Using ΔK280 TauRD Folding Reporter Cells to Screen TRKB Agonists as Alzheimer’s Disease Treatment Strategy

Misfolded aggregation of the hyperphosphorylated microtubule binding protein Tau in the brain is a pathological hallmark of Alzheimer’s disease (AD). Tau aggregation downregulates brain-derived neurotrophic factor (BDNF)/tropomycin receptor kinase B (TRKB) signaling and leads to neurotoxicity. Therefore, enhancement of BDNF/TRKB signaling could be a strategy to alleviate Tau neurotoxicity. In this study, eight compounds were evaluated for the potential of inhibiting Tau misfolding in human neuroblastoma SH-SY5Y cells expressing the pro-aggregator Tau folding reporter (ΔK280 TauRD-DsRed). Among them, coumarin derivative ZN-015 and quinoline derivatives VB-030 and VB-037 displayed chemical chaperone activity to reduce ΔK280 TauRD aggregation and promote neurite outgrowth. Studies of TRKB signaling revealed that ZN-015, VB-030 and VB-037 treatments significantly increased phosphorylation of TRKB and downstream Ca2+/calmodulin-dependent protein kinase II (CaMKII), extracellular signal-regulated kinase 1/2 (ERK) and AKT serine/threonine kinase (AKT), to activate ribosomal S6 kinase (RSK) and cAMP response element-binding protein (CREB). Subsequently, p-CREB enhanced the transcription of pro-survival BDNF and BCL2 apoptosis regulator (BCL2), accompanied with reduced expression of anti-survival BCL2-associated X protein (BAX) in ΔK280 TauRD-DsRed-expressing cells. The neurite outgrowth promotion effect of ZN-015, VB-030 and VB-037 was counteracted by a RNA interference-mediated knockdown of TRKB, suggesting the role of these compounds acting as TRKB agonists. Tryptophan fluorescence quenching analysis showed that ZN-015, VB-030 and VB-037 interacted directly with a Pichia pastoris-expressed TRKB extracellular domain, indirectly supporting the role through TRKB signaling. The results of up-regulation in TRKB signaling open up the therapeutic potentials of ZN-015, VB-030 and VB-037 for AD.


Introduction
A number of neurodegenerative diseases, including Alzheimer's disease (AD), progressive supranuclear palsy, corticobasal syndrome and frontotemporal dementia, are characterized by accumulation of β-sheet-rich misfolded Tau proteins [1]. Encoded by MAPT, Tau participates in microtubule dynamics and assembly, and also plays a role in axonal transport [2][3][4]. Various genetic mutations have been attributed to inducing Tau protein misfolding. For example, deletion of a highly conserved lysine 280 (∆K280) in the conserved 18-amino acid repeat domain of Tau (Tau RD ) was found in patients with tauopathies [5,6]. Overexpression of ∆K280 Tau RD in HEK293T cells increased the formation of misfolded aggregates, suggesting its potential to accelerate Tau misfolding [7]. Of note, the misfolded Tau aggregation down-regulates brain-derived neurotrophic factor (BDNF) signaling pathways [8,9], implying the potential involvement of BDNF in the pathogenesis of Tau-mediated neurodegeneration. BDNF is a neurotrophic factor involving neuronal growth and survival [10]. Postmortem studies demonstrate reduced expression of BDNF in hippocampus, cortex and basal nucleus of Meynert in patients with AD, even in the pre-clinical stages [11,12]. The binding of BDNF to tropomyosin-related kinase B (TRKB) induces autophosphorylation, and then activates phosphoinositide 3-kinase (PI3K)-AKT serine/threonine kinase (AKT), extracellular signal-regulated kinase (ERK) and phospholipase-C-γ (PLC-γ) pathways [13]. The PI3K-AKT pathway suppresses cell apoptosis by reducing the translocation of apoptotic B-cell lymphoma 2 (BCL2)-associated X protein (BAX) from the cytoplasm to mitochondria [14]. The 90-kDa ribosomal S6 kinase (RSK) involved in cell survival is a downstream effector of ERK and both ERK and RSK phosphorylate the transcription factor cAMP responsive element-binding protein 1 (CREB) [15]. Activation of CREB promotes transcription of the BDNF neurotrophin [16] and BCL2 anti-apoptosis regulator [17], to regulate neuronal growth and differentiation, synaptic plasticity, spatial memory and long-term memory formation, as well as to ensure neuronal survival [18]. The PLC-γ pathway also mediates CREB phosphorylation through Ca 2+ influx in cortical neurons [19], and decreased phosphorylation of Ca 2+ /calmodulin-dependent protein kinase II (CaMKII) has been observed in SH-SY5Y cells expressing pro-aggregator Tau [20]. In AD animal models, overexpression of BDNF exerts the potential to rescue deficits in learning and memory [21,22].

Biochemical Fluorescence-Based Tau RD Aggregation Assay
The anti-aggregation activity of congo red (positive control) and ZN/VB compounds (5-20 µM) was assessed using E. coli-expressed ∆K280 Tau RD protein [33] and thioflavin T (Sigma-Aldrich) fluorescence assay. The trend of ∆K280 Tau RD amyloid aggregate formation depicted with enhanced and red-shifted fluorescence was recorded by a FLx800 microplate fluorescence reader (BioTek Instruments), where the procedure was previously reported in detail [34].

Real-Time PCR Analysis
The cDNA was synthesized (SuperScript™ III reverse transcriptase; Invitrogen) using total RNA extracted by TRI Reagent™ (Sigma-Aldrich). The real-time quantitative PCR was conducted with primers that were specific to DsRed and hypoxanthine phosphoribosyltransferase 1 (HPRT1, an endogenous control) as described [7] on the StepOnePlus™ Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The fold change of the expressed Tau-DsRed mRNA was calculated using formula 2 ∆Ct , as ∆C T = C T (HPRT1) − C T (DsRed), where C T indicates the cycle threshold.

Caspase-1/6 and Acetylcholinesterase (AChE) Activity Assays
After being compounds-treated and RA-differentiated, cells were collected for caspase-1/6 and AChE activity measurements. Cells were lysed in lysis buffer with repeated freeze/thaw cycles and protein concentration in cell lysates quantitated. Types of substrates applied to detect each caspase's activity were: YVAD-AFC for caspase-1 and VEID-AFC for caspase-6 (BioVision, Milpitas, CA, USA). The reaction mixtures were incubated for 1.5 h at 37 • C according to the manufacturer's instructions. The fluorescence intensities with excitation/emission wavelengths at 400/505 nm were recorded by a FLx800 microplate fluorescence reader (BioTek Instruments). An AChE activity assay was performed according to the manufacturer's instructions (Sigma-Aldrich), as described in the Ref. [28]. The mixture was incubated for 2-10 min at room temperature, and absorbance at 412 nm wavelength was measured by Multiskan™ GO Microplate Spectrophotometer (Thermo Scientific).

Western Blot Analysis
Cells were collected and protein extracted as described [34]. The extracted protein (20 µg) was mixed with a 6X SDS sample buffer and boiled for 5 min. The denatured protein samples were then separated by 10-12% SDS-polyacrylamide gel and transferred to a polyvinylidene difluoride (PVDF) membrane (Pall Corporation, Port Washington, NY, USA). The blotted membrane was blocked with 3% bovine serum albumin in Tris-buffered saline with 0. The secondary antibodies used were horseradish peroxidase-conjugated goat anti-mouse (#GTX213111-01) or goat anti-rabbit (#GTX213110-01) IgG (1:5000; GeneTex, Irvine, CA, USA). Protein signals were detected with chemiluminescent substrate (Millipore) using the ImageQuant™ LAS 4000 imager and analysis software (GE Healthcare, UK).

RNA Interference
Knockdown of TRKB was performed using TRKB-specific lentiviral short hairpin RNA (shRNA) TRCN0000002243, TRCN0000002245 and TRCN0000002246, and a scrambled negative control (TRC2.Void) (RNAi core facility of Academia Sinica, Taipei, Taiwan), as previously described [34]. At day 2 after cell seeding, multiplicity of infection of 3 for each shRNA was used to infect cells with the existence of polybrene (8 µg/mL; Sigma-Aldrich). The cultured media were changed at day 3, following by 8 h compound (10 µM) pre-treatment and doxycycline (2 µg/mL) induction of ∆K280 Tau RD -DsRed expression. Cells were collected at day 9 for TRKB protein analysis or imaged for neurite outgrowth analysis as described above.

Tryptophan Fluorescence Quenching Assay
The interaction between the N-terminal extracellular domain of TRKB (TRKB-ECD) and 7,8-DHF (as positive control; [35]), ZN-015, VB-030 or VB-037 was evaluated with tryptophan fluorescence titration. TRKB-ECD-His protein was expressed in the Pichia expression system (Invitrogen) and purified as stated [8]. Intrinsic tryptophan fluorescence spectra with titration of a different concentration of compounds (0-1000 nM) were recorded by fluorescence spectrophotometer F-7000 (Hitachi, Tokyo, Japan), with excitation wavelength 295 nm and emission wavelength in the range of 300-400 nm. The calculation of dissociation constant (K D ) between the 7,8-DHF or ZN/VB compound and TRKB was described in a previous report [8].

Statistical Analysis
Data are presented as mean ± standard deviation. Statistical analysis of data was evaluated using the two-tailed Student's t-test or one-way analysis of variance (ANOVA) with Tukey's post hoc adjustment. Significance was considered as a p-value < 0.05.

Tryptophan Fluorescence Quenching Assay
The interaction between the N-terminal extracellular domain of TRKB (TRKB-ECD) and 7,8-DHF (as positive control; [35]), ZN-015, VB-030 or VB-037 was evaluated with tryptophan fluorescence titration. TRKB-ECD-His protein was expressed in the Pichia expression system (Invitrogen) and purified as stated [8]. Intrinsic tryptophan fluorescence spectra with titration of a different concentration of compounds (0-1000 nM) were recorded by fluorescence spectrophotometer F-7000 (Hitachi, Tokyo, Japan), with excitation wavelength 295 nm and emission wavelength in the range of 300-400 nm. The calculation of dissociation constant (KD) between the 7,8-DHF or ZN/VB compound and TRKB was described in a previous report [8].

Statistical Analysis
Data are presented as mean ± standard deviation. Statistical analysis of data was evaluated using the two-tailed Student's t-test or one-way analysis of variance (ANOVA) with Tukey's post hoc adjustment. Significance was considered as a p-value < 0.05.

Chemical Chaperone and Antioxidant Activities of ZN/VB Compounds
Next, we examined the effects of congo red and ZN/VB compounds on Tau misfolding using His-tagged ∆K280 Tau RD proteins purified from E. coli cells [33] in thioflavin T spectroscopic assay. As shown in Figure 3A, after 48 h incubation at 37 • C, markedly increased thioflavin T fluorescence was observed with ∆K280 Tau RD protein (from 193 to 12,215 arbitrary unit (AU); p < 0.001), and the increase was significant compared to wild-type Tau RD protein (12,215 versus 3202 AU; p < 0.001). ∆K280 Tau RD aggregation was significantly reduced by congo red (20 µM

Neuroprotective Effects of ZN/VB Compounds
The linkage between pathologically modified Tau and Tau aggregates in neurodegeneration has been widely addressed [36]. The activity of caspase-6 protease, associated with axonal degeneration [37], is positively regulated by caspase-1 [38]. Therefore, we assessed the neuroprotective effects of the tested ZN/VB compounds by examining neurite outgrowth and caspase-1/6 activity using congo red as a positive control. As did not elevate AChE activity, treatment with ZN-015 and VB-030 significantly suppressed endogenous AChE activity (from 90% to 66-59%, p = 0.040-0.007). Taken together, the results of neurite outgrowth and caspase 1 activity revealed the neuroprotective potentials of ZN-015, VB-030 and VB-037. High-content assessment of neurite outgrowth (TUBB3 stain, green) (n = 3). The relative length, process, and branch of neurite of uninduced cells was normalized (100%). Nuclei were counterstained with DAPI (blue). The multi-colored mask was applied on segmented fluorescence images to assign

Targets of ZN/VB Compounds on TRKB Pathway
We next examined the therapeutic targets in TRKB signaling for neuroprotection of ZN/VB compounds. LM-031, a coumarin derivative displaying neuroprotective potential by up-regulating CREB-dependent BDNF/BCL2 pathway in pro-aggregatory Tau SH-SY5Y cells [33], was included for comparison. As shown in Figure 5, auto-phosphorylation of TRKB at residues Y516 and Y817 decreased upon the induction of ∆K280 Tau

Binding Affinity of ZN-015, VB-030 and VB-037 with TRKB-ECD
As ZN-015, VB-030 and VB-037 increased TRKB Y516/Y817 auto-phosphorylation, we examined their binding affinity to TRKB. The 398-amino-acid extracellular domain of TRKB (TRKB-ECD, Figure 7A) was expressed in yeast Pichia pastoris [8] and used to assess the binding of tested compounds to TRKB-ECD by tryptophan fluorescence assay. A reported TRKB receptor agonist 7,8-DHF [40] was included for comparison. Upon the binding of ligand, the changed TRKB-ECD conformation affects tryptophan microenvironment to result in quench of intrinsic tryptophan fluorescence [35]. The tryptophan fluorescence changes of TRKB-ECD in the presence of test compounds (1-1000 nM) were recorded after excitation at 295 nm at room temperature. Similar to 7,8-DHF, ZN-015, VB-030 and VB-037 displayed decreasing fluorescence in a concentration-dependent manner, and the fluorescence quenching was maximized at 1000 nM of test compounds ( Figure 7B). Based on the quantitative analysis of fluorescence change, binding affinity KD of each compound was calculated ( Figure 7B). As a positive control, TRKB-ECD binding affinity of 7,8-DHF was 13.4 ± 8.4 nM, which was closed to the reported value (12.1 ± 1.6 nM, [35]). The observed KD of 12.2 ± 9.4 nM (ZN-015), 11.0 ± 2.4 nM (VB-030) and 4.3 ± 5.2 nM (VB-037) demonstrated the high binding affinities of selected ZN/VB compounds to the extracellular domain of TRKB receptor. The unexpected raised-fluorescence spectra of VB-030 (500-1000 nM) in the range from 360 to 400 nm at 295 nm excitation was due to the fluorescence self-absorption of the compound itself ( Figure 7C).

Binding Affinity of ZN-015, VB-030 and VB-037 with TRKB-ECD
As ZN-015, VB-030 and VB-037 increased TRKB Y516/Y817 auto-phosphorylation, we examined their binding affinity to TRKB. The 398-amino-acid extracellular domain of TRKB (TRKB-ECD, Figure 7A) was expressed in yeast Pichia pastoris [8] and used to assess the binding of tested compounds to TRKB-ECD by tryptophan fluorescence assay. A reported TRKB receptor agonist 7,8-DHF [40] was included for comparison. Upon the binding of ligand, the changed TRKB-ECD conformation affects tryptophan microenvironment to result in quench of intrinsic tryptophan fluorescence [35]. The tryptophan fluorescence changes of TRKB-ECD in the presence of test compounds (1-1000 nM) were recorded after excitation at 295 nm at room temperature. Similar to 7,8-DHF, ZN-015, VB-030 and VB-037 displayed decreasing fluorescence in a concentration-dependent manner, and the fluorescence quenching was maximized at 1000 nM of test compounds ( Figure 7B). Based on the quantitative analysis of fluorescence change, binding affinity K D of each compound was calculated ( Figure 7B). As a positive control, TRKB-ECD binding affinity of 7,8-DHF was 13.4 ± 8.4 nM, which was closed to the reported value (12.1 ± 1.6 nM, [35]). The observed K D of 12.2 ± 9.4 nM (ZN-015), 11.0 ± 2.4 nM (VB-030) and 4.3 ± 5.2 nM (VB-037) demonstrated the high binding affinities of selected ZN/VB compounds to the extracellular domain of TRKB receptor. The unexpected raised-fluorescence spectra of VB-030 (500-1000 nM) in the range from 360 to 400 nm at 295 nm excitation was due to the fluorescence self-absorption of the compound itself ( Figure 7C).

Discussion
Given that aberrant aggregation of the Tau protein is one of the pathological hallmarks of AD, treatments targeting tauopathy-induced neurodegeneration may be also beneficial to AD. ∆K280 mutation in the Tau gene influences Tau mRNA splicing and results in a pro-aggregator Tau protein [41]. Similarly, our previous and present studies have shown that ∆K280 mutation induces Tau aggregation, oxidative stress, cytotoxicity, and impaired TRKB signaling [8,34]. The low cytotoxicity of the tested heterocyclic compounds ZN-006 (flavone), -013 (benzofuran), -014 and -015 (coumarin), and VB-030, -037 and -041 (quinoline) suggests their potential as candidate therapeutic compounds for neurodegenerative diseases. We used the established ∆K280 Tau mutation model to show that ZN-015, VB-030 and VB-037 significantly increased DsRed fluorescence, indicating their ability to reduce Tau aggregation. ZN-015, VB-030 and VB-037 reduced Tau fluorescence in the thioflavin T binding assay, indicating their chemical chaperon activity. ZN-015 and VB-030 further decreased ROS in ∆K280 Tau RD -DsRed-expressing SH-SY5Y cells. Only ZN-015 displayed an oxygen radical absorbance capacity in biochemical antioxidant assays. Furthermore, ZN-015, VB-030 and VB-037 promoted neurite outgrowth and reduced caspase-1 activity, ZN-015 decreased caspase-6 activity, and ZN-015 and VB-030 lowered AChE activity in ∆K280 Tau RD -DsRed-expressing SH-SY5Y cells. Taken together, these results indicate that ZN-015 and VB-030 possess aggregation-inhibitory and anti-oxidative features to provide protection against tauopathy-mediated cytotoxicity. Although not reducing ROS in ∆K280 Tau RD -DsRed-expressing SH-SY5Y cells, the antioxidant effect of VB-037 has been demonstrated in cell models for Parkinson's disease [42]. This neuroprotective effect may be mediated by hormetic responses through the activation of nuclear factor erythroidderived 2-like 2 antioxidant response elements or heat shock protein-mediated signaling pathways [43][44][45]. Further research is needed to explore the relationship between these signaling pathways and our candidate compounds.
A previous study has shown that over-expression of the wild-type or mutant Tau protein down-regulates BDNF expression in cellular and animal models of AD [9], which was also demonstrated by the present study. Impaired TRKB signaling contributes to the neurodegeneration of tauopathy or AD. Indeed, TRKB reduction exaggerates cognitive impairments and signal dysfunctions in 5XFAD mice [46]. In accordance with the previous report, our study shows that mutant ∆K280 Tau impairs TRKB signaling and neurite outgrowth. Compounds ZN-015, VB-030 and VB-037 activate p-TRKB, p-AKT, p-CaMKII, p-ERK, p-RSK, and p-CREB, elevate BDNF and BCL2, and decrease BAX and CASP3 expression in ∆K280 Tau RD -DsRed-expressing SH-SY5Y cells to rescue neurite outgrowth deficits. These results suggest these compounds may act as TRKB agonists to achieve neuroprotection effects. The neurite outgrowth promotion effects are attenuated by the knockdown of TRKB, which provides evidence that these compounds exert their beneficial effects through activating the TRKB signaling pathway. Furthermore, the tryptophan fluorescence assay demonstrates high binding affinities of ZN-015, VB-030 and VB-037 to TRKB-ECD, which are compatible with that of the reported TRKB agonist, 7,8-DHF [26,27].
In this study, we have shown that all of these pathological changes induced by mutant ∆K280 Tau can be rescued by ZN-015, VB-030 and VB-037 via activating TRKB and its downstream signaling, including PI3K/AKT, ERK/CREB, ERK/RSK, PLC-γ/CaMKII, PLC-γ/CREB, and CREB/BDNF pathways [13]. Indeed, these pathways play an important role in neuronal survival, and impairment of them causes AD pathology and cognitive decline. The PI3K/AKT pathway plays a crucial role in regulating cell survival, proliferation, differentiation, intracellular trafficking, and neurite outgrowth [47]. It has been shown that Aβ oligomers attenuate PI3K/AKT signaling significantly and activation of the PI3K/AKT pathway may promote neuronal survival [48]. A previous study has shown that Tau overexpression induced by extrasynaptic N-methyl-D-aspartate (NMDA) receptor activation causes neuronal death through suppressing survival signaling ERK phosphorylation [49]. Furthermore, inhibition of the compensatory increases of the BDNF-ERK-CREB pathway and exacerbates cognitive impairment in vascular dementia associated with obesity [50].
Deficiency of CREB signaling underlies cognitive deficits in aging and also contributes to neurodegeneration in AD, and activation of CREB may be a potential therapeutic strategy in dementia [18,51,52]. PLC-γ/CaMKII promotes intracellular Ca 2+ release and activates long-term potentiation and synaptic plasticity [53]. In addition, Ca 2+ further activates CaMKII to phosphorylate CREB and subsequently induces BDNF transcription [54]. RSK is an effector of ERK, and ERK and RSK cooperate in regulation of several proteins, including Fos proto-oncogene, AP-1 transcription factor subunit (FOS), GSK-3β, eukaryotic elongation factor 2 (eEF2) kinase, tuberin and eukaryotic translation initiation factor 4B (eIF4B), to stimulate cell survival [15]. Taken together with our study findings, activating TRKB downstream pathways provides neuroprotection in tauopathy or AD models. Although we have shown that compounds ZN-015, VB-030 and VB-037 protect differentiated SH-SY5Y cells via ameliorating Tau aggregation-related disturbance in the TRKB signal pathway, we did not provide direct evidence of activating TRKB by the tested compounds. Future studies showing direct binding of these compounds to the TRKB are necessary.
Although over-expression of BDNF improves cognitive deficits in AD animal models [21,22], bioavailability and blood-brain barrier permeability of BDNF is poor, and it is invasive and not practical to inject the BDNF-expressing lentivirus into the human brain. Therefore, compounds working as TRKB agonists have much more potential in terms of clinical treatments. Kwon et al. found that plant Zizyphus jujuba var. spinosa works as a TRKB agonist to attenuate Aβ-induced synaptic long-term potentiation deficits [55]. In addition to 7,8-DHF, Chen et al. also showed that a synthetic derivative CF 3 CN binds with TRKB, activates TRKB signaling, reduces AD pathologies, and ameliorates cognitive dysfunctions in 5XFAD mice [56]. Recently, Wang et al. developed a TRKB agnost antibody that can penetrate into the brain, prevent Aβ-induced cell death, and rescue memory deficits in the APP/PS1 mouse model [57]. We have previously shown that quercetin, apigenin and coumarin derivatives LMDS-1 and -2 activate TRKB to reduce Tau aggregation and protect cells against Tau-induced neurotoxicity [8,34]. In this study we have shown that coumarin derivative ZN-015 and quinoline compounds VB-030 and VB-037 also act as TRKB agonists to protect SH-SY5Y cells from cytotoxicity induced by mutant ∆K280 Tau. However, these results should be confirmed by future animal studies.

Conclusions
In conclusion, we demonstrated that ZN-015, VB-030 and VB-037 have potentials in anti-aggregation and promoting neuronal survival in pro-aggregator Tau-expressing SH-SY5Y cells. In addition, our results revealed that ZN-015, VB-030 and VB-037 can bind to TRKB-ECD to activate TRKB signaling through AKT, CaMKII, ERK and RSK kinases, all of which promote neurite outgrowth. These multi-functional molecules provide more diversity in pursuance of finding an efficient therapeutic strategy for neurodegenerative diseases such as AD.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.