key: cord-0253959-47fodbig
authors: Engstrom, Amanda K.; Walker, Alicia C.; Moudgal, Rohitha A.; Myrick, Dexter A.; Kyle, Stephanie M.; Bai, Yu; Rowley, M Jordan; Katz, David J.
title: The inhibition of LSD1 via sequestration contributes to tau-mediated neurodegeneration
date: 2020-10-14
journal: bioRxiv
DOI: 10.1101/745133
sha: 68a0856928b1a29ed2378e3dcab9376adb016dec
doc_id: 253959
cord_uid: 47fodbig

Tauopathies are a class of neurodegenerative diseases associated with pathological tau. Despite many advances in our understanding of these diseases, the direct mechanism through which tau contributes to neurodegeneration remains poorly understood. Previously, our lab implicated the histone demethylase LSD1 in tau-induced neurodegeneration by showing that LSD1 localizes to pathological tau aggregates in Alzheimer’s disease cases, and that it is continuously required for the survival of hippocampal and cortical neurons in mice. Here, we utilize the P301S tauopathy mouse model to demonstrate that pathological tau can exclude LSD1 from the nucleus in neurons. In addition, we show that reducing LSD1 in these mice is sufficient to highly exacerbate tau-mediated neurodegeneration and tau-induced gene expression changes. Finally, we find that overexpressing LSD1 in the hippocampus of tauopathy mice, even after pathology has formed, is sufficient to significantly delay neurodegeneration and counteract tau-induced expression changes. These results suggest that inhibiting LSD1 via sequestration contributes to tau-mediated neurodegeneration. Thus, LSD1 is a promising therapeutic target for tauopathies such as Alzheimer’s disease. SIGNIFICANCE STATEMENT We have made the novel discovery that pathological tau functions through the histone demethylase LSD1 in the Alzheimer’s disease pathway. Thus, we have identified a mechanism that links tau to the downstream neuronal dysfunction pathways. This step can potentially be targeted therapeutically, after the onset of dementia symptoms, to block the progression of dementia in Alzheimer’s disease patients.

Tauopathies such as corticobasal degeneration, progressive supranuclear palsy, and 55 frontotemporal lobar degeneration with tau inclusions are neurodegenerative diseases 56 pathologically defined by different forms of tau positive intraneuronal deposits (1) (2) (3) (4) (5) . In addition 57 to these primary tauopathies, neuropathological observations of postmortem Alzheimer's disease 58 (AD) brains show the presence neurofibrillary tangles (NFTs) of hyperphosphorylated tau 59 protein, as well as plaques containing β-amyloid (Aβ) peptide (6-9). AD is the leading cause of 60 age-related dementia, resulting from neuronal cell death in the frontal and temporal cortices, as 61 well as the hippocampus (7). As dementia progresses, the spatial pattern of tau pathology highly 62 correlates with the level of cognitive impairment (10-13). In addition, Aβ oligomers and/or 63 plaques can enhance tau pathology in various mouse models (14, 15) , and there is increasing 64 evidence that accumulation of Aβ plaques can contribute to tau pathology (3, 16, 17) . The most 65 well-defined physiological role of tau is in stabilizing microtubules, particularly in neuronal 66 axons (2). However, in the pathological state, tau becomes aberrantly phosphorylated (2, 18, 19) , 67 truncated (1, 4), and aggregates into oligomers and larger insoluble filaments (20, 21) . This 68 pathology is thought to trigger synaptic loss, dramatic genome-wide expression changes, 69 increased inflammatory response, and neuronal cell death (22) (23) (24) (25) . These data suggest that 70 pathological tau may be a downstream mediator of the neurotoxic effects leading to neuronal 71 degeneration in AD. 72 Previously, our lab demonstrated that deleting the histone demethylase Lsd1 in adult mice 73 leads to significant neuronal cell death in the hippocampus and cortex with associated learning 74 and memory defects (26) . In this mouse model, loss of Lsd1 induces genome-wide expression 75 changes that significantly overlap with those observed in the brains of postmortem human AD 76 cases, but not other neurodegenerative diseases, such as Parkinson's disease or amyotrophic 77 lateral sclerosis (ALS) cases. Consistent with this overlap, we observed LSD1 protein 78 mislocalized to cytoplasmic NFTs, but not associated with Aβ plaques in AD cases or Lewy 79 bodies of a-synuclein in Parkinson's disease cases (26). In control cases, LSD1 remains strictly 80 confined to the nucleus (26), due to it's well-defined nuclear localization signal (27) . These data 81 highlight the requirement for LSD1 in neuronal survival and suggest that the nuclear function of 82 the histone demethylase LSD1 could be disrupted by mislocalization to pathological tau. 83 To investigate how LSD1 may contribute to tau-mediated neurodegeneration, we utilized 84 the PS19 P301S tauopathy mouse model (hereafter referred to as PS19 Tau). PS19 Tau mice 85 express a P301S mutated form of the human tau protein, originally identified in a frontotemporal 86 dementia with parkinsonism (FTDP-17) patient, driven by the prion promoter throughout the 87 nervous system (28). When expressed in mice, the P301S tau protein is prone to 88 hyperphosphorylation and somatodendritic aggregation, without the presence of Aβ plaques.

PS19 Tau mice develop a heavy pathological tau burden and have been well characterized for the 90 temporal progression of tau pathology and disease-related phenotypes (29, 30) . However, the 91 mechanism of neuronal cell death caused by pathological tau is still unknown. 92 Here, we provide functional data that the inhibition of LSD1 function contributes to tau 93 induced neurodegeneration. We demonstrate in PS19 Tau mice that pathological tau sequesters 94 LSD1 in the cytoplasm of neurons throughout the brain. This results in depletion of LSD1 from 95 the nucleus. Additionally, we provide genetic and molecular evidence that pathological tau Overall, we observed sequestration of LSD1 in 6 out of 7 mice analyzed. In each of the 6 mice, 123 there were varying levels of sequestration ranging from LSD1 found in both the nucleus and that has little sequestration (denoted by the ^ in Fig. 1M ) has extensive tau pathology, but this 126 pathology is largely extracellular rather than cytoplasmic, indicating that neurons with extensive 127 LSD1 sequestration may have died and been cleared. Consistent with this, we observed a 20% reduction in transcript levels ( Fig. S2 B) and a 35% 142 reduction in protein levels (Fig. S2 C, D) in Lsd1 Δ/+ mice compared to their Lsd1 +/+ littermates.

143 Surprisingly, PS19 Tau mice have a 20% increase in LSD1 protein levels compared to Lsd1 +/+ 144 littermates. It is possible that this is due to some type of compensation because LSD1 is being 145 sequestered. Nevertheless, consistent with the reduction in LSD1 that we observe in Lsd1 Δ/+ 146 mice, PS19;Lsd1 Δ/+ mice have a similar 14% reduction in transcript levels ( Fig. S2 B) and a 31% 147 reduction in protein levels ( Fig. S2 C,D) compared to PS19 Tau littermates. This enables us to 148 directly assess the effect of reducing LSD1 on tau-induced neurodegeneration. In addition, all 149 genotypes were born at normal Mendelian ratios with equal male/female ratios.

To determine if reducing LSD1 protein levels accelerates tau-induced depletion of LSD1 of the population remaining 83 days earlier than PS19 Tau mice, and all but one of the last 25% 173 of PS19;Lsd1 Δ/+ mice died between 11.5-13.5 months, compared to 13.5-19 months in PS19 Tau 174 mice. As a result, 28% of PS19 Tau mice were still alive after all but one of the PS19;Lsd1 Δ/+ 175 mice had died ( Fig. 2A ).

PS19 Tau mice develop paralysis starting with hind limb clasping which progresses until 177 they are unable to feed (28). In our hands, PS19 Tau mice displayed intermittent hind limb 178 clasping starting at approximately 6 months of age. At 12 months, these mice had a severe clasp, 179 but were still mobile. This is delayed compared to what was originally reported by Yoshiyama 180 and colleagues (28). PS19;Lsd1 Δ/+ mice also displayed intermittent hind limb clasping beginning 181 at approximately 6 months of age. However PS19;Lsd1 Δ/+ mice became terminally paralyzed at a 182 faster rate compared to PS19 Tau mouse littermates. At 12 months, when PS19 Tau mice were 183 still mobile, PS19;Lsd1 Δ/+ mice were severely paralyzed and typically terminal (Movie S1). To 184 quantitatively assess paralysis we performed rotarod and grid performance tests. In the rotarod, 185 we assessed the ability of the mice to stay on the rotating rod (latency to fall) ( Tau pathology is not affected by change in LSD1 levels 234 Since LSD1 is a chromatin regulator, it is possible that reducing LSD1 protein levels affects the 235 PS19 Tau transgene. However, we confirmed that there was no difference between PS19 Tau 236 mice and PS19;Lsd1 Δ/+ mice in the endogenous mouse Mapt RNA expression, nor in the human 237 P301S MAPT transgene expression (Fig. S6 A) . It is also possible that LSD1 affects tau 238 pathology. To test this, we performed immunohistochemistry staining for a hyperphosphorylated 239 form of tau (AT8). As expected, we did not observe any AT8 positive staining in Lsd1 Δ/+ at 6, 8, The functional interaction between tau pathology and LSD1 inhibition is specific 253 To test the specificity of the functional interaction between tau pathology and LSD1, we Tau mice compared to Lsd1 +/+ (Fig S8 A,B) , and 295 significant gene expression changes 265 in PS19;Lsd1 Δ/+ mice compared to Lsd1 +/+ (Fig S8 C,D) . Importantly, Lsd1 Δ/+ mice had only 4 266 gene expression changes observed genome-wide (Fig. S8 E,F) , indicating that the partially 267 reduced level of LSD1 expression had very little effect on transcription on its own. This is 268 consistent with the lack of phenotype in these animals. 269 We first examined the relationship between tau-induced expression changes and the are very similar (Fig. S8 G-J) . However, the genes affected in both sets of mice tended to be 282 further exacerbated in PS19;Lsd1 Δ/+ mice compared to PS19 Tau mice. Amongst the 77% of 283 genes that changed in the same direction in both PS19 Tau mice and PS19;Lsd1 Δ/+ mice, 75% of 284 these transcripts had a higher fold-change in PS19;Lsd1 Δ/+ mice compared to PS19 Tau mice 285 (Fig. 4B) . Importantly, the gene expression pathways that are induced by the PS19 transgene 286 overlap substantially with those observed in human AD case (43). This suggests that reduction of 287 Lsd1 exacerbates human AD pathways. only the HA tag (hereafter referred to as PS19-HA inj). Additionally, to control for the effects of 297 viral injection, we injected Wild Type littermates with the HA only expressing virus (hereafter 298 referred to as WT-HA inj). All injections were performed directly into the hippocampus at 8-8.5 299 months, when tau pathology is already present throughout the nervous system. Immunolabeling 300 for the HA tag demonstrated that the virus is specific to NeuN+ neurons (Fig. S9 A-D) , with no 301 HA expression observed in IBA+ microglia (Fig. S9 E-H) , or GFAP+ astrocytes (Fig. S9 I-L) . It 302 also confirmed that virally expressed LSD1 is nuclear (Fig. S9 M) and confined to the 303 hippocampus ( Fig. S9 N) , even at 11-11.5 months, when LSD1 is normally becoming 304 sequestered to the cytoplasm (Fig. S9 O-R) . After 3 months of overexpression, 11-11.5 month 305 old mice were euthanized, and the brains were analyzed. Injections resulted in a ~6-fold increase 306 in expression of LSD1 in the hippocampus compared to endogenous LSD1 in the PS19-HA inj 307 mice, but no increase in the cerebellum (Fig. S9 S,T) . As expected, because the viral injections 308 were restricted to the hippocampus, the mice injected with LSD1 still developed paralysis 309 (Movie S3). This confirms that the tau transgene is expressed and functioning. Additionally, we 310 did not observe a difference in total levels of AT8 positive tau immunoreactivity (Fig. S9 U-X) .

Therefore, any modulation to the phenotype was not due to changes in tau pathology. 

Of note, the one PS19-LSD1 inj mouse where we did observe increased glial cells, similar to 329 PS19-HA inj mice, had the lowest neuronal cell count (74% of WT-HA inj CA1 neurons). It is 330 possible that this mouse was already undergoing neurodegeneration prior to the injection of the 331 LSD1 overexpression virus.

Although the number of hippocampal neurons in PS19-LSD1 inj mice did not differ 333 from WT-HA inj controls, in 6 of the 10 PS19-LSD1 inj mice we observed cells with abnormal 334 blebbed nuclei at varying numbers throughout the hippocampus (Fig. 5E ). These abnormal cells 335 are rare in PS19-HA inj mice, which have a reduced overall number of cells in the pyramidal 336 cell layer compared to WT-HA inj control mice. One possibility is that these abnormal cells 337 with blebbing nuclei represent an intermediate state between a healthy neuron and a dying 338 neuron that is prolonged by rescue via LSD1 overexpression. Interestingly, these abnormal cells 339 also differed in the localization of HA-tagged LSD1. The four mice with normal nuclei had HA-340 tagged LSD1 protein localized uniformly throughout the nucleus (Fig. 5F ). In contrast, the six 341 mice with abnormal nuclear blebbing had some HA-tagged LSD1 that was mislocalized to the 342 cytoplasm (Fig. 5G) . This includes the one PS19-LSD1 inj mouse that had an elevated number 343 of astrocytes and TRL2 positive microglia. Thus, the blebbing state correlates with when the 344 viral produced LSD1 begins to be sequestered in the cytoplasm, similar to the endogenously 345 produced LSD1.

To test the specificity of the rescue that we observe when LSD1 is overexpressed in the 347 hippocampus of PS19 Tau mice, we performed RNA sequencing on the hippocampus of rescued 348 11-11.5 month old PS19-LSD1 inj mice versus PS19-HA inj mice. The RNA-seq analysis 349 detected 144 significant gene expression changes PS19-HA inj mice compared to WT-HA inj 350 control mice (Fig. S11 A,B) , and 57 significant gene expression changes in PS19-LSD1 inj mice 351 compared to WT-HA inj control mice (Fig. S11 C,D) . All of the 144 significant gene expression 352 changes induced by tau are ameliorated by overexpression of LSD1 (Fig. 6A) . As a result, 116 353 out of the 144 significantly changed genes in the PS19 Tau-HA mice are no longer significantly 354 changed in the rescued PS19-LSD1 inj mice. Based on this specific rescue of the differentially 355 expressed genes induced by tau, we further compared the expression changes between PS19-356 LSD1 inj mice and PS19-HA inj mice genome-wide (Fig. 6B) . Amongst the genome-wide 357 expression changes in PS19-HA inj mice compared to WT-HA inj control mice, 76% were 358 changed in the same direction (either up or down), and of these 69% were ameliorated. It should 359 be noted, that this is the opposite of what we observe when we reduce LSD1 (Fig. 4) . Thus, tau- In this study we investigate a potential downstream mediator of tau pathology in 368 neurodegenerative disease. We find that modulation of the chromatin modifying enzyme LSD1 369 can alter neurodegeneration in a tauopathy mouse model. Previously, we showed that LSD1 370 colocalizes with tau pathology in the cell body of neurons in AD cases (26). This suggested that 371 LSD1 might be disrupted in tauopathies such as AD, by being excluded from the nucleus. To 372 address this directly, we utilize the PS19 tauopathy mouse model. In these mice, we find that 373 LSD1 is sequestered in the cytoplasm, in some cases being completely depleted from the 374 nucleus. This provides the first cytological evidence that pathological tau can prevent LSD1 from 375 properly localizing to the nucleus in hippocampal and cortical neurons, where we have 376 previously shown it is continuously required (44).

Based on the ability of pathological tau to sequester LSD1, we hypothesized that 378 neuronal cell death may be due, at least partly, to LSD1 being sequestered in the cytoplasm and 379 depleted from the nucleus. In this case, reducing LSD1 levels should make it easier for tau to 380 deplete LSD1 from the nucleus, resulting in a faster progression of neurodegeneration and/or a 381 more severe neurodegenerative phenotype. We find that reducing LSD1 in the PS19 Tau Our previous data implicated LSD1 in the tau-mediated neurodegeneration pathway.

Utilizing the PS19 Tau mouse model, we now show a functional interaction between 419 pathological tau and LSD1. Importantly, because the tau mutation utilized in PS19 Tau mice 420 comes from an FTD patient, the functional interaction that we detected may also be relevant to To address this, we overexpressed LSD1 in the hippocampal neurons of PS19 Tau mice.

Overexpression of LSD1 specifically in hippocampal neurons rescues the neuronal cell death and 442 limits the inflammatory response. This rescue is neuronal specific, suggesting that the functional 443 interaction between LSD1 and tau is occurring in neurons. In addition, this rescue occurs despite 444 there being no effect on tau aggregation. This negates the possibility that the tau transgene is 445 simply not functioning when LSD1 is overexpressed. The ability of LSD1 overexpression to 446 overcome tau-mediated neurodegeneration in the presence of pathological tau aggregates, 447 provides further evidence that pathological tau is functioning through the inhibition of LSD1.

A priori, we would not expect the gene expression changes induced by overexpression of 449 LSD1 to be related to the gene expression changes induced by pathological tau. However, the 450 data presented here, along with our previous work (44) likely be necessary to permanently disrupt the interaction between pathological tau and LSD1.

This work is currently ongoing in the lab. Overall, our data establish LSD1 as a major 469 downstream effector of tau-induced neurodegeneration. Based on these data, we propose that the 470 LSD1 pathway is a potential late stage target for intervention in tauopathies, such as AD. bring deposited in the GEO and will be available upon publication. Results from DESEQ2 are 508 available in Data S1. Correspondence and requests for materials should be addressed to D.J.K.

(djkatz@emory.edu). 

A Brief Overview of Tauopathy: Causes

Trends in pharmacological sciences 38

Tau in physiology and pathology

Tau: From research to clinical development

Isolation of a fragment of tau derived from the core of the paired 520 helical filament of Alzheimer disease

Cloning and 522 sequencing of the cDNA encoding a core protein of the paired helical filament of 523 Alzheimer disease: identification as the microtubule-associated protein tau

The amyloid hypothesis of Alzheimer's disease: progress and 526 problems on the road to therapeutics

Amyloid cascade hypothesis: Pathogenesis and 528 therapeutic strategies in Alzheimer's disease

The Trigger and Bullet in Alzheimer Disease 530

Classification and basic pathology of Alzheimer 533 disease

Neurofibrillary 535 tangles but not senile plaques parallel duration and severity of Alzheimer's disease

The amyloid cascade hypothesis for Alzheimer's 538 disease: an appraisal for the development of therapeutics

Neuronal loss correlates with but exceeds neurofibrillary tangles in 541

Alzheimer's disease

Tau pathology and neurodegeneration contribute to cognitive 543 impairment in Alzheimer's disease

A{beta} accelerates the spatiotemporal progression of tau pathology 545 and augments tau amyloidosis in an Alzheimer mouse model

Amyloid-beta plaques enhance Alzheimer's brain tau-seeded pathologies by 548 facilitating neuritic plaque tau aggregation

A three-dimensional human neural cell culture model of Alzheimer/'s 550 disease

Frequency of Stages of Alzheimer-Related Lesions in Different Age 552

New 554 phosphorylation sites identified in hyperphosphorylated tau

Alzheimer's disease brain using nanoelectrospray mass spectrometry

Hyperphosphorylation of tau in PHF

Traumatic Brain Injury, and 562 Diagnostic Challenges

Altered expression of synaptic proteins occurs early during progression 564 of Alzheimer's disease

Early Abeta accumulation and progressive synaptic loss, gliosis, and 566 tangle formation in AD brain

Inflammation as a central mechanism in Alzheimer's disease

Alzheimer's & Dementia

A review: inflammatory process in Alzheimer's 571 disease, role of cytokines

LSD1 protects against hippocampal and cortical 573 neurodegeneration

Nuclear import of human histone lysine-specific demethylase LSD1

Synapse Loss and Microglial Activation Precede Tangles in a 577 P301S Tauopathy Mouse Model

Induction of Inflammatory Mediators and Microglial Activation in Mice 579

Transgenic for Mutant Human P301S Tau Protein

Tau pathology spread in PS19 tau transgenic mice following locus 582 coeruleus (LC) injections of synthetic tau fibrils is determined by the LC's afferent and 583 efferent connections

Generation of a germ cell-specific 585 mouse transgenic Cre line, Vasa-Cre

Opposing LSD1 complexes function in developmental gene activation 588 and repression programmes

An epigenetic trap stabilizes singular olfactory receptor expression

Lysine-specific demethylase 1 regulates the embryonic transcriptome 592 and CoREST stability

Loss of LSD1 (lysine-specific demethylase 1) suppresses growth and alters 594 gene expression of human colon cancer cells in a p53-and DNMT1(DNA 595 methyltransferase 1)-independent manner

Aberrant stress-induced phosphorylation of perikaryal 598 neurofilaments

Alternation of neurofilaments in immune-mediated injury of spinal cord 600 motor neurons

Chromatolysis: Do injured axons regenerate poorly when ribonucleases 602 attack rough endoplasmic reticulum, ribosomes and RNA?

What is the signal for chromatolysis?

Effects of naproxen on immune responses in 606 a colchicine-induced rat model of Alzheimer's disease

Compensatory Motor Neuron Response to Chromatolysis in the Murine 609 hSOD1(G93A) Model of Amyotrophic Lateral Sclerosis

Lysosomal abnormalities in degenerating 612 neurons link neuronal compromise to senile plaque development in Alzheimer disease

Meta-Analysis of the Alzheimer's Disease Human Brain Transcriptome 615 and Functional Dissection in Mouse Models

LSD1 protects against hippocampal and cortical 617 neurodegeneration

Alzheimer's Disease

Nuclear Pore Complex 621 Proteins in Alzheimer Disease

The Galaxy platform for accessible, reproducible and collaborative 626 biomedical analyses: 2016 update

The Galaxy platform for accessible, reproducible and collaborative 628 biomedical analyses

Marc 631 Schwartz, Bill Venables

WEB-based GEne SeT AnaLysis Toolkit 634 (WebGestalt): update 2013

WebGestalt 2017: a more 636 comprehensive, powerful, flexible and interactive gene set enrichment analysis toolkit

WebGestalt 2019: gene set analysis 639 toolkit with revamped UIs and APIs

WebGestalt: an integrated system for exploring gene sets 641 in various biological contexts

Lifespan curve showing that 665 no Lsd1 Δ/+ mice died before 18 months (orange, n=20). PS19;Lsd1 Δ/+ mice (purple, n=44) have a significant 666 reduction in survival compared to PS19 Tau mice with wild-type levels of Lsd1 (green, n=37)(Log-rank Mantle-Cox 667 test ***P<0.005). B-D, Rotarod testing of latency to fall

and distance traveled (in centimeters) (D) for mice at age 6, 8 and 10 months. Lsd1 Δ/+ (orange, n=10

Values are mean ± SEM (two-way analysis of 670 variance (ANOVA) with Tukey's post hoc test *

Reduction of Lsd1 exacerbates neurodegeneration in PS19 Tau mice. A,B, Average nuclei per area in the 674 CA1 (A) and CA3 (B) regions of the hippocampus in 12 month old Lsd1

Lsd1 D/+ 675 mice. Quantification from histology represented in Fig. S5 A-H. Values are mean ± SD (A, n=13 & B, n=9). C

Representative image of the brains of 12 month old Lsd1 D/+ , PS19 Tau, and PS19

For all graphs: one-way 678 analysis of variance (ANOVA) with Tukey's post hoc test (two-sided) *P<0.05,****P<0.001. E-J, Representative 679 image of T2-weighted RARE coronal MRI taken from 6 months (E-G) and 10 months (H-jJ) of age

Arrow denotes region of hippocampal atrophy

Molecular overlap between loss of LSD1 function and tauopathy. A, Histogram (log2 fold change) of the 684 112 genes that have significant changes in gene expression in the PS19 Tau mice (green, n=4) and their 685 corresponding gene expression change in PS19

B, Scatter plot showing the correlation 686 between the genome-wide log2 fold change in gene expression between PS19 Tau and PS19;Lsd1 Δ/+ . The most 687 significantly changed genes in PS19

All 688 other genes are shown in grey. Dotted line represents 1:1 relationship between gene expression changes in PS19 Tau 689 vs

Exacerbated genes fall to the right of the dotted line in the positively correlated quadrant and to 690 the left of the dotted line in the negatively correlated quadrant. Genes with correlated expression changes are found 691 in the top right and bottom left quadrants

Square brackets 697 denote thickness of pyramidal layer of the CA1 of the hippocampus and curvy brackets denote hippocampal region 698 with or without infiltrating cells. D, Quantification of the average number of nuclei in the pyramidal layer of the 699 hippocampus per area per mouse from histology represented in Fig. 5A-C. Values are mean ± SD (n=10, one-way 700 analysis of variance (ANOVA) with Tukey's post hoc test, **p<0.01, ns=not significant). E, Representative H&E of 701 PS19-LSD1 inj mouse with abnormal nuclei blebbing in the CA1 region of the hippocampus

LSD1) in 11 month PS19-LSD1 inj mice. HA is 704 either localized specifically to the nucleus in all nuclei (F) or in only a few nuclei while it is partially sequestered in 705 the cytoplasm in others (G). F'-F''', High magnified image of cells denoted by arrows in Fig. 5F of nuclear HA 706 localization in individual nuclei

Histogram (log2 fold change) of the 711 144 genes that have significant changes in expression in the PS19 Tau-HA inj mice (green, n=3) and their 712 corresponding gene expression changes in the PS19 Tau-LSD1 inj mice (blue, n=3). B, Scatter plot showing the 713 correlation between the genome-wide log2 fold change in gene expression between

All other genes are shown in grey. Dotted line represents 1:1 relationship between gene 716 expression changes in PS19 Tau-Ha inj and PS19 Tau-LSD1 inj. Rescued genes fall to the left of the dotted line in 717 the positively correlated quadrant and to the right of the dotted line in the negatively correlated quadrant. Genes with 718 correlated expression changes are found in the top right and bottom left quadrants