Increased Neuronal Activity in Motor Cortex Reveals Prominent Calcium Dyshomeostasis in Tauopathy Mice

Abstract

Perturbed neuronal Ca²⁺ homeostasis is implicated in Alzheimer’s disease, which has primarily been demonstrated in mice with amyloid-β deposits but to a lesser and more variable extent in tauopathy models.

In this study, we injected AAV to express Ca²⁺ indicator in layer II/III motor cortex neurons and measured neuronal Ca²⁺ activity by two photon imaging in awake transgenic JNPL3 tauopathy and wild-type mice. Various biochemical measurements were conducted in postmortem mouse brains for mechanistic insight and a group of animals received two intravenous injections of a tau monoclonal antibody spaced by four days to test whether the Ca²⁺ dyshomeostasis was related to pathological tau protein.

Under running conditions, we found abnormal neuronal Ca²⁺ activity in tauopathy mice compared to age-matched wild-type mice with higher frequency of Ca²⁺ transients, lower amplitude of peak Ca²⁺ transients and lower total Ca²⁺ activity in layer II/III motor cortex neurons. While at resting conditions, only Ca²⁺ frequency was increased. Brain levels of soluble pathological tau correlated better than insoluble tau levels with the degree of Ca²⁺ dysfunction in tauopathy mice. Furthermore, tau monoclonal antibody 4E6 partially rescued Ca²⁺ activity abnormalities in tauopathy mice after two intravenous injections and decreased soluble pathological tau protein within the brain. This correlation and antibody effects strongly suggest that the neuronal Ca²⁺ dyshomeostasis is causally linked to pathological tau protein.

These findings also reveal more pronounced neuronal Ca²⁺ dysregulation in tauopathy mice than previously reported by two-photon imaging that can be partially corrected with an acute tau antibody treatment.

Introduction

Alzheimer’s disease (AD) is characterized by depositions of amyloid-β peptide (Aβ) and tau protein in various brain regions and associated loss of synapses and neurons, resulting in cognitive impairments. The degree of tau pathology correlates better with cognitive deficits than Aβ lesions (Nelson et al., 2012), suggesting that when clinical symptoms have manifested, targeting tau is likely to be more effective for therapy than clearing Aβ.

Perturbed neuronal Ca²⁺ homeostasis has for many years been implicated in AD (Alzheimer’s Association Calcium Hypothesis, 2017; Bezprozvanny and Mattson, 2008), and increased Ca²⁺ influx has been shown to have a close link to tau pathology (Furukawa et al., 2003; Mattson, 1990; McKee et al., 1990; Nixon, 2003), but this connection has not been well examined in animal models. Advances in imaging techniques now allow detection of intracellular signal in live animals by two-photon imaging. A few prior studies have examined Ca²⁺ signaling in tauopathy mouse models (Busche et al., 2019; Decker et al., 2015; Kopeikina et al., 2013; Kuchibhotla et al., 2014; Marinkovic et al., 2019; Overk et al., 2015). The results are quite variable, reporting no changes in Ca²⁺ homeostasis (Kopeikina et al., 2013; Kuchibhotla et al., 2014), Ca²⁺ dysregulation (Decker et al., 2015), reduction in Ca²⁺ transient frequency (Busche et al., 2019; Marinkovic et al., 2019), and decreased Ca²⁺ activity (Overk et al., 2015). These variable findings may be explained by differences in experimental design, such as the model, age, brain region, Ca²⁺ indicators and parameters, and if the mice are under awake or anesthetized conditions

Hence, we decided to focus on awake tauopathy animals and analyze their neuronal Ca²⁺ activity profile with two-photon imaging under resting and running conditions, compared to age-matched wild-type mice of the same strain background. By performing Ca²⁺ imaging of layer II/III pyramidal neurons in the motor cortex, our studies reveal different and more pronounced neuronal Ca²⁺dysregulation than previously reported by two-photon imaging in tauopathy mice. We also show that Ca²⁺dyshomeostasis relates better to soluble than insoluble levels of pathological tau protein, and that an acute tau antibody treatment partially corrects the Ca²⁺ abnormalities and reduces soluble tau protein.

Materials and methods

Animals

The JNPL3 mouse line (10–12 month old, homozygous females, human 0N4R with P301L mutation, Taconic, (Lewis et al., 2000)) and age-matched wild-type mice of the same strain background were used in this study. We used females because in this model they have more consistent pathology (unpublished observations), and thereby increase the likelihood of detecting significant differences between tauopathy mice and wild-type controls. Eight wild-type and eight JNPL3 mice were enrolled for their comparison (Figure 2, 3, 4, Suppl Figure 1, 2), except in the synaptosome analysis (n = 6 per group, Suppl Figure 6, 7). Seven IgG treated JNPL3 mice and seven 4E6 treated JNPL3 mice were enrolled for their comparison (Figure 6, 7, 8, 9, Suppl Figure 3, 4, 5), except in the synaptosome analysis (n = 6 IgG and n=7 4E6, Suppl Figure 8). All animals were housed at NYU School of Medicine animal facilities and cared for by the veterinary staff in AAALAC-approved facilities. All the procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the university, and are in accordance with NIH Guidelines, which meet or exceed the ARRIVE guidelines.

Figure 2. Neuronal Ca²⁺ activity abnormalities in tauopathy mice under resting condition.

A-B. Representative Ca²⁺ traces from a WT mouse (A) and a JNPL3 mouse (B) under resting condition. C. The frequency of neuronal Ca²⁺ transients was increased in JNPL3 mice compared to WT mice under the resting condition. D. Frequency distribution of Ca²⁺ transients during 100 sec imaging period showed clear difference between WT and JNPL3 mice (p =0.0009, Kolmogorov-Smirnov test). X-axis shows the number of Ca²⁺ transients per 100 sec imaging period. Y-axis shows the number of neurons with different the number of Ca²⁺ transients per 100 sec imaging period. There were no differences between WT and JNPL3 mice in the amplitude of the peak Ca²⁺ transient (E) or total Ca²⁺ activity (AUC; F). Data in C, E-F are presented per animal as mean ± SEM. n = 998 neurons from 8 WT mice and n = 842 neurons from 8 JNPL3 mice were analyzed, *: p < 0.05, unpaired t-test, F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

Figure 3. Neuronal Ca²⁺ activity abnormalities in tauopathy mice under running condition.

A-B. Representative Ca²⁺ traces from a WT mouse (A) and a JNPL3 mouse (B) under running condition. C. The frequency of neuronal Ca²⁺ transients was increased in JNPL3 tauopathy mice compared to WT mice. D. Frequency distribution of Ca²⁺ transients during 100 sec imaging period showed clear difference between WT and JNPL3 mice (p = 0.0006, Kolmogorov-Smirnov test). Amplitude of the peak Ca²⁺ transient (E) and total Ca²⁺ activity (AUC; F) were decreased in JNPL3 mice, compared to WT mice. Data in C, E-F are presented per animal as mean ± SEM. n = 754 neurons from 8 WT mice and n = 829 neurons from 8 JNPL3 mice were analyzed, *: p< 0.05, **: p < 0.01, unpaired t-test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

Figure 4.

Increased neuronal activity affects calcium parameters differently in wild-type vs tauopathy mice.

A-C. In WT mice, running did not increase Ca²⁺ transient frequency but amplitude of the peak Ca²⁺ transient (B) and total Ca²⁺ activity (AUC; C) were increased, compared to resting. (D). In JNPL3 mice, running increased frequency of neuronal Ca²⁺ transients but not amplitude of the peak Ca²⁺ transient (E) or total Ca²⁺ activity (AUC; F), compared to resting. *: p < 0.05, ***: p < 0.001, paired t-test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

Figure 6. Tau mAb 4E6 has no effect on neuronal Ca²⁺ activity in tauopathy mice under resting condition when analyzed per animal.

A. Schematic diagram for tau mAb 4E6 administration in JNPL3 tauopathy mice. AAV-Syn-GCaMP6s viral construct (1–2 µl; >2×10¹² genome copies per ml) was injected into JNPL3 mice to express the Ca²⁺ indicator in neurons in layer II/III of the motor cortex. A glass window was created for two-photon Ca²⁺ imaging. Four weeks later, control IgG or 4E6 (100 µg each time, n = 7 for each group) was injected into the femoral vein of JNPL3 mice on Day 0 and Day 4. Two-photon imaging was collected on Day 0 before the first mAb injection and on Day 8 (four days after second mAb injection) under resting or running condition. Representative Ca²⁺ traces from the IgG pre-injection (pre) (B) and IgG post-injection Day 8 (C) mice under resting condition. D-F. When JNPL3 mice were resting on the treadmill, control IgG had no effect on frequency of Ca²⁺ transients (D), amplitude of the peak Ca²⁺ transient (E), or total Ca²⁺ activity (AUC; F). Representative Ca²⁺ traces from the 4E6 pre-injection (G) and 4E6 post-injection Day 8 (H) mice under resting condition. I-K. Tau mAb 4E6 had no effect on frequency of Ca²⁺ transients (I), or amplitude of the peak Ca²⁺ transient (J), or total Ca²⁺ activity per animal (AUC; K). The gray dash line indicates the respective Ca²⁺ activity level in age-matched WT mice. n = 516 neurons from 7 IgG injected JNPL3 mice and n = 574 neurons from 7 4E6 injected JNPL3 mice were analyzed, paired t-test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

Figure 7. Tau mAb 4E6 partially rescues abnormal Ca²⁺ activity in tauopathy mice under running condition.

Representative Ca²⁺ traces from the IgG pre-injection (pre) (A) and IgG post-injection Day 8 (B) mice under running condition. C-E. When JNPL3 mice were running on the treadmill, control IgG had no effect on frequency of Ca²⁺ transients (C), amplitude of the peak Ca²⁺ transient (D), or total Ca²⁺ activity (AUC; E).

Representative Ca²⁺ traces from the 4E6 pre (F) and 4E6 Day 8 (G) mice under running condition. H-J. Tau mAb 4E6 had no effect on frequency of Ca²⁺ transients (H), but increased the amplitude of the peak Ca²⁺ transient (I) and total Ca²⁺ activity (AUC; J). The gray dash line indicates the respective Ca²⁺ activity level in age-matched WT mice. n = 516 neurons from 7 IgG injected JNPL3 mice and n = 574 neurons from 7 4E6 injected JNPL3 mice were analyzed, *: p < 0.05, paired t-test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

Figure 8. Tau mAb 4E6 decreases soluble pathological tau, but has no effect on insoluble pathological tau.

A. Western blots for soluble protein fraction (human tau) and GAPDH (loading control). stained with PHF1 (phospho-tau), CP27 B-E. Quantification of protein levels in A. B. Soluble PHF1/GAPDH level was decreased after two 4E6 femoral vein injections, compared to control mice (p = 0.0197). C. No change was detected in soluble CP27/GAPDH levels (p = 0.7823). D. Soluble PHF1/CP27 was decreased in the treated mice compared to controls (p = 0.0202). E. No change was detected in GAPDH (p = 0.6489). F. Western blots for insoluble protein fraction stained with PHF1 (phospho-tau) and CP27 (human tau). G-H. Quantification of protein levels in F. Insoluble PHF1 levels (G; p = 0.3879) or insoluble CP27 levels (H; p = 0.6927) did not differ between the treated and control groups. All data are presented as mean ± SEM. n = 7 for each group. *: p < 0.05, unpaired t-test.

Figure 9. Soluble pathological tau level PHF1 correlates with the degree of Ca²⁺ activity abnormalities.

A-F. Correlation analysis of soluble pathological tau PHF1 levels with different Ca²⁺ parameters under resting and running condition in JNPL3 tauopathy mice. Under resting condition, total Ca²⁺ activity (AUC) correlated significantly with soluble PHF1 levels (C, R² = 0.343, p = 0.0278). Likewise under running condition, this tau parameter correlated significantly with frequency of the Ca²⁺ transients, (D, R² = 0.420, p = 0.0123), amplitude of the peak Ca²⁺ transient (E, R² = 0.587, p = 0.0014), and total Ca²⁺ activity (AUC) (F, R² = 0.482, p = 0.0059). R is Pearson correlation coefficient. n = 7 from each group were analyzed.

Animal surgery

Surgery was performed using aseptic techniques under isoflurane anesthesia (3% for induction, 1.5–2% for surgery). Briefly, the mouse was placed on a heated pad to maintain a body temperature of ~37 °C. The depth of anesthesia was monitored frequently during the surgery. The animal was placed in a stereotaxic frame, incision was made and the skull exposed to reveal the landmarks (bregma, lambda and midline) needed for cranial window. Subsequently, a section of the skull, about 3 mm in diameter, over the right motor cortex was removed.

To express Ca²⁺ indicator, a total of 1 – 2 µL of AAV5-Syne-GCaMP6s virus (2.52 × 10¹² genome copies per mL, Penn Vector Core) in an artificial cerebrospinal fluid (ACSF: 125 mM NaCl, 2.5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 25 mM NaHCO₃, 1.25 mM NaH₂PO₄, 25 mM glucose, pH 7.4) was slowly injected by glass micropipette into layer II/III of the motor cortex (0 – 1.5 mm posterior from bregma and 0 – 1.5 mm lateral from midline) over 10 – 15 min. Then the removed skull section was replaced by a precut #1 round cover glass (Denville Scientific, M0720). Dental cement was used to seal the edges of the cover glass and embed a head holder composed of two parallel micro-metal bars. This holder was used to attach the mouse securely on a custom-built free-floating treadmill (Bai et al., 2017; Cichon and Gan, 2015; Li et al., 2017). After surgery, the mice were individually housed. Four weeks later, when the window had cleared and neuronal GCaMP6 expression was robust, the mice were ready for two-photon imaging.

Antibodies and intravenous injections

Monoclonal antibody (mAb) 4E6G7 (referred to as 4E6) was purified from our hybridomas by Genscript (Paramus, NJ). IgG1κ (referred to as IgG) of the same isotype as the tau mAb 4E6 served as control (eBioscience, 16–4714). IgG and 4E6 were injected (100 µg each time) twice into the femoral vein of the JNPL3 mice on Day 0 and Day 4 (Fig. 6A).

In vivo two-photon imaging

A custom built free-floating treadmill (101 cm × 58 cm × 44 cm) was used for awake animal two-photon imaging (Bai et al., 2017; Cichon and Gan, 2015; Li et al., 2017). This free-floating treadmill allowed head-fixed mice to move their forelimbs freely to perform motor running tasks. To minimize motion artifact during imaging, the treadmill was constructed so that all the moving parts (motor, belt and drive shaft) were isolated from the microscope stage and the supporting air-table. Only mice with intact motor function that could successfully run on the treadmill were used in the study. The motor phenotype in the JNPL3 mice we used has now shifted to older age and it not typically seen at the age used here (Lin et al., 2019).

Two-photon imaging was collected with an Olympus Fluoview 1000 two-photon system (920 nm) equipped with a Ti:Sapphire laser (MaiTai DeepSee, Spectra Physics). Ca²⁺ signals were recorded at 1 Hz using a 25X objective (NA 1.05, 512 × 512 pixels). The mice were forced to run (referred to as running) at the speed of 1.67 cm/s or allowed to rest (referred to as resting) on a treadmill with the head fixed on top of the custom-built free-floating treadmill. Two-photon Ca²⁺ images were collected for five trials (100 secs for each trial) at different focal planes (motor cortex containing GCaMP6-positive neurons at the depth of 150 – 300 µm below the pial surface were imaged) under resting or running conditions. Resting and running tests were performed within the same testing session. The same focal planes (same cells) were imaged during these two conditions. For before vs. after treatment, the same focal planes were imaged but because these images were collected at different days, there may have been a slight shift in the plane so the same cells may not have been analyzed as reflected in the statistical analysis (unpaired comparison).

The use of GCaMP6s and 1 Hz sampling rate does not capture certain physiological responses such as patterns of neuronal action potentials. We chose to use the slow GCaMP6s, instead of the fast GCaMP6f, because it is more sensitive for detecting small changes in Ca²⁺ between WT and JNPL3 mice. The 1Hz sampling rate was chosen to match the slow kinetics of GCaMP6s and to minimize the potential phototoxicity. GCaMP6s and 1Hz sampling rate were used in our previous studies to reveal Ca²⁺ dynamics under different experimental conditions (Bai et al., 2017; Cichon and Gan, 2015). This particular Ca²⁺ indicator has also be used previously for Ca²⁺ imaging in tauopathy mice (Marinkovic et al., 2019).

Image analysis

Neuronal Ca²⁺ activity, indicated by GCaMP6 fluorescence changes, was analyzed post hoc using Image J software (NIH). The GCaMP6 fluorescence (F) during resting and running was measured by averaging pixels within each soma of GCaMP6 positive neurons. The frequency of the Ca²⁺ transients, amplitude of the peak Ca²⁺ transient, and total Ca²⁺ activity (area under the curve, AUC) (Bruni et al., 2017; Overk et al., 2015; Reznichenko et al., 2012; Schmunk et al., 2017) were analyzed by using R/RStudio software. A time window that spanned 10% of the trace length was slid across the trace, and the window with minimal standard deviation (SD) was selected automatically by the software as the baseline interval (F0). F0 differs between animals. Therefore, F is normalized to F0 and changes reported as ΔF/F0. F0 did not appear to be influenced by mouse strain (JNPL3 vs. WT) but it was the same for resting and running conditions because the same neuron was analyzed. F0 and threshold were determined as the mean and three times of the SD of the F values within the baseline interval. The trace was then normalized as ΔF/F0 = (F-F0)/F0. The frequency of Ca²⁺ transients was calculated as the number of Ca²⁺ transients per 100 seconds for each soma. The amplitude of the peak Ca²⁺ transient was the highest amplitude value of the Ca²⁺ transients. The total Ca²⁺ activity (area under the curve, AUC) was the average of all the integrals over the time periods (100 seconds) above the threshold. See Supplementary Tables 1–2.

Tissue Processing

After imaging, the brains were collected for biochemical and immunohistochemistry assays. Briefly, the mice were perfused with PBS and the brains extracted. The right hemisphere with a sliver of the left hemisphere to maintain tissue integrity was fixed in 4% paraformaldehyde at 4°C overnight, followed by at least 24 h immersion in 2% DMSO and 20% glycerol in phosphate buffer at 4°C, and then cryo-sectioned coronally at 40 µm for immunohistochemistry (see below). The left hemisphere of the brain was homogenized in (5 × vol/w) modified RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, pH 7.4) with protease and phosphatase inhibitors (1 µg/mL of protease inhibitor mixture (cOmplete, Roche), 1 mM NaF, 1 mM Na₃VO₄, 1 nM PMSF and 0.25% sodium deoxycholate). The brain homogenate was centrifuged (20,000 × g) for 20 min at 20 °C and supernatants were collected as soluble tau fraction LSS (Low Speed Supernatant). For the sarkosyl insoluble fraction, equal amounts of protein from LSS were mixed with 10% sarkosyl solution to a final 1% sarkosyl concentration, and the sample mixed for 30 min at room temperature, then centrifuged at 100,000 × g for 1 h at 20 °C. The pellet was then washed with 1% sarkosyl solution and centrifuged again at 100,000 × g for 1 h at 20 °C. The sarkosyl pellet (SP) fraction was then air dried for 30 min, mixed with 50 µL of modified O+ buffer (62.5 mM Tris- HCl, 10% glycerol, 5% β-mercaptoethanol, 2.3% SDS, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 1 mM Na₃VO₄, 1 nM PMSF, and 1 µg/mL of the protease inhibitor mixture). The LSS was eluted with O+ buffer (1:5).

Synaptosome protein extraction

Synaptosome protein was extracted from mouse brain (10–12 month old, n=6 per group) as described by others (Kamat et al., 2014) with modification. Briefly, freshly isolated mouse brain was homogenized in 10% (w/v) of 0.32 M sucrose HEPES buffer (145 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 5 mM glucose, 5 mM HEPES). The homogenate was centrifuged for 10 min at 600 × g at 4 °C. The supernatant was then diluted 1:1 with 1.3 M sucrose HEPES buffer and centrifuged at 20,000 × g for 30 min at 4 °C. The pellet was then resuspended in RIPA buffer mixed with 1 µg/ml protease inhibitor mixture and 1 nM PMSF.

Western blots

All of the samples were boiled, and electrophoresed on 12% SDS-PAGE gels and transferred to nitrocellulose membranes. The blots were blocked in Superblock® (Thermo Fisher Scientific), incubated with PHF1 (1:1000, gift from Peter Davies), CP27 (1:500, gift from Peter Davies), GAPDH (1:3000, Cell Signaling Technology, D16H11), GluR1 (1:50, Antibodies Incorporated, 73–327), GluR2 (1:50, Antibodies Incorporated, 73–002), NR1 (1:50, Antibodies Incorporated, 73–272), GluK2 (1:1500, Alomone Labs, Jerusalem, Israel, AGC-009), Ca_v 1, 2 (1:1500, Alomone,Labs, ACC-022), synaptophysin (1:1000, clone svp-38, Sigma, s-5768), and PSD-95 (1:1000, Antibodies Incorporated, 75–028). The blots then were incubated with HRP-conjugated anti-Fab secondary antibody (1:3000, Thermo Fisher Scientific) and detected with ECL substrates (Thermo Fisher Scientific), or incubated with IRDye 800CW Goat anti-Rabbit/ IRDye® 680RD Goat anti-Mouse (1:10000, LI-COR Biosciences) and detected with Odyssey® CLx imaging system (LI-COR Biosciences). Images of immunoreactive bands were quantified by MultiGauge software (Fujifilm Corporation) or Image J software (NIH).

Immunohistochemistry

Free floating brain sections were washed in PBS, followed by permeabilization and blocking in PBS containing 5% bovine serum albumin and 0.3% Triton X-100 at room temperature for 2 hours. The sections were then incubated with diluted primary antibodies in blocking solution without Triton X-100 at 4°C overnight. The dilution of the primary antibodies was as follows: PHF1, 1:1000 (generously provided by Peter Davies); anti-GFP, 1:1000 (Abcam ab290). Then the brain sections were washed thoroughly in PBS and incubated with goat anti-mouse Alexa Fluor 568 (PHF1) or goat anti-rabbit Alexa Fluor 488 (anti-GFP) antibodies (1:1000) at room temperature for one hour. DAPI was used to stain the nuclei. After washing, the sections were mounted in ProLong Gold antifade reagent (ThermoFisher, P10144), and imaged by an LSM 700 Zeiss confocal laser scanning microscope at 5x or 10x magnification. The images were processed by Fiji imageJ software (https://imagej.net/Fiji).

Statistics

Data are presented as mean ± SEM. Shapiro-Wilk normality test was used to test whether datasets were normally distributed. Data that passed that test was analyzed by the parametric unpaired (Fig. 2C, E, F, 3C, E, F, 8, S6–S8) or paired (Fig. 4, 6, 7) Student’s t-test. Data that failed the normality test was analyzed by the non-parametric Mann-Whitney test (Fig. S1–S4). All tests were two-tailed. Likewise, correlation analysis was performed by the Pearson test (Fig. 9, S5). Frequency distribution was compared by the two-sample Kolmogorov-Smirnov test (Fig. 2D, 3D). Significance level was set at p ≤ 0.05. All statistical analyses were performed using GraphPad Prism 8.

Results

Tauopathy mice showed abnormal somatic Ca²⁺ activity in layer II/III motor cortex neurons compared to age-matched wild-type (WT) mice under resting condition.

To explore if tau pathology affects neuronal Ca²⁺ activity, 10–12 month old JNPL3 mice and age-matched WT mice were examined by two-photon Ca²⁺ imaging. AAV-Syn-GCaMP6s virus was injected into WT or JNPL3 mice to express the Ca²⁺ indicator GCaMP6s in neurons in layer II/III of the motor cortex. Four weeks later, somatic Ca²⁺ activity was collected by two-photon imaging. There was a limited Ca²⁺ activity when the animals were under anesthesia. Therefore, the two-photon Ca²⁺ images were collected when the mice were awake, resting or running on a treadmill. The relative fluorescence change (ΔF/F0) versus time traces was analyzed for each neuron within the imaging field in WT or JNPL3 mice (Fig. 1A–1D). The frequency distribution of Ca²⁺ transients, amplitude of the peak Ca²⁺ transient and average total Ca²⁺ activity (area under the curve (AUC)) were measured (Fig. 1E).

Figure 1. Imaging somatic neuronal Ca²⁺ activity in tauopathy mice.

Neurons in layer II/III M1 region were imaged in vivo in WT and JNPL3 tauopathy mice. A. Representative two-photon Ca²⁺ image from a WT mouse (11 month). B. Representative somatic Ca²⁺ transients in A. C. Representative two-photon Ca²⁺ image from a JNPL3 mouse (11 month). D. Representative somatic Ca²⁺ transients in C. E. Illustration of quantified Ca²⁺ activity parameters. A time window that spanned 10% of the trace length was slid across the trace, and the window with minimal standard deviation (SD) was selected as the baseline interval. F0 was determined as the mean of the baseline interval. Threshold was determined as F0 plus three times the SD of the baseline interval. The trace was then normalized as ΔF/F0 = (F-F0)/F0. The frequency of Ca²⁺ transients was calculated as the number of Ca²⁺ transients per 100 seconds for each soma (indicated by the gray arrows). The amplitude of the peak Ca²⁺ transient was the highest amplitude value of the Ca²⁺ transients. The total Ca²⁺ activity (area under the curve, AUC) was the average of all the integrals over the time periods (100 seconds) above the threshold.

Under the resting condition, the JNPL3 mice showed higher frequency of Ca²⁺ transients per neuron than WT mice (Fig. 2C (per animal), p < 0.05, Suppl. Fig. 1A (per neuron), p = 0.0041, n = 998 neurons from 8 WT mice and n = 842 neurons from 8 JNPL3 mice). Likewise, the frequency distribution of Ca²⁺ transients showed clear differences between WT and JNPL3 mice (Fig. 2D, p = 0.0009). There were no differences between JNPL3 and WT mice in amplitude of the peak Ca²⁺ transient (Fig. 2E (per animal), Suppl. Fig. 1B (per neuron)), or average total Ca²⁺ activity (area under the curve (AUC); Fig. 2F (per animal), Suppl. Fig. 1C (per neuron)).

Tauopathy mice showed abnormal somatic Ca²⁺ activity of neurons in layer II/III motor cortex compared to age-matched WT mice under running condition.

When the animals were forced to run on a treadmill, we detected more severe Ca²⁺ activity abnormalities per neuron in JNPL3 mice compared to WT mice (Fig. 3 (per animal), Suppl. Fig. 2 (per neuron), n = 754 neurons from 8 mice WT mice and n = 829 neurons from 8 JNPL3 mice). Compared to the age-matched WT mice, JNPL3 mice had higher frequency of Ca²⁺ transients (Fig. 3C (per animal), p < 0.01, Suppl. Fig. 2A (per neuron), p < 0.0001) and the frequency distribution of Ca²⁺ transients showed clear differences between WT and JNPL3 mice (Fig. 3D, p = 0.0006). Furthermore, a lower amplitude of the peak Ca²⁺ transient (Fig. 3E (per animal), p < 0.01, Suppl. Fig. 2B (per neuron), p < 0.0001), and lower total Ca²⁺ activity (AUC) (Fig. 3F (per animal), p < 0.05, Suppl. Fig. 2C (per neuron), p < 0.0001) were found in JNPL3 mice compared to WT mice. Higher amplitude of the peak Ca²⁺ transient and total Ca²⁺ activity (AUC) had been detected under the running condition in WT mice (Fig. 4A–C). However, the JNPL3 mice were not able to adapt to the strong stimulation under running condition (Fig. 4D–F).

The abnormal somatic Ca²⁺ activity in the tauopathy mice is not associated with extensive tau deposition.

Limited tau deposition was detected in two-photon imaged brain region in the JNPL3 mice, suggesting that the in vivo neuronal Ca²⁺ abnormalities were likely related to soluble pathological tau (Fig. 5).

Figure 5. Limited tau deposition in two-photon imaged brain region.

A-D. Representative images of GCaMP6s (green) and PHF1 (red) staining in JNPL3 (A, B) and wild-type mice (C, D). The box shown in merged image in A and C is magnified in B and D, respectively. DAPI (blue) stains nuclei.

A, B. Limited tau deposition is visible in the two-photon imaged motor cortex in JNPL3 mice. C, D. As expected, no PHF1 immunoreactivity was detected in the two-photon imaged motor cortex in wild-type mice. Scale bar: 100 μm.

Tau mAb 4E6 increased total Ca²⁺ activity in tauopathy mice under resting condition when analyzed per neuron.

Our previous study showed that tau mAb 4E6 crosses the blood-brain-barrier and is primarily detected within neurons in the brain, where it colocalizes with tau aggregate dye FSB and leads to clearance of such labelled tau assemblies (Wu et al., 2018). To assess the effects of tau mAb 4E6 on neuronal Ca²⁺ activity in live tauopathy mice, tau mAb 4E6 and isotype control IgG1κ (100 µg each time) were injected twice into the femoral vein of the JNPL3 mice on Day 0 and Day 4, and two-photon Ca²⁺ images were collected before and after antibody injections on Day 0 and Day 8 under resting and running conditions (Fig. 6A).

Analyzed per animal and compared with IgG control (Fig. 6B–F (per animal), Suppl. Fig. 3A–C (per neuron)), tau mAb 4E6 (Fig. 6G–K (per animal), Suppl. Fig. 3D–F (per neuron)) had no effect on frequency of Ca²⁺ transients (Fig. 6I (per animal), Suppl. Fig. 3D (per neuron)), amplitude of the peak Ca²⁺ transient (Fig. 6J (per animal), Suppl. Fig. 3E (per neuron)), or total Ca²⁺ activity (AUC) (Fig. 6K (per animal)) in tauopathy mice under resting condition after two femoral vein injections. However, when analyzed per neuron, 4E6 increased the total Ca²⁺ activity (AUC) (Suppl. Fig. 3F, p < 0.0001).

Tau mAb 4E6 partially rescued abnormal Ca²⁺ activity abnormalities in tauopathy mice under running condition.

Under the running condition, IgG control had no effects on the three Ca²⁺ parameters (Fig. 7A–E (per animal), Suppl. Fig. 4A–C (per neuron)). Tau mAb 4E6 had no effect on the frequency of Ca²⁺ transients (Fig. 7H (per animal), Suppl. Fig. 4D (per neuron)), but it rescued the amplitude of the peak Ca²⁺ transient (Fig. 7I (per animal, p < 0.05, Suppl. Fig. 4D (per neuron), p < 0.0001) and total Ca²⁺ activity (AUC) (Fig. 7J (per animal), p < 0.05, Suppl. Fig. 4D (per neuron), p < 0.001) in tauopathy mice after two femoral vein injections.

Tau mAb 4E6 decreased soluble pathological tau levels in tauopathy mice after two femoral vein injections.

Following two-photon imaging, the mouse brains were collected for biochemical analysis to examine levels of soluble (Fig. 8A–E) and insoluble (Fig. 8F–H) tau protein fractions. PHF1 and CP27 were used to detect phospho-tau and total human tau, respectively. Consistent with efficacy in previous studies (Congdon et al., 2013; Congdon et al., 2016; Gu et al., 2013; Wu et al., 2018), two femoral vein injections of 4E6, compared to IgG control, decreased soluble PHF1 (Fig. 8B, p = 0.0197) and PHF1/CP27 (Fig. 8D, p = 0.0202) levels. Tau mAb 4E6 had no effect on insoluble tau levels (Fig. 8F–H), which again fits our prior findings using a similar acute therapeutic approach (Congdon et al., 2016).

Pathological tau levels correlated with the degree of Ca²⁺ activity abnormalities in tauopathy mice.

Next, we explored whether pathological tau levels correlated with the degree of Ca²⁺ activity abnormalities in treated and control tauopathy mice (n = 7 per group). Under resting condition, correlation analysis showed significant correlation between soluble PHF1 levels and total Ca²⁺ activity (AUC) (Fig. 9C, R² = 0.343, p = 0.0278). Under running condition, the same analysis showed stronger correlation that was significant for more comparisons. Specifically, for soluble PHF1 and: 1) frequency of Ca²⁺ transients (Fig. 9D, R² = 0.420, p = 0.0123); 2) amplitude of the peak Ca²⁺ transient (Fig. 9E, R² = 0.587, p = 0.0014); and 3) total Ca²⁺ activity (AUC) (Fig. 9F, R² = 0.482, p = 0.0059). Insoluble pathological tau PHF1 had less significant correlation with the degree of Ca²⁺ activity abnormalities (Suppl. Fig. 5). Correlation analysis showed significant correlation between insoluble PHF1 levels and: 1) total Ca²⁺ activity (AUC) (Suppl. Fig. 5C, R² = 0.369, p = 0.0212) under resting condition; and 2) frequency of Ca²⁺ transients (Suppl. Fig. 5D, R² = 0.306, p = 0.0400) under running condition.

AMPA receptor subunits and synaptophysin levels were decreased in tauopathy mice compared to WT mice but 4E6 was unable to reverse these deficits.

Since Ca²⁺ signaling involves glutamatergic transmission, key glutamate receptors were analyzed. Their levels are too low to measure in western blots from low speed fraction of brain homogenate, but are enriched in synaptosomes. Synaptosomes were extracted from JNPL3 (n = 6) and WT (n = 6) mouse brains (10–12 months old) and analyzed on western blots for AMPA subunit GluR1, GluR2, NMDA subunit NR1, synaptophysin, and PSD-95 (Suppl. Fig. 6).

The expression of AMPA subunit GluR1 (Suppl. Fig. 6D, p = 0.0414) and GluR2 (Suppl. Fig. 6E, p = 0.0460) decreased in JNPL3 mice brain compared to age-matched WT mice. NR1 levels were not changed (Suppl. Fig. 6F). In addition, synaptophysin levels were decreased (Suppl. Fig. 6B, p = 0.0473) in the JNPL3 mice compared to WT controls. PSD-95 levels did not differ between tauopathy and WT mice (Suppl. Fig. 6C). Additional analysis showed that a major kainate receptor subunit GluK2 and L-type voltage gated ion channels were not changed in JNPL3 mice compared to wild-type mice (Suppl. Fig. 7).

Subsequently, the effect of tau mAb 4E6 on the same synaptosome proteins was examined (Suppl. Fig. 8). Tau mAb 4E6-mediated clearance of soluble pathological tau was confirmed (Suppl. Fig. 8A–B, p = 0.0315). Compared to the IgG control (n = 6), acute 4E6 treatment (n = 7) had no effect on the expression of AMPA subunit GluR1, GluR2, NMDA subunit NR1, or on synaptophysin and PSD-95.

Discussion

In this study, two-photon Ca²⁺ imaging revealed abnormal neuronal Ca²⁺ activity in layer II/III motor cortex in awake tauopathy mice, compared to age-matched wild-type (WT) mice, which was more pronounced in running vs. resting animals. These Ca²⁺ abnormalities may relate to reduction in AMPA receptor subunits in the tauopathy mice compared to WT mice. Furthermore, acute administration of a tau monoclonal antibody (mAb), 4E6, partially rescued Ca²⁺ activity abnormalities in the tauopathy mice, and decreased brain levels of soluble pathological tau protein. Finally, the levels of soluble pathological tau correlated better with the degree of Ca²⁺ dyshomeostasis than levels of insoluble tau in the tauopathy mice.

The reasons for the neuronal dysfunction in the JNPL3 mice compared to the WT mice are likely complex and not attributable to one specific component. It is more likely linked to soluble pathological tau than tau deposition because limited tau aggregates were detected in the imaged region (Fig. 5). We can speculate that the increased frequency reflects more erratic firing of presynaptic neurons related to tau pathology. In addition, the fact that the peak amplitude and overall neuronal activity in the JNPL3 mice is less responsive to running stimulation may reflect weakened responses of postsynaptic neurons (Fig. 4). Loss of presynaptic input and AMPA receptors as reflected by decreased synaptophysin and GluR1 and GluR2 in JNPL3 mice compared to WT mice (Suppl. Fig. 6) are likely a factor in this dysfunction. A caveat to this interpretation, however, is that we are analyzing synaptosomes from one hemisphere instead of only the imaged motor cortex. The reason being that the small imaged brain region does not yield sufficient amount of synaptosomes for analysis. Therefore, the data gathered from the whole hemisphere sample may not be entirely representative of the observed outcome in cortex.

Ca²⁺ dyshomeostasis has long been recognized in AD (Alzheimer’s Association Calcium Hypothesis, 2017), and has been extensively studied in models linked to Aβ and/or presenilin pathology (Busche et al., 2012; Busche et al., 2008; Busche et al., 2015; Grienberger et al., 2012). Potential link between tau and Ca²⁺ abnormalities has been less examined and the findings are variable. A report examining the rTg4510 tauopathy mouse model at 8–10 months of age did not detect changes in Ca²⁺ homeostasis under anesthetized conditions compared to controls (Kopeikina et al., 2013). This model has a more aggressive tauopathy phenotype than the JNPL3 model although it expresses the same familial P301L tau mutation. A subsequent report revealed that neurofibrillary tangle (NFT)-bearing neurons in the same model at the same age have normal Ca²⁺ levels in awake animals, that received visual stimulation (Kuchibhotla et al., 2014). More recently, a strong reduction in frequency of cortical Ca²⁺ transients was found in the same rTg4510 mouse model at a similar age range (6–12 months) or in older animals (17–24 months) under anesthesia (Busche et al., 2019). Another report by a different group detected reduced Ca²⁺ transient frequency in layer II/III in frontal cortex in awake 2–3 months old P301S mice, which also have a more severe progression of tauopathy than the JNPL3 model (Marinkovic et al., 2019). Furthermore, the P301S mice received an intracerebral injection of preformed tau fibrils, which further enhances tau pathology. Importantly, none of these prior studies measured the peak amplitude of the neuronal Ca²⁺ signal or total Ca²⁺ activity (AUC), which we show to be decreased in tauopathy mice, compared to WT controls, when neuronal activity is increased (running condition). Others have shown Ca²⁺ dysregulation in synaptic boutons in hippocampal slices from a different tauopathy transgenic mouse model that expresses the tau repeat domain with a pro-aggregant mutation, K280 (Decker et al., 2015). Studies in yet another tauopathy mouse model, expressing two familial Pick’s disease mutations (L266V and G272V), showed decreased AUC in some layers of barrel cortex with or without electric whisker pad stimulation, compared to the WT controls (Overk et al., 2015). Related studies in different models have indicated dysregulation of Ca²⁺ evoked by electrical stimulation in induced pluripotent stem cell derived tauopathy neurons (Imamura et al., 2016), and that intracellular Ca²⁺ release can be triggered by preterminal injection of recombinant tau in the squid giant synapse (Moreno et al., 2016).

In our hands, cortical Ca²⁺ activity was low under our experimental conditions when the mice were anaesthetized, which likely will make it more difficult to detect differences compared to WT mice. That may explain some of the findings by others as detailed above. Other differences within those studies and compared to ours may be explained by the different experimental conditions, including models, age and activity level of the animal, Ca²⁺ indicators, and types of Ca²⁺ activity analyzed. For example, some of the phenotype of the rTg4510 model is likely to be caused by transgene disruption of other genes rather than tauopathy per se (Gamache et al., 2019). Importantly as well, none of the prior studies examined mice under running conditions. The enhanced neuronal activity during such a state shows more pronounced neuronal Ca²⁺ dysfunction, and allows for better detection of therapeutic efficacy of antibody treatment.

In our study, we report our findings both based on overall number of neurons analyzed and per animal. As expected, more significant differences are detected with the former mode of analysis but the pattern is the same for the latter approach. Prior articles that analyzed in vivo neuronal Ca²⁺ in tauopathy mice, in other related neurodegenerative mouse model or in non-disease mouse models did these measurements per dendrite and spine (Bai et al., 2017; Cichon and Gan, 2015; Kopeikina et al., 2013; Li et al., 2017), per neuron (Bai et al., 2017; Grienberger et al., 2012; Kuchibhotla et al., 2014), per animal (Busche et al., 2012; Busche et al., 2015; Busche et al., 2019; Marinkovic et al., 2019), and in one case both per neuron and animal (Overk et al., 2015). Some of the reported differences between the tauopathy studies may be explained by these different forms of analysis. Ideally, both should be reported (per unit of measure and per animal) as we do here, for better insight into the severity of the tauopathy-induced Ca²⁺ dysfunction. The number of units (dendrites or neurons) or animals analyzed differs also between these studies, ranging from 2000–3000 dendrites or 150–900 neurons in total per group, or 3–9 animals per group, with our numbers being at the higher end (516–998 neurons analyzed by group in 7–8 animals per group). Overall, these differences and others in the experimental design (model, age, brain region, calcium indicator and parameters, awake or anesthetized) likely explain the different outcomes of these studies, including ours. However, it is not surprising that increased neuronal activity reveals more pronounced calcium dyshomeostasis in tauopathy mice than previously reported.

Ca²⁺ imaging in these mice was made feasible by an AAV-mediated expression of a Ca²⁺ indicator, GCaMP6s, which we have used in previous studies to reveal Ca²⁺ dynamics under different experimental conditions (Bai et al., 2017; Cichon and Gan, 2015). It has also be used previously for Ca²⁺ imaging in tauopathy mice (Marinkovic et al., 2019). It should be noted that this particular Ca²⁺ indicator does not capture certain physiological responses such as patterns of neuronal action potentials. We chose to use the slow GCaMP6s, instead of the fast GCaMP6f, because it is more sensitive for detecting small changes in Ca²⁺ between WT and JNPL3 mice. It also allows for a slower sampling rate to minimize potential phototoxicity that could otherwise occur during frequent imaging. In future studies, we plan to examine the feasibility of using GCaMP6f to reveal Ca²⁺ abnormalities in tauopathy models.

How the other hallmark of AD, amyloid-β (Aβ) influences Ca²⁺ signaling has also been explored, reporting increased neuronal Ca²⁺ activity with two-photon imaging in anesthetized Aβ plaque mice (Busche et al., 2012; Busche et al., 2008), which may impair neuronal circuits involved in cognitive information processing (Grienberger et al., 2012). Surprisingly, two different Aβ mAbs, which reduced brain Aβ burden, increased numbers of hyperactive neurons (Busche et al., 2015), suggesting long-term detrimental effect of Aβ targeting therapy in this context. In contrast, our acute tau antibody therapy, reversed or attenuated tau pathology induced neuronal Ca²⁺ abnormalities. Previously, tau mAb 4E6 has been found effective in reducing pathological tau in culture cells, brain slices and animal models (Congdon et al., 2019; Congdon et al., 2013; Congdon et al., 2016; Gu et al., 2013; Wu et al., 2018). Importantly, it can acutely improve cognitive function in tauopathy mice (Congdon et al., 2016). Compared to the IgG control, tau mAb 4E6 increased the amplitude of the peak Ca²⁺ transient under running condition (Fig. 7I, Suppl Fig. 4E), and total Ca²⁺ activity under running conditions (AUC; Fig. 7J, Suppl. Fig. 4F). This suggests that 4E6 can rescue Ca²⁺ activity abnormalities and restore Ca²⁺ homeostasis in JNPL3 mice, and its beneficial effects are more obvious when mice exhibit high neuronal activity during running.

Consistent with our previous studies, the tau mAb 4E6 significantly decreased PHF1-reactive soluble pathological tau after two femoral vein injections (Fig. 8A–E) (Congdon et al., 2019; Congdon et al., 2013; Congdon et al., 2016; Gu et al., 2013; Wu et al., 2018). Interestingly, the degree of Ca²⁺ activity abnormalities correlated with PHF1-reactive tau levels, and in more ways with its soluble- than its insoluble fraction. Soluble PHF1-reactive tau correlated negatively with total Ca²⁺ activity (AUC) under both resting and running conditions (Fig. 9C, 9F), positively with frequency of the Ca²⁺ transients (Fig. 9D) and negatively with amplitude of the peak Ca²⁺ transient (Fig. 9E) under the running condition. Insoluble PHF1-reactive tau levels correlated negatively with total Ca²⁺ activity (AUC) under the resting condition (Suppl. Fig. 5C) and positively with the frequency of Ca²⁺ transients under the running condition (Suppl. Fig. 5D). However, many of the other correlations analyzed for the insoluble tau fraction showed the same direction as for the soluble levels and were close to being significant (Suppl. Fig. 5B, p = 0.0897, 5E, p = 0.0934, 5F, p = 0.1159). Studies from numerous labs, including ours, suggest that soluble forms of pathological tau may be more toxic than insoluble tau (Castillo-Carranza et al., 2014; Congdon et al., 2016; Ghag et al., 2018; Goedert and Spillantini, 2017; Jiang et al., 2018; Kopeikina et al., 2012; Lasagna-Reeves et al., 2011; Ma et al., 2013; Santacruz et al., 2005). In particular, 4E6’s efficacy has been linked to its neutralization of toxic soluble tau (Congdon et al., 2019; Congdon et al., 2013; Congdon et al., 2016; Gu et al., 2013), which fits with its therapeutic effects in this study, and highlights why soluble pathological tau correlates better than insoluble tau with Ca²⁺ abnormalities in the mice. However, these two pools of pathological tau relate closely to each other and the nature of the tau fractionation process may increase variance in measured insoluble tau levels, and thereby decrease the significance of its correlation with the various Ca²⁺ parameters.

With regard to how tau pathology may induce Ca²⁺ dyshomeostasis, several possibilities exist although the actual mechanisms remain unknown. These include alteration in Ca²⁺ entry, intracellular release and/or intracellular sequestration. The ability of 4E6 to correct at least some of these abnormalities likely relates to its documented effect on tau pathology as shown here and previously (Congdon et al., 2019; Congdon et al., 2013; Congdon et al., 2016; Gu et al., 2013; Wu et al., 2018). Excitatory amino acids have for many years been closely linked to neurodegenerative diseases, such as tauopathies, and their receptors are involved in Ca²⁺ signaling. To examine their possible involvement in the JNPL3 model, brain levels of AMPA, NMDA and kainate receptor subunits were compared to age-matched WT mice, in addition to L-type Ca²⁺ channels. The analysis revealed significant reduction in AMPA receptor subunits but not in NMDA, kainate or L-type Ca²⁺ channels in synaptosomes isolated from JNPL3 tauopathy mice compared to WT mice (Suppl. Fig. 6–7). AMPA receptor downregulation has been previously found in a different tauopathy model, rTg4510 (Hoover et al., 2010), and in APP/PS1 mice (Chang et al., 2006), and impaired AMPA signaling has been reported in 3xTg mice (Baglietto-Vargas et al., 2018). However, the acute treatment with tau mAb 4E6 did not reverse the reduction of AMPA receptor GluR1, GluR2 or synaptophysin (Suppl. Fig. 8), although it did reduce pathological tau and significantly rescue tauopathy induced abnormalities in neuronal Ca²⁺ activity. As mentioned above, Ca²⁺ dyshomeostasis involves several pathways, of which changes in AMPA receptor numbers is only one component and may not be evident under the relatively acute treatment conditions employed here. Regardless, the beneficial effect of acute 4E6 treatment on Ca²⁺ homeostasis is highly significant and fits nicely with our prior studies showing its therapeutic potential in culture and in vivo, including cognitive improvements (Congdon et al., 2019; Congdon et al., 2013; Congdon et al., 2016; Gu et al., 2013; Wu et al., 2018).

Conclusions

Our findings show a more pronounced neuronal Ca²⁺ dysregulation in tauopathy mice than previously reported using two-photon imaging that relates better to soluble than insoluble levels of pathological tau. Importantly, this neuronal dysfunction is primarily observed during increased neuronal activity, and can be corrected with a tau antibody that targets soluble hyperphosphorylated tau. The repeated in vivo assessment of neuronal function and integrity conducted herein, before and after therapy provides a valuable insight into functional deficits associated with tau pathology and the mechanisms of tau immunotherapies.

Supplementary Material

1. Supplementary Figure 1. Neuronal Ca²⁺ activity abnormalities in tauopathy mice under resting condition, analyzed per neuron.

A. The frequency of neuronal Ca²⁺ transients was increased in JNPL3 mice compared to WT mice. There were no differences between WT and JNPL3 mice in the amplitude of the peak Ca²⁺ transient (B) or total Ca²⁺ activity (AUC; C). All the data are presented as mean ± SEM. n = 998 neurons from 8 WT mice and n = 842 neurons from 8 JNPL3 mice were analyzed, **: p < 0.01; Mann-Whitney test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

2. Supplementary Figure 2. Neuronal Ca²⁺ activity abnormalities in tauopathy mice under running condition, analyzed per neuron.

A. The frequency of neuronal Ca²⁺ transients was increased in JNPL3 mice compared to WT mice. Amplitude of the peak Ca²⁺ transient (B) and total Ca²⁺ activity (AUC; C) were decreased in JNPL3 mice, compared to WT mice. All the data are presented as mean ± SEM. n = 998 neurons from 8 WT mice and n = 842 neurons from 8 JNPL3 mice were analyzed, ****: p < 0.0001, Mann-Whitney test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

3. Supplementary Figure 3. Tau mAb 4E6 increases total Ca²⁺ activity in tauopathy mice under resting condition, analyzed per neuron.

When JNPL3 mice were resting on the treadmill, control IgG had no effect on frequency of Ca²⁺ transients (A), amplitude of the peak Ca²⁺ transient (B), or total Ca²⁺ activity (AUC; C). Tau mAb 4E6 had no effect on frequency of Ca²⁺ transients (D) and amplitude of the peak Ca²⁺ transient (E). But it increased the total Ca²⁺ activity (AUC; F). The gray dash line indicates the respective Ca²⁺ activity level in age-matched WT mice. n = 516 neurons from 7 IgG injected JNPL3 mice and n = 574 neurons from 7 4E6 injected JNPL3 mice were analyzed, ****: p < 0.0001, Mann-Whitney test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

4. Supplementary Figure 4. Tau mAb 4E6 partially rescues abnormal Ca²⁺ activity in tauopathy mice under running condition, analyzed per neuron.

When JNPL3 mice were running on the treadmill, control IgG had no effect on frequency of Ca²⁺ transients (A), amplitude of the peak Ca²⁺ transient (B), or total Ca²⁺ activity (AUC; C). Tau mAb 4E6 had no effect on frequency of Ca²⁺ transients (D), but increased the amplitude of the peak Ca²⁺ transient (E) and total Ca²⁺ activity (AUC; F). The gray dash line indicates the respective Ca²⁺ activity level in age-matched WT mice. n = 516 neurons from 7 IgG injected JNPL3 mice and n = 574 neurons from 7 4E6 injected JNPL3 mice were analyzed, ***: p < 0.001, ****: p < 0.0001, Mann-Whitney test. F0: Fluorescence intensity at baseline, ΔF: Fluorescence intensity per time point minus fluorescence intensity at baseline.

5. Supplementary Figure 5. Insoluble pathological tau level PHF1 correlates with the degree of Ca²⁺ activity abnormalities.

A-F. Correlation analysis of insoluble pathological tau PHF1 levels with different Ca²⁺ parameters under resting and running condition in JNPL3 tauopathy mice. Under resting condition, total Ca²⁺ activity (AUC) correlated significantly with insoluble PHF1 levels (C, R² = 0.369, p = 0.0212). Under running condition, insoluble PHF1 levels correlated significantly with frequency of the Ca²⁺ transients, (D, R² = 0.306, p = 0.0400). R is Pearson correlation coefficient. n = 7 from each group were analyzed.

6. Supplementary Figure 6. Reduced expression of AMPA receptor subunits and synaptophysin in tauopathy mice.

Synaptosome protein was extracted from JNPL3 and WT mouse brains and analyzed on western blots. A. Western blots stained for GluR1, GluR2, NR1, synaptophysin, PSD-95 and GAPDH (loading control). B-F. Quantification of protein levels in A. AMPA receptor subunits synaptophysin (B), GluR1 (D) and GluR2 (E) are decreased in JNPL3 mice compared to WT mice but NMDA receptor subunit NR1 (F) or postsynaptic marker PSD-95 (C) do not differ between the two groups. Data in B-F are presented as mean ± SEM, n = 6 for each group. *: p < 0.05, unpaired t test.

7. Supplementary Figure 7. Expression of kainate receptor subunit (GluK2) or L-type Ca²⁺ channel (Ca_v 1.2) was comparable in tauopathy vs wild-type mice.

Synaptosome protein was extracted from JNPL3 and WT mouse brains and analyzed on western blots. A. Western blots stained for GluK2, Ca_v 1.2 and GAPDH (loading control). B-D. Quantification of protein levels in A. Kainate receptor subunit GluK2 (B), L-type Ca²⁺ channel (Ca_v 1.2; C) and GAPDH levels (D) were comparable in JNPL3 and wild-type mice. Data in B-D are presented as mean ± SEM, n = 6 for each group.

8. Supplementary Figure 8. Tau mAb decreases soluble pathological tau but cannot reverse reduction of AMPA receptor subunits and synaptophysin.

A. Western blots for soluble protein fraction stained with PHF1, and synaptosome protein stained for GluR1, GluR2, NR1, synaptophysin, PSD-95 and GAPDH (loading control). B. 4E6 decreased soluble pathological PHF1 immunoreactive tau after 4E6 injections (p = 0.0315). C-G. Synaptosome protein was extracted from JNPL3 after two femoral vein injections of IgG (n = 6) or 4E6 (n = 7) and analyzed on western blots. 4E6 had no effects on AMPA receptor subunit GluR1 (C), GluR2 (D), NMDA receptor subunit NR1 (E), synaptophysin (F) and PSD-95 (G) expression. All data are presented as mean ± SEM, *: p < 0.05, unpaired t test.

9. Supplementary Table 1. Neuronal Ca²⁺ activity per neuron in WT and tauopathy mice under resting and running conditions.

Quantitative summary of neuronal Ca²⁺ activity in WT and tauopathy mice with/without antibody administration under resting and running conditions. Data is presented as average ± standard error of the mean. TRT = treatment.

10. Supplementary Table 2. Neuronal Ca²⁺ activity per animal in WT and tauopathy mice under resting and running conditions.

Quantitative summary of neuronal Ca²⁺ activity in WT and tauopathy mice with/without antibody administration under resting and running conditions. Data is presented as average ± standard error of the mean. TRT = treatment.

Highlights.

• Ca²⁺ imaging revealed cortical neuronal dysfunction in awake tauopathy mice

• These neuronal abnormalities were more pronounced in running vs. resting animals

• Ca²⁺ dyshomeostasis correlated better with soluble than insoluble tau protein

• Tau antibody cleared soluble tau and partially rescued the neuronal deficits

Abbreviations:

Aβ

amyloid-β

AD

Alzheimer’s disease

AUC

area under the curve

BBB

blood-brain barrier

FSB

1-fluoro-2,5-bis (3-carboxy-4-hydroxystyryl) benzene

LSS

Low Speed Supernatant

mAbs

monoclonal antibodies

NFT

neurofibrillary tangle

SNs

synaptosomes

SP

sarkosyl pellet