Materials. Compound TAU1 was purchased from SYNCOM (custom synthesis) and used without further purification. The chemical identity of compounds was assessed by re-running Nuclear Magnetic Resonance spectroscopy (NMR) experiments and proved to be in agreement with the literature data reported for this compound. The purity, checked by reversed-phase High Performance Liquid Chromatography (HPLC), was approximately 95%. When not specified reagents and solvents were purchased from commercial suppliers (MERCK Life Science, TCI Chemicals, Eurisotop) and were used without further purification. UV/vis spectra were recorded on a Jasco V-750 spectrophotometer and fluorescence spectra were obtained using a Shimadzu RF-6000 spectrofluorophotometer. Melting points were recorded with Büchi melting point B-545 apparatus in open capillaries and are not corrected. NMR spectra have been acquired with a Bruker Avance/Ultra ShieldTM 400 spectrometer operating at 400.13 MHz for 1H and 100.62 MHz for 13C at room temperature, using tetramethylsilane (TMS) as internal standard and 5 mm diam-eter glass tubes. Chemical shifts (d) are reported in parts per million (ppm) and coupling constants (J) in hertz (Hz), approximated to 0.1 Hz. The residual solvent peak was used as an internal reference for 1H and 13C NMR spectra and is referenced to CD2Cl2 (δ = 5.26 ppm for 1H, δ = 53.84 ppm for 13C). Data for 1H NMR are reported as follows: chemical shift, multiplicity (br = broad, ovrlp = over-lapped, s = singlet, d = doublet, t = triplet, q =quartet, m = multiple, dd = doublet of doublets), coupling constant, integral. All 13C NMR spectra were obtained with complete proton decoupling. Spectra were processed with the pro-gram MestReNova version 12.0.0-20080, FT and zero filling at 64K. Chromatography was carried out on 60 Å silica gel (40-63 µm, 230–400 mesh). All reactions were monitored by thin-layer chromatography (TLC), and 60 Å silica gel on TLC plates were used. The compounds on TLC were revealed by quenching fluorescence at 365 nm using a 4W UV lamp because the fluorescent markers of the invention are characterized by an excitation wavelength of 350 to 650 nm and an emission wavelength of 390 to 800 nm. Electron spray ionization mass spectra (ESI-MS) were per-formed on Bruker BioApex Fourier transform ioncyclotron resonance (FT-ICR) mass spectrometer.
Molecular modelling. Ligands were designed in 2D with PICTO version 18.104.22.168 (OpenEye Scientific Software, Santa Fe, NM)43 and converted into 3D format by OMEGA version 22.214.171.124 (OpenEye Scientific Software, Santa Fe, NM)44,45 The most prevalent ligand protonation form at pH 7.4 was assigned by QUACPAC version 126.96.36.199 (OpenEye Scientific Software, Santa Fe, NM)46, while energy minimization was carried out by SZYBKI version 188.8.131.52 (OpenEye Scientific Software, Santa Fe, NM) 44 using the MMFF94S force field47. The 6-mer model of the most conserved channel formed by four adjacent β-sheets was prepared as described in Verwilst et al, 31 starting from the PDB-ID 5K7N49. Molecular docking was carried out with AutoDock4.2 using default settings50. Since common molecular modeling software such as OpeEye and AutoDock do not provide force field parameters for boron atoms, in our study it was replaced by a carbon atom having sp3 hybridization.
Synthesis of 3. Trans-4-[2-(4-dimethylaminophenyl)vinyl]benzaldehyde was synthesized via Heck reaction starting from 4-bromobenzaldehyde (1) and 4-dimethylaminostyrene (2) according to a modified literature procedure11. The suitable catalyst, chosen to promote the stereoselectivity of the reaction, was prepared in situ: 16.8 mg of palladium acetate (II) (0.075 mmol) and 19.7 mg of triphenylphosphine (0.075 mmol) have been soluble in DMF. After 10 minutes, a solution of 202 mg of 4-bromobenzaldehyde 1 (1.5 mmol), 264.6 mg of 4-dimethylminostyrene 2 (1.8 mmol) and 414 mg of potassium carbonate (3.00 mmol) in 3 mL of DMF has been added to the catalyst solution. The reaction was left in agitation at 80°C for 4 h, after which 50 mL of a saturated NH4Cl solution in water was added. Later, the mixture was extracted with CH2Cl2 (3 × 100 mL) and the organic layers were combined, dried with Na2SO4 anhydrous and concentrated under vacuum. Compound 3 (1,074 mmol) was obtained by cold hexane crystallization as a yellow solid (270 mg, 72%). Mp: 218.0 – 220.0°C. 1H NMR (400 MHz, CH2Cl2) δ 9.94 (s, 1H), 7.82 (d, J = 8.3 Hz, 2H), 7.62 (d, J = 8.3 Hz, 2H), 7.45 (d, J = 8.8 Hz, 2H), 7.23 (d, J = 16.3 Hz, 1H), 6.96 (d, J = 16.3 Hz, 1H), 6.72 (d, J = 8.8 Hz, 2H), 3.00 (s, 6H). 13C NMR (101 MHz, CH2Cl2) δ 191.3, 150.8, 144.5, 134.6, 132.4, 130.1, 128.1, 126.1, 124.5, 122.4, 112.3, 79.5, 40.1. ESI-MS(m/z): [M+H]+ calcd. For C17H18NO, 252.13; found 252.17
Synthesis of BT1. A solution of 100 mg (0.45 mmol) of commercially available 1, 112.95 mg (0.45 mmol) of compound 3, in the presence of 0.35 mL of acetic acid (3.5 mmol) and 0.35 mL of piperidine (6.12 mmol) in 10 mL of toluene was heated under Dean−Stark conditions for 4 h. The reaction was allowed to cool to room temperature, and 50 mL of a saturated aqueous NH4Cl solution was added. The mixture was extracted with CH2Cl2 (3 × 100 mL), and the organic layers were combined, dried over Na2SO4 anhydrous and concentrated at reduced pressure. Column chromatography (silica, EtOAc/hexane, 1:9 → CH2Cl2) resulted in 18 mg (0.04 mmol,18%) of BT1 as black solid. Mp: 257-262°C. 1H NMR (400 MHz, CH2Cl2) δ 7.63-7.57 (m, 4H), 7.53 (d, J = 8.4 Hz, 2H), 7.46 – 7.42 (m, 3H), 7.23 (s, 1H), 7.16 (d, J = 16.4 Hz, 1H), 6.97 – 6.91(m, 2H), 6.83 (bs, 1H), 6.72 (d, J = 8.9 Hz, 2H), 6.48-6.47 (m, 1H), 2.99 (s, 6H), 2.34 (s, 3H). 13C NMR (101 MHz, CH2Cl2) δ 151.4, 149.8, 141.2, 141.0, 139.0, 134.9, 131.3, 129.3, 128.9, 128.7, 127.2, 126.4, 125.9, 123.9, 123.6, 121.3, 118.2, 118.1, 117.0, 113.1, 41.0, 12.1. ESI-MS(m/z): [M+H]+ calcd. For C28H27BF2N3, 453.22; found 454.33.
Human iPSCs Maintenance. Induced pluripotent stem cell line (WT#1)39,40 was maintained in mTeSR Plus medium (STEMCELL Technologies) on growth factor-reduced Matrigel-coated (Corning; dilution 1:100) plates at 37°C in 5% CO2 and iPSC colonies were passaged with PluriS-TEM Dispase-II (Merck Life Science) when 80% confluent.
iPSCs Differentiation to cortical neurons. iPSC WT#1 was differentiated with a two-step protocol based on doxycycline-induced expression of human NGN2 transcript as previously described. Briefly, human iPS cells were treated with 1X Accutase (Thermo Fisher Scientific) and plated onto growth factor-reduced Matrigel-coated plates at a density of 1000 cells/mm2 in mTeSR Plus containing 10 µM Rock-inhibitor Y-27632 (STEMCELL Technologies) and 2 µg/mL doxycycline (Merck Life Science). The day of seeding is set as day minus 3 (D-3). One day after seeding (D-2), the medium is switched to N2 medium consisting of DMEM/F12 [1:1], 1% N2 supplement, 1% NEAA, 1% GlutaMAX (Thermo Fisher Scientific) supplemented with 2 µg/mL doxycycline (Merck Life Science) to sustain hu-man NGN2 expression. N2 medium was refreshed every day. Three days after (D0), the early born neurons were dis-sociated with Accutase and plated onto PDL/laminin-coated (Merck Life Science) dishes at a density of 500 cells/mm2 in maturation medium consisting of Neurobasal, 2% B27 with vitamin A, 1% GlutaMAX (Thermo Fisher Scientific), 0,5 µg/mL laminin (Merck Life Science), 20 ng/mL BDNF (Peprotech), 20 ng/mL ascorbic acid (Peprotech), 10 ng/mL GDNF (Peprotech) supplemented with 2 µg/mL doxycycline, 10 µM Rock-inhibitor Y-27632 and 10 µM DAPT (STEMCELL Technologies). After 24 hours, doxycycline and Y-27632 were removed and the medium was refreshed every three days until day 10. Optionally, 2 µM Ara-C (Merck Life Science) was added to the medium on day 4 and day 7 to remove proliferative cells. Thereafter, the maturation medium was half changed weekly until the experimental window was reached around D30.
RT-PCR and RT-qPCR. Total RNA was extracted with the EZNA Total RNA Kit I (Omega Bio-Tek) and retrotranscribed using the iScript Reverse Transcription Supermix for RT-qPCR (Bio-Rad). Real-time RT-PCR was performed with iTaq Universal SYBR Green Supermix (Bio-Rad) on a ViiA 7 Real-Time PCR System (Applied Biosystems) and housekeeping gene ATP5O (ATP synthase, H+ transporting, mitochondrial F1 complex, O sub-unit) was used as an internal control. A complete list of primers is provided in the supplementary material (Supplementary Table 3).
Induction of a tau hyperphosphorylated state and staining with BODIPY-base probes. iPSC-derived cortical networks were treated with 50 nM Okadaic acid (OA; Merck Life Science) for 2 hours at 37°C in 5% CO2. Afterward, untreated and treated neuronal cultures were incubated with either 100 µM TAU1 probe or 100 µM BT1 probe for 30 minutes at 37°C and thus fixed for 15 minutes at room temperature with cool and fresh-made 4% PFA.
Immunostaining. Fixed iPSC-derived cortical neurons were permeabilized with 0.2% Triton X-100 (Merck Life Science) in 1X TBS and incubated for 1 hour in blocking solution containing 1X TBS, 0.2% Triton X-100, and 5% goat serum (Merck Life Science). Afterwards, the cells were incubated in blocking solution containing primary antibodies overnight at 4°C. The primary antibodies employed in this study were anti PHF-tau Ser202/Thr205 (AT8; dilu-tion 1:200; Thermo Fisher Scientific) and anti-oligomeric tau (T22; dilution 1:200; Merck Life Science) followed by incubation with secondary antibody (dilution 1:1000) for 1 hour at room temperature. A complete list of antibodies is provided in the supplementary material (Supplementary Table 2).
Images were acquired with an inverted microscope equipped with the X-Light V2 Spinning Disk Confocal module (Crest Optics) with a 40×/NA 0.75 objective lens in stack with z-step of 0.4 µm. Image processing was per-formed through custom numeric codes implemented in MATLAB environment.
Image analysis. Image processing was performed through custom numeric codes implemented in MATLAB environment. The algorithm consists of three main steps: image correction, image binarization and colocalization.
Correction (Figure 6A). For each channel, the maximum z-projection obtained from the z-stack was examined. Noise and background affecting each channel were removed before carrying out the quantitative analysis. A moving average filter (5x5 pixels window) was used to reduce the noise, the background affecting each channel was automatically retrieved and subtracted through the statistical analysis of global pixel intensity distribution (negative values have been set to zero). More precisely, a histogram shape-based method was used, which identifies background levels beyond the peaks of the smoothed histograms. Furthermore, to exclude unwanted signals, a neurite structure mask, obtained by thresholding pixel values on the AT8 channel, was applied to each channel.
Binarization (Figure 6B) Meaningful pixels were selected for each corrected image. For this purpose, removing background and thresholding, unfortunately, was not sufficient: low threshold values lead to an overestimation of the signal, and higher values could cause information loss. To overcome this problem, the following iterative procedure was designed. Starting from a low threshold value, 20% of the maximum signal, at each step the threshold was increased by 15%, until enough pixels were no longer selected. A different binary image was obtained for each threshold: the final binary image was obtained by combining them and deleting the largest of the overlapping areas (more than 70% overlapping). Finally, the area covered by the fluorescence signal and the corresponding integrated density were calculated for each channel.
Colocalization (Figure 6C) To quantify the colocalization of the probe with the anti-bodies T22 and AT8 the following procedure was followed. Two of the most used colocalization coefficients were exploited: Pearson’s correlation coefficient (PC) and Manders’ overlap coefficient (M1), which measure correlation and co-occurrence respectively. An easy way to visualize the dependence of pixels in dual-channel images is to consider a pixel distribution diagram called scatter plot or fluorogram (orange and green plots in Figure 6C), where x-coordinates are given by the pixel intensities of the probe image and y-coordinates are given by the pixel intensities of the antibody image. This scatter plot provides the first intuitive evidence of colocalization: in a complete colocalization the points on the diagram are distributed around a line (e.g. green plot in Figure 5C) and the spread with respect to the line is measured by the correlation coefficient PC. More precisely, PC quantifies pixel intensity spatial correlation and, being calculated on corrected not binary images, it is independent of the binarization step. However, in a more general scenario (e.g. orange plot in Figure 5C), in addition to scatter plots and PC it is necessary to evaluate colocalization by using the corrected binary images and calculating the coefficient M1. M1 is defined as the ratio of the ‘summed intensities of pixels from the antibody image for which the intensity in the probe channel is above zero’ to the ‘total intensity in the antibody channel’. Hence M1 measures the fraction of antibody signal coinciding with probe signal, an ideal tool to quantify probes efficiency. PC and M1 can be considered reliable colocalization coefficients since images were properly corrected, similar acquisition and thresholding conditions were applied, and a large set of images was compared.
Statistical data analysis. Statistical analysis, graphs and plots were generated using GraphPad Prism 6 (GraphPad Software) and MATLAB 2016b (MathWorks). To verify whether our data sets were reflecting normal distribution, the Shapiro-Wilk normality test was performed. Where the normality distribution was not fulfilled, statistical significance analysis was performed using the non-parametric two-sided Mann–Whitney test (MW test, P=0.05). In all other cases, whether not stated otherwise, t-Student test (P=0.05) was performed, and data set are given as mean ± standard error of the mean (s.e.m.).