Tau is required for progressive synaptic and memory deficits in a transgenic mouse model of α-synucleinopathy

Parkinson’s disease dementia (PDD) and dementia with Lewy bodies (DLB) are clinically and neuropathologically highly related α-synucleinopathies that collectively constitute the second leading cause of neurodegenerative dementias. Genetic and neuropathological studies directly implicate α-synuclein (αS) abnormalities in PDD and DLB pathogenesis. However, it is currently unknown how αS abnormalities contribute to memory loss, particularly since forebrain neuronal loss in PDD and DLB is less severe than in Alzheimer’s disease. Previously, we found that familial Parkinson’s disease-linked human mutant A53T αS causes aberrant localization of the microtubule-associated protein tau to postsynaptic spines in neurons, leading to postsynaptic deficits. Thus, we directly tested if the synaptic and memory deficits in a mouse model of α-synucleinopathy (TgA53T) are mediated by tau. TgA53T mice exhibit progressive memory deficits associated with postsynaptic deficits in the absence of obvious neuropathological and neurodegenerative changes in the hippocampus. Significantly, removal of endogenous mouse tau expression in TgA53T mice (TgA53T/mTau−/−), achieved by mating TgA53T mice to mouse tau-knockout mice, completely ameliorates cognitive dysfunction and concurrent synaptic deficits without affecting αS expression or accumulation of selected toxic αS oligomers. Among the known tau-dependent effects, memory deficits in TgA53T mice were associated with hippocampal circuit remodeling linked to chronic network hyperexcitability. This remodeling was absent in TgA53T/mTau−/− mice, indicating that postsynaptic deficits, aberrant network hyperactivity, and memory deficits are mechanistically linked. Our results directly implicate tau as a mediator of specific human mutant A53T αS-mediated abnormalities related to deficits in hippocampal neurotransmission and suggest a mechanism for memory impairment that occurs as a consequence of synaptic dysfunction rather than synaptic or neuronal loss. We hypothesize that these initial synaptic deficits contribute to network hyperexcitability which, in turn, exacerbate cognitive dysfunction. Our results indicate that these synaptic changes present potential therapeutic targets for amelioration of memory deficits in α-synucleinopathies. Electronic supplementary material The online version of this article (10.1007/s00401-019-02032-w) contains supplementary material, which is available to authorized users.

Suppl. Figure 1 -Aged TgWT and TgA30P mice do not display deficits in long-term spatial learning and memory via Barnes Maze. a. Depiction of animals used: wild-type animals (nTg) do not express human aS. TgWT animals are heterozygous for a transgene that expresses wild-type human aS. TgA30P animals are heterozygous a transgene that expressed human mutant A30P aS. b. Barnes maze (BM) training trials for 12month-old nTg, TgA30P, and TgWT animals demonstrating duration of training trials per group during each of the four training days. TgWT training: two-way repeated measures ANOVA with Geisser-Greenhouse correction and Sidak's posthoc analysis revealed a significant effect of training (F (2.545,40.72) = 36.09, p < 0.0001), no effect of haS WT genotype (F (1,16) = 3.249, p = 0.0903), and no significant training day* haS WT interaction (F (3,48) = 0.1449, p = 0.9325). TgA30P training: two-way repeated measures ANOVA with Geisser-Greenhouse correction and Sidak's posthoc analysis revealed a significant effect of training (F (1.838,29.40) = 143.9, p < 0.0001), no effect of haS A30P genotype (F (1,16) =2.806, p = 0.1113), and no significant training day* haS A30P interaction (F (3,48) = 1.645, p = 0.1915). c. Time (seconds) spent occupying the target (goal) quadrant of the Barnes maze during the probe trial. Unpaired t test with Welch's correction. TgWT: t = 1.472, df = 14.91; TgA30P: t = 1.439, df = 15. TgWT: N = 9 animals/genotype. TgA30P: n nTg = 7; n TgA30P = 9. d. Left column: probe test BM occupancy heat maps obtained by averaging the location of all animals in each genotype and cohort. Right column: representative individual animal traces tracking movement during the BM probe test. BM orientation shown in Fig 1c, yellow shading indicates goal quadrant. The findings here demonstrate that postsynaptic deficits, as observed in TgA53T mice but not in TgWT or TgA30P, are required for cognitive decline by 12 months of age as TgWT and TgA30P do not display such deficits at this time point. ns: not significant. Error bars represent mean ± standard error of the mean (S.E.M).
Suppl. Figure 2 -Barnes Maze quadrant analysis demonstrating progressive deficits in long-term spatial learning and memory in TgA53T mice in the absence of overt locomotor deficits. a and b. Analysis of time animals spent in each quadrant of the Barnes Maze (BM) during the probe test of the BM at 6 months (6M) (a) and 12 months (12M) (b). 12M BM goal quadrant: F (3,29) = 10.47, p < 0.0001, one-way ANOVA with Tukey's posthoc analysis. c and d. Total distance traveled on the BM by animals during the probe test at 6M (c) and 12M (d). 12M: F (3,31) = 0.9377, p = 0.4347 by one-way ANOVA with Tukey's posthoc analysis. 6M: n nTg = 13; n TgA53T = 11; n TgA53T/mTau-/-= 10; n mTau-/-= 8. 12M: n nTg = 9; n TgA53T = 7; n TgA53T/mTau-/-= 11; n mTau-/-= 8. e. BM diagram for testing and probe trial, with yellow shading indicating goal quadrant and dark grey showing escape box location in that quadrant. f. Representative individual animal traces tracking movement during the BM probe test. Yellow shading indicates goal quadrant. These results demonstrate that TgA53T mice have age-dependent deficits in spatial learning and memory that are dependent on endogenous mouse tau expression and precede onset of locomotor abnormalities. One-way ANOVA: *** p < 0.001. ns: not significant. Error bars represent mean ± S.E.M. Figure 3 -TgA53T mice show impaired short-term spatial learning and memory in a tau-dependent manner. a. Diagram depicting the Y maze (YM) orientation and experimental testing paradigm. YM testing was divided into two discrete phases: first, a Learning trial (Novel arm blocked off), followed by a Recognition test (all arms open, including Novel). Walls at the end of each arm were marked by different patterns: start with triangles, familiar with square checkerboard, and novel with stripes. b. Time spent in novel (N) and familiar (F) arms of the Y maze during the 300-second-long recognition trial. Unpaired t test with Welch's correction was used compare time animals spent in N versus F arms during recognition trial within genotype. nTg: t = 5.967, df = 22, p < 0.0001; TgA53T: t = 0.9192, df = 20.40, p = 0.3687; TgA53T/mTau -/-: t = 3.425, df = 18.34, p = 0.0030; mTau -/-: t = 4.291, df = 11, p = 0.0011. c. Left panel: average speed (meters/second) of mice during the recognition trial: F (3,39) = 2.609, p = 0.0652, one-way ANOVA with Tukey's posthoc analysis. Right panel: total distance traveled (meters) by animals during the recognition trial. F (3,39) = 2.632, p = 0.0635, one-way ANOVA with Tukey's posthoc analysis. d. Top row: heat maps demonstrating time all animals tested spent occupying areas of the YM during the recognition trial. Bottom row: individual trace of representative animal for each group tested during the recognition trial. Orientation of YM is preserved from diagram in a. Taken together, this experiment demonstrates that TgA53T mice present with tau-dependent deficits in spatial learning and memory via YM prior to developing deficits that can be detected via Barnes Maze. t test and one-way ANOVA: ** p < 0.01 and **** p < 0.0001. ns: not significant. Error bars represent mean ± S.E.M.
Suppl. Figure  animals/genotype. Densitometry shows that while b-synuclein and synapsin isoforms are decreased in TgA53T mice, they are not altered by loss of tau expression. Further, the levels of synaptophysin and PSD95 are comparable in all animals, indicating a lack of synaptic loss. One-way ANOVA: * p < 0.05, *** p < 0.001, and **** p < 0.0001. Error bars represent mean ± S.E.M.

Suppl.
Suppl. Figure 9 -Tau-dependent biochemical changes in postsynaptic glutamatergic (AMPA and NMDA) receptors are region-specific. a. Representative western blot images for AMPAR (GluA) and NMDAR (GluN) subunits of cortical lysates from 12-month-old animals. b. Densitometry quantifying protein expression in isolated cortices. For all, values were normalized to the average densitometric values of nTg samples within each gel. For all densitometry: one-way ANOVA with Tukey's posthoc analysis. GluN1: F (3,20) = 11.49, p < 0.0001. Compared to nTg mice cortices, cortical AMPA and NMDA receptor expression is not decreased in TgA53T mice at 12 months of age, when TgA53T mice have established synaptic deficits, evidence of network abnormalities, and memory loss. One-way ANOVA: * p < 0.05, ** p < 0.01, and *** p < 0.001. Error bars represent mean ± S.E.M.
Suppl. Figure 10 -Tau-dependent synaptic dysfunction in TgA53T neurons is not associated with increases in PrP expression and signaling. a. Graphical depiction of cellular prion protein (PrP C ) primary structure and antibody binding regions. b and c. Representative western blot images of hippocampal (b) and cortical (c) lysates for PrP C expression at 3, 6 and 12 months of age (3M, 6M, and 12M, respectively). d and e. Densitometry of protein expression in hippocampal (HIP) (d) and cortical (CTX) (e) lysates, normalized to the average densitometric values of nTg samples within each gel. For 3M and 6M densitometry, unpaired t test with Welch's correction. For 12M densitometry: one-way ANOVA with Tukey's posthoc analysis. 12M HIP 6D11: F (3,20) = 3.193, p = 0.0458. 12M HIP 8B4: F (3,20) = 3.685, p = 0.0292. These results demonstrate that activation of PrP C signaling is not required for the age-dependent synaptic and memory deficits in TgA5T mice. t test and one-way ANOVA: * p < 0.05. Error bars represent mean ± S.E.M. Figure 11 -Tau-mediated, mutant aS-induced synaptic and cognitive deficits are independent of postsynaptic Fyn and GluN2B activation. a and b. Western blot images for activation status of both Fyn (via pTyr416 Src) (a) and NMDA receptor subunit GluN2B (via pTyr1472 GluN2B) (c) 3 (left) and 6 (right) months of age (3M and 6M, respectively) through analysis of total and active phosphorylated states. b and d. Densitometry quantifying 3M and 6M cortical lysates, normalized to the average densitometric values of nTg samples within each gel. Activation status was determined through examining ratio of values of activating phosphorylation site: pTyr416 for Fyn (b) and pTyr1472 for GluN2B (d), normalized to respective total levels of protein. For 3M and 6M densitometry, unpaired t test with Welch's correction. 3M pTyr-Src/Fyn: t = 1.761, df = 8, p = 0.1163. 3M pTyr-GluN2B/GluN2B: t = 2.539, df = 8.066, p = 0.0345. These findings build on Suppl. Figure 9, further showing that TgA53T mice do not display increased activation of the PrP C -Fyn-GluN2B signaling cascade as compared to agematched nTg littermates. Together, these results suggest that haS A53T -driven and impairments requiring tau are mediated independent of this pathway. t test: * p < 0.05. ns: not significant. Error bars represent mean ± S.E.M.

Suppl. Figure 12 -Activity-dependent remodeling of hippocampal circuits is not observed in younger
TgA53T mice with intact cognition and appears to follow glutamatergic signaling deficits. a and b. Representative confocal images from dentate gyri and hippocampi of 3-month-old (3M, a) and 6-month-old (6M, b) nTg and TgA53T mice stained for c-Fos, NPY, and calbindin. 3M and 6M c-Fos, NPY, and calbindin quantification values are reported in Fig 9c-e. c-Fos scale bar: 300 µm. NPY and calbindin scale bar: 250 µm. c-e. Quantification of immunoreactivity (IR) via cell counting (c: c-Fos) or densitometry (d: NPY in the Molecular Layer "Molecular", ML; NPY in the Mossy Fiber pathway, "Mossy", MF; and, e: calbindin) at 6M. 6M TgA53T animals were classified into two groups as some TgA53T mice displayed activity-related changes in c-Fos, calbindin, and NPY while others, the TgA53T INT (for "Intermediate") resembled their age-matched nTg littermate controls. c-Fos: F (2,9)= 22.16, p = 0.0003; NPY-Molecular: F (2,9) = 10.60, p = 0.0043; NPY-Mossy: F (2,9) = 20.74, p = 0.0004; Calbindin: F (2,9) = 4.174, p = 0.0522. All by one-way ANOVA with Tukey's posthoc analysis. n nTg = 6; n TgA53T = 3; n TgA53T-INT = 3. f-h. Three-dimensional (3D) X-Y-Z scatterplot of c-Fos, NPY (Mossy Fiber, "NPY-Mossy"), and calbindin immunostaining at 3 (f), 6 (g), and 12 (h) months (3M, 6M, and 12M, respectively). 3M TgA53T mice do not display evidence of network remodeling. The two distinct TgA53T populations present at 6M are clearly observed: one TgA53T group displayed activity-related changes in c-Fos, calbindin, and NPY while others, the TgA53T INT (for "Intermediate") more closely resembled their age-matched nTg littermate controls. By 12M, all aged TgA53T mice demonstrate signs of chronic hippocampal network hyperactivity. One-way ANOVA: * p < 0.05, ** p < 0.01, and *** p < 0.001. ns: not significant. Error bars represent mean ± S.E.M. Figure 13 -Proposed model for aS-mediated, tau-dependent synaptic and cognitive deficits in haS A53T mice. Neurons from cognitively-intact nTg control mice display intact AMPA receptor (AMPAR) expression and function, and thus display intact synaptic transmission. The presence of pathological species of aS, such as haS A53T in TgA53T mice leads to progressive postsynaptic synaptic deficits that underlie memory deficits in this model, and potentially PDD or DLB patients, without overt structural alterations at the synaptic or neuronal level. Pathogenic aS (haS A53T ) expression or accumulation (1) leads to tau phosphorylation and mislocalization to dendritic spines (2), producing physiological deficits in AMPAR -mediated signaling and synaptic plasticity, likely through reductions in AMPAR expression (3). These haS A53T -mediated AMPAR deficits culminate in reduced glutamatergic signaling, particularly at hippocampal pyramidal neuron synapses. The AMPAR deficits lead to more global, circuit-level abnormalities in the brains of TgA53T mice. For example, it is possible that haS A53T -mediated, tau-dependent synaptic depression may also affect inhibitory interneurons, leading to homeostatic imbalance between inhibitory and excitatory neurotransmission and aberrant excitatory activity. These epileptiform changes can then lead to homeostatic responses in hippocampal circuits that attempt to suppress this hyperactivity, representing a potential mechanism underlying the hippocampal remodeling observed in the brains of TgA53T mice. However, future studies are warranted to better elucidate the biochemical, physiological, and structural mechanisms connecting haS A53T -mediated synaptic dysfunction and aberrant network changes. Ultimately, we hypothesize that haS A53T -driven, tau-mediated abnormalities at the individual synapse lead to network-level perturbations that together contribute to and exacerbate memory deficits in TgA53T mice. PDD: Parkinson's disease dementia. DLB: dementia with Lewy bodies. Pre: presynaptic neuron. Post: postsynaptic neuron. A: AMPAR. a: haS A53T .