups were analysed using a two-tailed Student's t-test or one-way ANOVA (analysis of variance) with a post hoc Tukey test where appropriate. A p < 0:05 was considered to be statistically significant.

Tested Coumarin-Chalcone Compounds and Biochemical
Aβ/Tau Aggregation Inhibition. Coumarin-chalcone derivatives have been shown to upregulate the CREB pathway in 4 Oxidative Medicine and Cellular Longevity AD and polyQ cell models [20][21][22]. Here, we examined three synthetic coumarin-chalcone derivatives LM-016, LM-021, and LM-022 ( Figure 1(a)), all of which fulfilled Lipinski's criteria for oral bioavailability on the basis of molecular weight, hydrogen bond donors, hydrogen bond acceptors, and calculated octanol/water partition coefficient [29] ( Figure S1A). Given that all compounds have a polar surface area smaller than 90 Å 2 , they were predicted to diffuse across the BBB [30], which was also recommended by the online BBB predictor [31] ( Figure S1A). Inhibition of amyloid aggregation and oxidative stress are considered important treatment approaches for AD. The free radical-scavenging activity of these coumarinchalcone derivatives was examined using DPPH as a substrate. A natural antioxidant kaempferol [32] was used as a positive control. EC 50 values of kaempferol, LM-016, LM-021, and LM-022 were 28, 126, 181, and 143 μM, respectively ( Figure 1(b)). The Aβ or ΔK280 tau RD aggregationinhibitory effect of the test compounds was measured using Thioflavin T, a dye widely used to examine misfolded protein aggregates [33]. Curcumin or Congo red, known to reduce amyloid aggregation [34,35], was included as a positive control. As shown in Figure 1  Oxidative Medicine and Cellular Longevity (10 μM) ( Figure S1B). The results showed that LM-021 directly hampered both Aβ and ΔK280 tau RD aggregate formation.

Coumarin-Chalcone Derivatives Increase CRE-Mediated
Gene Expression. To monitor the activation of CREB, a plasmid with GFP reporter driven by the CRE motif-TATA-like promoter ( Figure S2A) was used to establish Flp-In 293 GFP reporter cells. As Ca 2+ influx triggers phosphorylation of CREB [36], the reporter cell was tested with 2 μM Ca 2+ ionophore for 3-24 h and 0.5-10 μM Ca 2+ ionophore for 5 h to obtain the optimized experimental condition. No significant cell death was observed after treatment with 2 μM Ca 2+ ionophore for 3-6 h (Figure 2(a)). Thus, Ca 2+ ionophore treatment for 5 h was selected for the following test. As shown in Figure 2(b), treatment with 0.5-2 μM Ca 2+ ionophore for 5 h stimulated CRE-motif-driven GFP expression. CRE-GFP reporter cells treated with 2 μM Ca 2+ ionophore for 5 h were then used to test the studied coumarin-chalcone derivatives ( Figure S2B). As shown in Figure 2(c), LM-021 (5-10 μM) significantly activated CRE-motif-driven GFP expression in the presence or absence of Ca 2+ ionophore in a dose-dependent manner. Forskolin (10 μM), an activator of adenylyl cyclase [37], significantly increased CRE-motif-driven GFP expression in the presence of Ca 2+ ionophore.    7 Oxidative Medicine and Cellular Longevity level through GFP fluorescence intensity, were used to determine the ability of coumarin-chalcone derivatives to inhibit Aβ aggregation (Figure 3(a)). Curcumin was used as a positive control. The green fluorescence intensity of 2.5-5 μM curcumin-pretreated cells was significantly higher than that of untreated cells. Treatment with LM-016 at 2.5-5 μM, LM-021 at 1.2-5 μM, or LM-022 at 2.5-5 μM also significantly augmented the green fluorescence intensity (Figure 3(b)). Curcumin, LM-016, LM-021, and LM-022 had an EC 50 value of 6.1, 7.0, 5.1, and 8.2 μM, respectively, in Aβ aggregation inhibition ( Figure S3A). In oxidative stress analysis, Aβ-GFP expression induced significantly elevated ROS levels in Aβ-GFP-expressing SH-SY5Y cells, while treatment with curcumin at 2.5-5 μM, LM-016 at 2.5-5 μM, LM-021 at 1.2-5 μM, or LM-022 at 2.5-5 μM effectively decreased the ROS levels caused by Aβ overexpression (Figures 3(c) and S3B). These results suggest that these three coumarin-chalcone derivatives not only impeded Aβ aggregation but also reduced ROS overproduced by Aβ. Treatment with curcumin, LM-016, LM-021, or LM-022 did not significantly modify Aβ-GFP RNA levels (Figure 3(d)), suggesting that the increases in fluorescence intensity and ROS were not caused by changes in gene expression.

Coumarin-Chalcone Derivatives Inhibit Tau Aggregation
and Promote Neurite Outgrowth. The ability of coumarinchalcone derivatives to inhibit tau aggregation and promote neurite outgrowth was tested by using Tet-On ΔK280 tau RD -DsRed SH-SY5Y cells [24] (Figure 4(a)). The poorly folded ΔK280 tau RD caused misfolding of fused DsRed, leading to decreased DsRed fluorescence [41]. For comparison, Congo red was included as a positive control. Treatment with Congo red at 10 μM and LM-021 at 5-10 μM significantly elevated the DsRed fluorescence intensity (Figure 4(b)). Based on the measured concentrations, Congo red, LM-016, LM-021, and LM-022 had extrapolated EC 50 values of 57, 70, 10, and 73 μM, respectively, in inhibiting ΔK280 tau RD aggregation ( Figure S4A). ΔK280 tau RD -DsRed expression elevated ROS levels, although the difference did not reach statistical significance. LM-021 at 10 μM effectively reduced the ROS level elevated by ΔK280 tau RD overexpression (Figures 4(c) and S4B). Treatment with Congo red or LM-021 did not significantly alter the ΔK280 tau RD -DsRed RNA level (Figure 4(d)).
The neuroprotective effects of Congo red and coumarinchalcone derivatives on caspase 1/AChE activities and neurite outgrowth were also examined. As shown in Figure 4(e), the overexpression of ΔK280 tau RD significantly enhanced caspase 1 activity, and treatment with Congo red or LM-021 (10 μM) lowered caspase 1 activity. In contrast, AChE activity was not significantly changed by ΔK280 tau RD overexpression or compound treatment (Figure 4(f)). ΔK280 tau RD overexpression significantly decreased neurite length    (Figures 4(g) and S4C).

Regulatory Targets of CREB Phosphorylation by LM-021 in CRE-GFP Reporter Cells.
To further investigate if LM-021 employs its effect by increasing the phosphorylation of CREB, the inhibitor H-89 (PKA inhibitor), KN-62 (CaMKII inhibitor), U0126 (ERK inhibitor), or wortmannin (PI3K inhibitor) (10 μM) was added to CRE-GFP reporter cells 4 h before addition of coumarin-chalcone derivatives (10 μM) and/or Ca 2+ ionophore (2 μM) ( Figure S5). As shown in Figure 5(a), forskolin (positive control) significantly activated CRE-motif-driven GFP expression in the presence of Ca 2+ ionophore over that of untreated cells, while H-89 treatment attenuated the GFP level. In addition, LM-021 significantly activated CRE-mediated transcription in the presence or absence of Ca 2+ ionophore. The kinase inhibitors H-89, KN-62, and U0126 but not wortmannin attenuated the GFP levels in the presence or absence of Ca 2+ ionophore.
The p-CREB (S133) and GFP protein levels in CRE-GFP reporter cells without Ca 2+ ionophore addition were further examined by immunoblotting using specific antibodies. As shown in Figure 5(b), the increased p-CREB and GFP levels were also significantly reduced in CRE-GFP reporter cells after treatment with H-89, KN-62, or U0126 but not wortmannin. The results demonstrate that LM-021 stimulated cAMP-mediated transcription through the PKA, CaMKII, and ERK pathways.
3.6. Regulation of CREB Phosphorylation by LM-021 in Aβ-GFP SH-SY5Y Cells. CREB plays an important role in cell survival and synaptic activity by upregulating BCL2 [42] and BDNF [36], following phosphorylation at Ser133. CREB-regulated BDNF and BCL2 pathways are compromised in the hippocampus of APP transgenic mice, and the overexpression of CREB protects rat primary hippocampal neurons against Aβ-induced apoptosis [15]. To further investigate the CREB-mediated neuroprotective potential of LM-021, we applied the kinase inhibitor H-89, KN-62, or U0126 to LM-021-treated Aβ-GFP-expressing SH-SY5Y cells (Figure 6(a)). Protein expression levels of PKA, CaM-KII, ERK, CREB, BDNF, BCL2, and BAX were examined. As shown in Figure 6(b), while the level of CaMKII was not notably affected, overexpression of Aβ-GFP downregulated PKA and ERK. LM-021 treatment rescued the reduction in PKA and ERK, although not significantly. The overexpression of Aβ-GFP reduced p-PKA, p-CaMKII, and p-ERK protein levels, and treatment with LM-021 significantly upregulated p-PKA, p-CaMKII, and p-ERK, whereas H-89 treatment mitigated the increase of p-PKA, KN-62 treatment mitigated the upregulation of p-CaMKII, and KN-62 or U0126 treatment normalized p-ERK.

Pharmacokinetics of LM-021.
Pharmacokinetic data for LM-021 were obtained by noncompartmental analysis using Phoenix WinNonlin. The calculated PK parameters in mice are summarized in Table 1. The elimination half-life (t 1/2 ) of LM-021 in plasma and brain was 2:54 ± 0:79 and 2:17 ± 0:67 h, respectively, suggesting that after 5 half-lives, 97% of LM-021 in plasma/brain will be eliminated. The mean resident time (MRT) of LM-021 was 3:54 ± 0:93 h, indicating the average time that LM-021 molecules are present in the body before being eliminated. The plasma clearance of LM-021 was 1:85 ± 0:12 ml/min/kg, which is about 0.5% of the cardiac output blood flow of mice (400 ml/min/kg), indicating low clearance in mice. The total apparent volume of distribution at steady state (V ss ) was 0:39 ± 0:08 l/kg, which is higher than the total plasma volume (0.05 l/kg) of mice, suggesting that LM-021 can distribute into tissues. After intravenous bolus injection, the systemic exposure (AUC 0-∞ ) of LM-021 in plasma and brain was 45:28 ± 2:83 and 2:38 ± 0:03 μg h/ml, respectively. LM-021 had a brain to plasma ratio of 5.3% in mice.

Discussion
Effective treatments to slow AD neurodegeneration are still unavailable. Lines of evidence have suggested that the CREB signaling pathway is a potential therapeutic target for AD. The decreased levels of phosphorylated CaMKII, ERK, and CREB have been shown in the hippocampus of old rats with impaired spatial memory [43][44][45]. Oligomeric Aβ treatment significantly downregulates CREB phosphorylation [16]. Accumulation of tau may cause synapse and memory impairment by inactivating CREB signaling and CaMKIV activity [46]. Here, we showed that the novel synthetic coumarin-chalcone derivative LM-021 activates CREB signaling. LM-021 provides neuroprotective effects through increasing phosphorylation of PKA, CaMKII, and ERK to activate CREB and enhancing the expression of its On day 1, cells were plated with retinoic acid (RA, 10 μM) being added to the culture medium. On day 2, LM-021 (5 μM) was added to the cells for 8 h, followed by treatment of doxycycline (Dox, 5 μg/ml) to induce Aβ-GFP expression. Kinase inhibitors (10 μM) were added to the cells on day 6. On day 8, p-PKA, p-CaMKII, p-ERK, p-CREB, BDNF, BCL, and BAX levels, as well as neurite outgrowth (TUBB3 staining), were measured. (b) p-PKA, p-CaMKII, and p-ERK protein levels determined by immunoblot using GAPDH as a loading control (n = 3). To normalize, the protein expression level in untreated cells was set as 100%. (c) Measurements of neurite outgrowth (length, process, and branch) (n = 3). p values: induced vs. uninduced cells ( # p < 0:05, ## p < 0:01, and ### p < 0:001), LM-021-treated vs. untreated cells ( * p < 0:05, * * p < 0:01, and * * * p < 0:001), or kinase inhibitor-treated vs. untreated cells ( & p < 0:05, && p < 0:01, and &&& p < 0:001) (one-way ANOVA with a post hoc Tukey test). 12 Oxidative Medicine and Cellular Longevity downstream genes in Aβ-GFP-and ΔK280 tau RD -DsRedexpressing SH-SY5Y cells. When Aβ and tau models were compared, the less reduction changes of phosphorylated PKA, CaMKII, and ERK in tau cells may explain that the effects of LM-021 on tau cells seem to be less significant than on Aβ cells.
The activation of CREB and expression of its downstream genes are regulated by protein kinases under the control of cAMP [10] or Ca 2+ [47,48]. CREB phosphorylation and nuclear translocation are largely mediated by Ca 2+ influx, the mobilization of calmodulin to the nucleus, and activation of CaMKII and CaMKIV [49][50][51]. To facilitate 13 Oxidative Medicine and Cellular Longevity CREB phosphorylation in our CRE-GFP 293 reporter cells, we treated the reporter cells with a Ca 2+ ionophore to increase cytosolic Ca 2+ . We observed that CRE-motifdriven GFP fluorescence and CREB phosphorylation were augmented by the Ca 2+ ionophore (Figures 2 and 5). These results are similar to those of other studies in PC12 cells and auditory neurons [52,53]. The ability of LM-021 to activate the CREB signaling pathway was also demonstrated in these reporter cells. Of note, LM-021 also upregulated the expression of CREB in Aβ-GFP-and ΔK280 tau RD -DsRedexpressing SH-SY5Y cells ( Figures S6A and S7A), further providing its neuroprotective effects by modulating the CREB signaling pathway.
To further understand the mechanism underlying the effect of LM-021 on CREB phosphorylation, we treated CRE-GFP reporter cells with H-89, KN-62, U0126, and wortmannin, which are inhibitors of PKA, CaMKII, ERK, and PI3K. The LM-021-increased phosphorylation of CREB and CRE-motif-driven GFP fluorescence were attenuated by the inhibition of PKA, CaMKII, and ERK, but not PI3K ( Figure 5). In Aβ-GFP-and ΔK280 tau RD -DsRed-expressing SH-SY5Y cells, the reduced phosphorylation of PKA, CaM-KII, ERK, and CREB indicates that these signaling pathways were all compromised by the overexpression of Aβ or tau protein (Figures 6 and 7). LM-021 improved neurite outgrowth and consistently increased the phosphorylation of PKA, CaMKII, and ERK, all of which in turn phosphorylate CREB and drive the expression of BDNF and BCL2. The decrease in CREB phosphorylation by inhibitors of PKA, CaMKII, and ERK counteracted the neuroprotective effects of LM-021. These results indicate that LM-021, working as a CREB enhancer by regulating the PKA, CaMKII, and ERK pathways, displays neuroprotective effects against Aβ and tau aggregation.
PKA is the main regulator of most cAMP-dependent physiological processes. The association of cAMP with a specific site on the surface of PKA leads to the release of catalytic subunits that phosphorylate downstream substrates [54]. PKA can remodel neurite outgrowth in developing neurons. Decreased PKA activity by adenylyl cyclase inhibitors results in deficient neurite outgrowth, while enhanced PKA activity by forskolin or dibutyryl-cAMP promotes neurite outgrowth [55]. In our study, the reduction in PKA activity caused by the specific inhibitor H-89 significantly counteracted the improved neurite outgrowth induced by LM-021. This observation is consistent with the reported neuritogenic activity mediated by cAMP/PKA in human SH-SY5Y cells [56]. On the other hand, while a basal level of PKA activity is required to promote neurite outgrowth, a high level of PKA activity shortens and simplifies neurites [57]. Thus, a proper level of PKA activity is critical for promoting optimal neurite growth.
CaMKII is regulated by the Ca 2+ /calmodulin complex [58]. Postmortem brain analyses show that CaMKIIexpressing neurons are selectively lost in the hippocampus of severe AD patients [59,60]. The phosphorylation of CaMKII is also reduced in the hippocampus of patients with mild cognitive impairment or severe AD [61,62]. Upregulation of CaMKII phosphorylation by spatial training can improve spatial learning and memory in an APP transgenic mouse model of AD [63]. In Aβ-GFP-and ΔK280 tau RD -DsRed-expressing cells, we consistently observed that a decrease in CaMKII phosphorylation by KN-62 diminished the neuroprotective effects of LM-021, suggesting a key role for CaMKII in the mechanism underlying AD neurodegeneration.
ERK is highly expressed in the central nervous system, and the ERK pathway plays multiple roles in the activitydependent regulation of neuron function [64,65]. ERK activity controls the proliferation and differentiation of neurons and glia during brain development [66,67]. Consistent with our findings, the phosphorylation of ERK by α-lipoic acid has been reported to promote neurite outgrowth [68]. Evidence suggests that ERK activity is markedly increased in AD [69][70][71]. ERK is significantly activated in APP transgenic mice [72], and tau can be hyperphosphorylated by ERK [73]. ERK colocalises with neurofibrillary tangles in neurons and with tau in subcellular compartments [74]. More details regarding the function of ERK in pathological conditions may provide insight into the pathogenesis of AD.
CREB phosphorylation enhances the expression of CREmediated genes, including BCL2 and BDNF, which may provide neuroprotection against apoptosis [75,76]. BCL2 inhibits the BAX activation and prevents it from translocation to the mitochondria, thereby decreasing apoptosis; conversely, downregulated BCL2 promotes mitochondrial BAX homooligomerization, resulting in subsequent induction of the cytochrome c-mediated apoptotic cascade [77][78][79]. In AD, Aβ induces the downregulation of BCL2 and upregulation of BAX [80]. The ratio of BCL2 to BAX has been shown to correlate negatively with the level of tau phosphorylation [81]. As an important member of the classic neurotrophin family of growth factors, BDNF promotes neuron survival, differentiation, and plasticity [82]. In SH-SY5Y cells, treatment with oligomeric Aβ significantly causes downregulation of BDNF expression [83]. Tau overexpression or hyperphosphorylation reduces BDNF expression in primary neurons and tau transgenic mice [84,85]. We also consistently observed that overexpression of Aβ-GFP or ΔK280 tau RD -DsRed reduces BCL2 and BDNF and increases BAX The brain tissue density was assumed to be 1 g/ml. b Brain-plasma ratio was calculated by the AUC (0-∞) ratios. NA: not applicable; C 0 : concentration at time zero; t 1/2 : elimination half-life; MRT: mean residence time; AUC (0-∞) : the area under the concentration-time curve from time 0 extrapolated to infinity; CL: body clearance; V ss : an estimate of the volume of distribution at steady state.
14 Oxidative Medicine and Cellular Longevity expression. In this study, we show that LM-021 normalizes the expression of these CREB-responsive genes. LM-021 displays good oral bioavailability, BBB penetration capability (Figure 1), and low toxicity in SH-SY5Y cells (Figures 3 and 4). Our pharmacokinetic study showed that LM-021 has a low clearance and a brain to plasma ratio of 5.3% in mice, which allows LM-021 to distribute into tissues, including the brain. Although 5.3% is a modest brain to plasma ratio, some of the drugs used to treat neurodegenerative diseases have an even lower ratio. For example, without an enzyme inhibitor, only about 1% of orally administered levodopa enters the brain [86]. Therefore, LM-021 shows the potential to cross the BBB and may be a suitable candidate compound for treating AD. LM-021 also demonstrated significant inhibitory effects on aggregation and ROS reduction in both Aβ-GFP-and ΔK280 tau RD -DsRed-expressing SH-SY5Y cells. As Aβ and tau pathologies are likely synergistic, the observation that LM-021 targets both of these proteins suggests that it may be particularly suitable for use in pleiotropic treatments for AD. LM-021 exhibited greater inhibitory effects against Aβ and tau aggregation than did curcumin or Congo red, indicating its strong potential as a therapeutic agent. Furthermore, LM-021 displayed neuroprotective effects not only by rescuing neurite outgrowth deficits but also by reducing caspase 1 activity in Aβ-GFPand ΔK280 tau RD -DsRed-expressing SH-SY5Y cells. Given that tau accumulations are commonly seen in other neurodegenerative diseases, including frontotemporal dementia, progressive supranuclear palsy, Pick's disease, and corticobasal syndrome, we expect that LM-021 may be promising for treating these tauopathies as well.
To determine whether the coumarin-chalcone derivatives have chemical chaperone activity, we used the Thioflavin T assay to assess misfolding of aggregated synthetic Aβ 42 and E. coli-derived ΔK280 tau RD . The results show that the coumarin-chalcone derivative LM-021 directly hinders Aβ 42 and tau aggregation (Figure 1). Whether the upregulation of CREB phosphorylation and its upstream pathways PKA, CaMKII, and ERK occurs subsequent to aggregate inhibition or is directly enhanced by LM-021, or both, necessitates further studies.
Finally, preconditioning signal leading to cellular protection through an important redox reaction has the hormesis feature, which can ameliorate aging associated with free radical accumulation and inflammatory responses involved in neurodegenerative/neuroprotective mechanisms [87]. Several reports have shown the relationship between polyphenol compounds, redox status, and the vitagene network and its possible biological relevance in neuroprotection [88,89]. The tested novel compound LM-021 displays significant antioxidative and neuroprotection effects and probably implicates its role as an agent involved in the vitagene network. Although our study has shown the maximal effective dose of inhibiting aggregation and neuroprotection effects among 3 different doses, we have not yet determined if higher concentrations induce significant cell toxicity and display less neuroprotection effect. Further studies will be demanded to explore optimised doses of LM-021 in treating and preventing AD, based on the hormesis feature of antioxidant.

Conclusions
In the present study, we show that the CREB signaling pathway in Aβ-GFP-and ΔK280 tau RD -DsRed-expressing SH-SY5Y cells is compromised, which leads to decreased BCL2 and BDNF, elevated BAX and oxidative stress, and subsequent neurite outgrowth deficits. LM-021 exerts neuroprotective effects in these cells by regulating the PKA, CaMKII, and ERK pathways, all of which subsequently increase CREB phosphorylation and neurite outgrowth. Given that multiple pathways are involved in AD pathogenesis, the targeting of multiple pathways by LM-021 to provide neuroprotection may be particularly promising in the development of drugs for treating AD. The neuroprotective effects of LM-021 observed in the tau cell model also suggest this agent as a novel candidate for treating other neurodegenerative tauopathies. The demonstration that LM-021 has good oral availability and BBB penetration suggests its use for targeting neurons in AD patients by oral administration. Nevertheless, it is important to note that the effects of LM-021 were only observed in cell models. Future studies in animal models are necessary to confirm these results before investigating its therapeutic application to clinical trials.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Supplementary Materials
Preparation and analysis of mouse plasma and brain homogenate samples. Figure S1: oral bioavailability prediction and TEM examination of Aβ and tau aggregates. Figure S2: nucleotide sequence of synthetic CRE fused to TATA-like promoter and CRE fluorescence reporter assay. Figure S3: dose-response curves based on GFP fluorescence and ROS images in Aβ-GFP cells. Figure S4: dose-response curves based on DsRed fluorescence and ROS images in ΔK280 tau RD -DsRed cells. Figure S5: experimental flow chart to examine LM-021-mediated kinase activation. Figure S6: regulation of CREB signaling pathway and neurite outgrowth images in Aβ-GFP cells. Figure S7: regulation of CREB signaling pathway and neurite outgrowth images in ΔK280 tau RD -DsRed cells. (Supplementary Materials)