Amyloid exacerbates tau and alpha‐synuclein pathologies, behavioral impairments, and neuroinflammation in a mixed dementia model

Abstract

INTRODUCTION

Neurodegenerative diseases often involve overlapping alpha‐synuclein (asyn), amyloid beta, and tau proteinopathies, yet the mechanisms, impact, and directionality of their interactions remain unclear.

METHODS

We induced brain‐wide neuronal asyn/tau pathologies via viral expression of wild‐type asyn, mutant asyn^E46K, mutant tau^A152T, or both asyn^E46K/tau^A152T in adult amyloidosis knock‐in mice and controls, either post‐plaque deposition (6 months old) or pre‐plaque (3 months old). Open‐field behavior was assessed baseline and 3 and 6 months post‐transduction, followed by neuropathology and neuroinflammation analyses.

RESULTS

Post‐plaque induction in amyloid mice increased asyn/tau total and phosphorylated levels and exacerbated amyloid‐related hyperlocomotion/anxiety. Pre‐plaque induction produced robust phosphorylated pathologies irrespective of amyloid, while causing similar amyloid‐dependent behavioral synergy. Tau pathology drove LGALS3⁺ inflammatory glial responses in white‐matter fibers.

DISCUSSION

Amyloid context gates vulnerability, with certain synergies manifesting across stages. White‐matter gliosis is a novel mechanism of tau^A152T risk. Together, our data argue for the development of stage‐aware, multitarget interventions and biomarkers.

Highlights

• Novel models: AAV‐driven asyn/tau in APP^NL‐G‐F/MAPT amyloid mice.

• Amyloid stage dictates susceptibility to asyn and tau accumulation.

• Triple pathology heightens hyperlocomotion and anxiety‐like behaviors.

• Tau^A152T pathology triggers LGALS3‐related glial inflammation in white matter.

• Stage‐specific, multi‐target therapeutic strategies are needed in mixed dementia.

1. BACKGROUND

The high frequency of mixed pathologies in the aging brain highlights the challenge in understanding and treating heterogeneous neurodegenerative diseases. For instance, in brains meeting clinicopathological criteria for Lewy body dementia (LBD), alpha‐synuclein (asyn)‐containing inclusions called Lewy bodies and Lewy neurites, frequently co‐exist with Alzheimer's disease (AD) lesions, including amyloid beta (Aβ) plaques in 60% to 80% of cases and hyperphosphorylated tau aggregates in ∼40%. ¹ , ² , ³ , ⁴ , ⁵ Conversely, Lewy pathologies have been reported in 40% to 60% of cases with a clinical diagnosis of AD, in addition to the core AD pathologies of amyloid plaques and tau tangles. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ Furthermore, accumulation of Lewy bodies is a core feature of Parkinson's disease (PD) and other “pure” synucleinopathies. In addition to the clinicopathological similarities, some of the strongest genetic risk factors for LBD also impart risk for AD (e.g., apolipoprotein E4 or APOE ε4) as well as PD (e.g., glucosylceramidase beta or GBA, and alpha‐synuclein or SNCA). ⁹ , ¹⁰ , ¹¹ This convergence of clinicopathological and genetic disease features suggests that amyloid, tau, and asyn proteinopathies may cooperatively affect their accumulation, propagation, and clinical consequences, depending on timing and context.

Very few studies report the effects of mixed amyloid/asyn and amyloid/asyn/tau co‐pathologies, though existing data indicate that amyloid exacerbates asyn and tau accumulation, neuronal dysfunction, and behavioral decline. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ There are also conflicting reports showing minimal interactions or even protective effects of asyn on Aβ deposition, ¹⁷ , ¹⁸ underscoring the complexity of these relationships and the need for more translatable systems. A caveat on the usage of transgenic mouse models overexpressing APP is the impact of APP overexpression alone (independent of amyloid pathology). ¹⁹ , ²⁰ , ²¹ , ²² Furthermore, transgenic crosses can compound issues stemming from random insertion location/copy number, off‐target effects, and lack of temporal control. Pre‐formed asyn or tau fibril seeding onto aged APP transgenic backgrounds partially addresses these concerns but requires technically demanding surgeries and experimental variability (e.g., fibril purity). As a result, it remains unclear how amyloid pathology modulates susceptibility to asyn and tau pathologies, or the reverse.

Here, we designed novel adeno‐associated virus (AAV) constructs to induce robust and persistent asyn and tau pathologies in the brains of APP^NL‐G‐F/microtubule‐associated protein tau (MAPT) double knock‐in (KI) mice that express physiologically appropriate levels of APP, humanized mutant Aβ, and all human tau isoforms. ²³ It is important to note that human MAPT expression in this model does not result in tau tangle pathology in the absence of additional tau seeding (through transgenic, fibril, or viral means). AAV.CAP‐B10 ²⁴ capsid and neuron‐specific transcriptional regulatory elements were selected to enable temporally controlled, brain‐wide expression of wild‐type (WT) asyn^WT, mutant asyn^E46K, mutant tau^A152T, or combined asyn^E46K/tau^A152T. These AAV transgenes are anchored in human genetics relevant to LBD. SNCA duplication and triplication cause dominantly inherited parkinsonism and LBD, respectively, establishing a WT asyn dose‐severity relationship. ²⁵ , ²⁶ The SNCA E46K mutation causes LBD and promotes more aggregation‐prone asyn fibrils, serving as an accelerated pathological driver compared to WT asyn. ²⁷ , ²⁸ Lastly, the MAPT A152T variant increases the risk for tauopathies as well as LBD by altering phosphorylation states, microtubule stability, and axonal transport, making it an ideal candidate to probe the contribution of tau to mixed‐pathology states. ²⁹ , ³⁰ , ³¹ , ³²

Using this combined AAV and amyloid KI platform, we characterized the neuropathological and behavioral impacts of introducing pathological asyn, tau, or both, either before or after robust amyloid deposition and neuroinflammation begins (i.e., pre‐plaque vs post‐plaque). We quantified total and phospho‐protein levels, glial responses, amyloid plaque load, and longitudinal behavior. We found that amyloid potentiated susceptibilities and amyloid/asyn/tau pathologies interacted synergistically to influence behavior. Additionally, we found that tau pathology was distinctly capable of inducing a white‐matter inflammatory response.

2. METHODS

2.1. Animals

Breeder APP^NL‐G‐F/MAPT double KI (APP) and MAPT KI (control) mice ²³ were procured from the RIKEN Center for Brain Science under the necessary material transfer agreement. A large breeder colony was established from initial stock, such that the entirety of animals in a given cohort were born within 1 to 2 weeks of each other. Roughly equal proportions of male and female mice up to 12 months of age were used in this study. Animals were housed under standard conditions with ad libitum access to food and water in a specific pathogen‐free environment at Mayo Clinic Arizona. All animals used in this study were treated under an approved protocol from the Mayo Clinic Institutional Animal Care and Use Committee, adhering to the provisions of the Animal Welfare Act, PHS Animal Welfare Policy, National Institutes of Health Guide for Care and Use of Laboratory Animals, and the policies and procedures of the Department of Comparative Medicine at the Mayo Clinic.

RESEARCH IN CONTEXT

1. Systematic review: We extensively reviewed the literature for studies of mixed amyloid, tau, and asyn proteinopathies in AD, PD, LBD, and related mouse models (Boschen et al., 2025). We discovered a lack of understanding on how disease stage and timing influence mixed‐pathology interactions.

2. Interpretation: Combining temporally controlled AAV‐induced asyn and tau pathologies with a KI amyloidosis mouse model, we show that amyloid stage gates susceptibility to protein accumulation and phosphorylation, combined pathologies synergistically perturb locomotor and anxiety‐like behavior, and tau is capable of independently inducing white‐matter inflammation.

3. Future directions: Key open questions include whether reducing amyloid or proteostatic stress normalizes vulnerability to intracellular proteinopathies, how white‐matter glia and vulnerable axonal tracts map at single‐cell and spatial resolution, and whether this platform can guide stage‐aware, multitarget therapeutic and biomarker strategies for individuals with mixed amyloid, tau, and asyn pathologies.

2.2. Study design

The aim of this study was to identify the neuropathological, neuroinflammatory, and behavioral impact of introducing single asyn or tau co‐pathologies, as well as mixed co‐pathologies (asyn plus tau), into a naive or mature amyloid brain milieu. To that end, we developed and leveraged novel AAV constructs packaged with AAV.CAP‐B10 to induce robust and stable neuronal expression of human WT asyn^WT, human mutant asyn^E46K, human mutant tau^A152T (of the 4R2N isoform), both tau and asyn mutants (asyn^E46K/tau^A152T), or enhanced green fluorescent protein (EGFP) in the brains of pre‐plaque (3‐month‐old) or post‐plaque (6‐month‐old) APP and age‐matched control mice. Pre‐treatment (baseline), 3 months post‐injection (mpi), and 6 mpi (endpoint) we assessed open field behaviors. Endpoint histological outcomes included total protein, phospho‐protein, amyloid plaque, and inflammation levels in the brain. Animal numbers and statistical analyses are detailed in figure legends.

2.3. AAV transfer plasmid preparation

pAAV‐hSyn‐EGFP was a gift from Bryan Roth (Addgene, Catalog No.: 50465). The other AAV transfer plasmids were created by subcloning cDNA inserts encoding asyn^WT, asyn^E46K, and tau^A152T into the backbone of the pAAV‐hSyn‐EGFP plasmid, such that they were flanked upstream at the 5’ end by the human synapsin 1 promoter (hSyn) ³³ and downstream at the 3’ end by the woodchuck hepatitis virus post‐transcriptional regulatory element (WPRE) ³⁴ and human growth hormone polyadenylation signal post‐transcriptional stabilizing sequences (Supplementary Material, Figure S1). The resulting pAAV transfer plasmids containing our targeted transgenes were sequence confirmed and sterile filtered prior to transfection.

2.4. AAV preparation

HEK 293T cells (ATCC, Catalog No.: CRL‐3216) were cultured in DMEM (Gibco, Catalog No.:10569044) supplemented with 10% fetal bovine serum (FBS) (Corning, Catalog No.:MT35015CV), 1× MEM Non‐Essential Amino Acids (Gibco, Catalog No.: 11140050), and 1% penicillin–streptomycin (Gibco, Catalog No.: 15070063). Twenty‐four hours prior to transfection, cells were seeded in multi‐layer CellSTACK chambers (Corning, Catalog No.: 3310; 3311) at a density of 1.0 × 10⁵ cells/cm² in transfection medium composed of DMEM supplemented with 5% FBS, 1× MEM Non‐Essential Amino Acids, and 0.5% penicillin–streptomycin. Triple transfection was performed using PEI MAX (Polysciences, Catalog No.: 49553‐93‐7) at 0.93 µg/cm² and plasmid DNA at 0.26 µg/cm², in a 1:4:2 ratio of transfer:capsid:helper plasmids, following a published AAV production protocol we optimized for large‐scale productions. ³⁵ Transfer plasmids were prepared as described earlier, while the capsid and helper plasmids were pUCmini‐iCAP‐AAV.CAP‐B10 (gift from Viviana Gradinaru; Addgene, Catalog No.: 175004) ²⁴ and pAdDeltaF6 (gift from James M. Wilson; Addgene, Catalog No.: 112867), respectively. Medium was replaced 24 h after transfection with fresh transfection medium.

At 96 h after transfection, conditioned medium was collected, and culture chambers were rinsed with DPBS, incubated with trypsin (Gibco, Catalog No.: 12604013), and agitated to detach cells. Trypsin was neutralized with added medium used to recover remaining cells. The resulting cell/medium mixtures were centrifuged at 2000 × g for 15 min at room temperature (RT), and clarified supernatants were stored at 4°C for polyethylene glycol (PEG) virus precipitation. Cell and PEG pellets were digested with salt active nuclease (SAN) nuclease (ArcticZymes, Catalog No.: 70910‐202) in SAN buffer (500 mM NaCl, 40 mM Tris base, 10 mM MgCl2, pH 9.5) at 37°C for 1 h, viral particles were purified by iodixanol density‐gradient ultracentrifugation, washed, concentrated with DPBS using centrifugal filters, and sterile filtered. Purified virus was titrated using quantitative PCR targeting the WPRE region as described. ³⁵ Titered virus was aliquoted and stored at −80°C until use.

2.5. Retro‐orbital injections of AAV

AAVs were administered systemically via retro‐orbital sinus injections as described previously, with minor modifications. ³⁶ Mice were injected at 3 and 6 months of age in the pre‐ and post‐plaque cohorts, respectively. Concentrated AAV stock solutions were thawed and diluted in sterile DPBS to 3.13×10¹² viral genomes/mL (vg/mL). All mice received an 80‐µL injection to a total dose of 2.5×10¹¹ vg (5×10¹¹ for AAV‐asyn^E46K/tau^A152T combination treated). Diluted AAVs were prepared fresh the day of treatment and kept on ice until immediately prior to injections. Animals from one cage received different treatments, generally resulting in all five treatments being represented in each cage. All mice from a cage were anesthetized in an isoflurane vaporizer (3%) chamber, then removed one at a time and placed prone on a small stack of paper towels with the head oriented toward the operator's dominant hand. Using the index finger and thumb of the non‐dominant hand, the skin above and below the eye was gently retracted to slightly protrude the globe. With the dominant hand, a pre‐loaded insulin syringe (31‐gauge × 5/16‐inch), bevel down, was inserted at a 30° to 45° angle through the medial canthus and conjunctival membrane. The needle was advanced to position the tip in the retro‐orbital sinus behind the globe, the solution was slowly injected, and the needle was withdrawn gradually. One to two drops of 0.5% proparacaine ophthalmic solution were applied to the corneal surface for local analgesia, and mild pressure was applied to the eyelid. After recovery from anesthesia, mice were returned to their home cages.

2.6. Behavior

Open field assay (OFA) was used to assess the impact of amyloid, asyn, and tau pathologies on ambulatory and rearing behaviors. Metrics measured for pre‐ and post‐plaque induction cohorts at baseline (pre‐treatment), 3 mpi, and 6 mpi (endpoint). Mice were acclimated to the testing room for 1 h prior to testing, and all females were tested prior to males. Mice were placed in the center of an open‐field space (40 × 40 × 30 cm, W × L × H; Stoelting, Catalog No.: 62000) with opaque walls and allowed to freely explore for 15 min. To detect rearing behavior, infrared photobeam sensors external to the apparatus were elevated 3 cm (apparatus base to bottom of sensor). Total distance, the ratio of total distance to center area distance traveled, and time spent rearing were determined by ANY‐maze version 7.48 software (Stoelting). All testing apparatus were cleaned in between tests with 70% ethanol, dried thoroughly, and further aired out for 5 min. Experimenters were blinded to the treatment status of all animals during testing and analysis.

2.7. Tissue harvest and processing

Mice were deeply anesthetized via intraperitoneal injections of 150 mg/kg compounded pentobarbital solution (Euthasol) and then transcardially perfused with 20 mL ice‐cold PBS supplemented with 1 mM EDTA. Brains were surgically removed and bisected down the midline. Left hemibrains were drop‐fixed in 10% neutral‐buffered formalin for 48 h at 4C, then transferred to 30% sucrose in PBS for storage at 4°C. The cerebellum was removed from right hemibrains and the resulting hemi‐forebrains were snap‐frozen on dry ice. Following the final tissue harvest, preserved hemibrains were sent to NeuroScience Associates for MultiBrain embedding and coronal sectioning to 35 um. Forty hemibrains per embedding were arrayed, splitting males and females so that all treatments, mouselines, and time points could be represented on each MultiBrain array. Sections were stored in cryoprotectant, returned to our lab and kept at −20°C.

2.8. Immunofluorescent staining

We processed 11 free‐floating brain section arrays (n = 1 section per animal) per stain. Unless noted, all steps were performed at RT on an orbital shaker. First, sections were rinsed of cryoprotectant with PBS. For phospho‐protein and glial stains, native EGFP signal was quenched prior to permeabilization to allow use of the 488 channel. We used the “tanning bed” method to quench autofluorescence and EGFP signal by submerging sections in bleaching solution (PBS, 3% H2O2, 40 mM ammonium hydroxide) for 2×45 min while sandwiched between two 10,000‐lumen LED panels. Sections were rinsed thoroughly with PBS before proceeding with permeabilization. Sections were permeabilized in PBS‐X (PBS + 0.3% Triton‐X) for 30 min and blocked for 2 h in blocking buffer (PBS‐X + 5% normal serum from secondary host species). To suppress biotin‐related background, the panel including the total tau stain underwent additional avidin/biotin blocking: sections were incubated for 15 min in avidin solution (PBS‐X + 0.1 mg/mL Neutravidin; Thermo Fisher Scientific, Catalog No.: 31000), washed 3×10 min, incubated for 15 min in biotin solution (PBS‐X + 0.5 mg/mL biotin), and washed again 3×10 min in PBS‐X. Sections were incubated with primary antibodies in blocking buffer overnight at 4°C with gentle agitation. Sections were washed 3×10 min in PBS‐X and incubated with secondary antibodies in blocking buffer overnight at 4°C with gentle agitation, except for the transgene stain for which secondaries were incubated for 2 h at RT. Sections were washed 3×10 min in PBS‐X. The glia stain included a 10‐min 1 ug/uL DAPI (Thermo Fisher Scientific, Catalog No.: D1306) counterstain in PBS, followed by two more washes in PBS‐X. Finally, sections were washed in PBS, mounted with ProLong Glass Antifade Mountant (Thermo Fisher Scientific, Catalog No.:P36984), and cured at RT for 48 h before imaging. Slides were imaged at 10× on a Zeiss Axioscan 7 slide scanning microscope running Zen version 3.1. Whole section images were automatically stitched and shading corrections applied.

2.8.1. Antibodies

Total‐protein primaries: rabbit anti‐asyn (1:500; Thermo Fisher Scientific, Catalog No.:PA5‐85791), mouse anti‐human tau (HT7)‐biotin (1:250; Thermo Fisher Scientific, Catalog No.: MN1000B), chicken anti‐NeuN (1:500; Millipore, Catalog No.: ABN91). Secondaries: Alexa Fluor Plus 555 (AF+555) goat anti‐rabbit (1:750; Thermo Fisher Scientific, Catalog No.:A32732), streptavidin‐AF647 (1:750; Thermo Fisher Scientific, Catalog No.: S32357), DyLight 755 goat anti‐chicken (1:750; Thermo Fisher Scientific, Catalog No.: SA5‐10075). Phospho‐protein primaries: rabbit anti‐phospho S129 asyn (1:1000; Abcam, Catalog No.: EP1536Y), mouse anti‐phospho tau (AT8)‐biotin (1:200; Thermo Fisher Scientific, Catalog No.:MN1020B), chicken anti‐NeuN (detailed above). Secondaries: AF488 goat anti‐chicken (1:1000; Thermo Fisher Scientific, Catalog No.: A‐11039), AF+555 goat anti‐rabbit (1:1000, Thermo Fisher Scientific, Catalog No.: A32732), streptavidin‐AF647 (1:500; Thermo Fisher Scientific, Catalog No.: S32357). Glia and plaque stain: goat anti‐galectin‐3 (LGALS3; 1:300; R&D Systems, Catalog No.: AF1197), rabbit anti‐Iba1 (1:500; Wako, Catalog No.:019‐19741), chicken anti‐glial fibrillary acidic protein (GFAP) (1:2000; Thermo Fisher Scientific, Catalog No.: PA1‐10004), mouse anti‐beta amyloid (MOAB‐2)‐Alexa Fluor 750 (1:500; Novus, Catalog No.: NBP2‐13075AF750). Secondaries: AF488 donkey anti‐goat (1:750; Thermo Fisher Scientific, Catalog No.: A32814), Cy3 donkey anti‐rabbit (1:500; JIR, Catalog No.: 711‐165‐152), AF647 donkey anti‐chicken (1:750; Thermo Fisher Scientific, Catalog No.: A78952).

2.9. Staining analysis

Image quantifications were conducted in QuPath version 0.6.0. ³⁷ Region‐of‐interest (ROI) annotations for whole‐section (“global”) measurements were created automatically using a pixel thresholder for markers highly and ubiquitously expressed across groups (total asyn, NeuN, and DAPI for the transgene, phospho‐protein, and glia/plaque stain, respectively). Annotations for cortical phospho‐protein measurements were manually drawn to capture the isocortex using an atlas for reference (atlas.brain‐map.org). Care was taken to exclude overly damaged sections, areas with folds, and other causes of artifacts.

2.9.1. Mean intensity quantifications

The mean intensities of total tau and total synuclein stains were measured over each ROI using QuPath's Add Intensity Features function, with pixel size set to 0.69 and other parameters left at the default setting. Due to some minor section‐array batch effect on tau background signal, tau mean intensity measurements were normalized to EGFP‐treated animals on the same section array.

2.9.2. Percentage area quantifications

The percentage area of a ROI covered by total amyloid plaque (MOAB2), microglia (IBA1), astrocytes (GFAP), and activated glia (LGALS3) stains was determined using QuPath's pixel classifier thresholder by selecting a threshold that accurately represented positive pixels for a channel over all groups. The thresholder computes the percentage area of positive pixels over the ROI as follows: (Area of stain in µm / Area of ROI in µm) × 100.

2.9.3. Detection‐based quantifications

To quantify cell‐based metrics, InstanSeg version 0.1.5, a pre‐trained machine learning segmentation algorithm, ³⁸ was used within QuPath to segment neuronal soma using the “fluorescence_nuclei_and_cells‐0.1.0” model and NeuN, phospho‐tau, and phospho‐asyn as channel inputs. Object classifiers were trained on a subset of images to identify detections positive for NeuN, phospho‐tau, and phospho‐syn independently and then combined into a composite classifier that was applied to all images. The percentage of neuronal soma positive for phospho‐asyn, phospho tau, or both was calculated as follows: (Number of single‐ or double‐positive detections / Total number of NeuN detections) × 100.

2.10. Statistical analysis

All statistical analyses were performed in GraphPad Prism version 10. Metrics involving multiple data points per mouse (including neuronal positivities, plaque size) were averaged on a per‐mouse basis before plotting and statistics. Male and female mice were pooled for all analyses, with approximately equal sex ratios per group, as indicated in figure legends. All tests were two‐tailed, and a p ≤ 0.05 was considered statistically significant. Statistical tests and animal numbers are indicated in figure legends.

In summary, open‐field behaviors (total distance, center:total distance ratio, rearing) were analyzed at each time point (baseline, 3 mpi, 6 mpi) by two‐way ANOVA with mouseline and treatment as factors and Fisher's least significant difference for comparisons versus EGFP within each time point and mouseline. Data are displayed as line graphs with points indicating mean ± SEM. For histological and plaque measures including both mouseline APP and control animals, group effects were assessed using two‐way ANOVA with fixed‐factor mouseline and AAV plus interaction, followed by Tukey's multiple‐comparisons test. For analyses restricted to the controls or APP only (LGALS3 area subset and plaque metrics, respectively) one‐way ANOVA across treatments was followed by Dunnett's test versus EGFP. Data are displayed as violin plots (median and interquartile range) or bar graphs (mean ± SEM) with individual mice data points displayed.

3. RESULTS

3.1. AAV‐mediated asyn and tau accumulate more in mature amyloid environment

To determine the impact a mature amyloid milieu has on single asyn or tau co‐pathologies, as well as mixed co‐pathologies (asyn plus tau), we generated novel AAVs with hSyn promoter to induce the neuronal expression of asyn^WT, mutant asyn^E46K, mutant tau^A152T, or EGFP (Figure S1). Constructs were packaged into the AAV.CAP‐B10 viral serotype that exhibits high central nervous system neuronal tropism and minimal peripheral tissue transduction following systemic administration in adult mice. ²⁴ AAVs were administered via intravenous retro‐orbital injections to 6‐month‐old, “post‐plaque” deposition, APP^NL‐G‐F/MAPT double KI, and MAPT‐only mice (hereafter referred to as APP and control mice, respectively; Figure 1A). ²³ Here, “post‐plaque” refers to the high amyloid plaque load and significant plaque‐associated gliosis present at this induction age.

FIGURE 1.

AAV alpha‐synuclein and tau levels are increased in a mature amyloid milieu following post‐plaque induction. (A) Experimental schema for post‐plaque induction: 6‐month‐old APP and control mice received equal amounts of AAV encoding asyn^WT, asyn^E46K, tau^A152T, asyn^E46K/tau^A152T (co‐induction; equal viral genomes each), or AAV‐EGFP and brains were collected for histology at 6 mpi (12 months of age).  (B) Representative immunofluorescence images of coronal hemibrain sections stained for total asyn (endogenous mouse + AAV‐mediated human asyn) and total tau (endogenous human + AAV‐mediated human tau). (C and D) Quantification of whole‐section mean intensity and fold change relative to EGFP (indicated above violin plots) of (C) asyn and (D) tau signals across treatments in control and APP mice. Intensities of tau were normalized to EGFP‐treated mice. Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Tukey post hoc test for multiple comparisons (***p ≤ 0.001, ****p ≤ 0.0001). Violins show median (solid line) and quartiles (dotted) with individual animal data points shaped by sex. N = 6 to 16 mice/mouseline/treatment, roughly equal male:female. Scale bar = 100 µm.

At 6 mpi, when animals were 12 months of age, we immunostained for total asyn (endogenous mouse asyn plus AAV‐mediated human asyn) and total tau (endogenous human tau plus AAV‐mediated human tau). Brain‐wide transduction was evident, and the strongest signals were observed in the cortex (Figure 1B) with a preference for layers 2/3 and 5 (Figure S2A). Other key disease‐relevant areas that showed prominent asyn and tau signals include the CA2, CA3, and dentate gyrus subregions of the hippocampus (Figure S2B,C), as well as the lateral hypothalamus including connecting white‐matter fiber tracts (Figure S2D).

Quantification of global asyn signal in non‐amyloid controls that received AAV‐asyn^WT, asyn^E46K, and asyn^E46K/tau^A152T combination treatments produced ∼1.5fold increases in total asyn protein levels relative to endogenous (Figure 1C). However, APP mice of the same treatments exhibited total asyn levels substantially higher than controls, at approximately two‐fold over endogenous. Interestingly, AAVtau^A152T and combination treatments increased total tau by only up to ∼1.2fold in controls and ∼1.5fold in APP (∼30% higher), the former not reaching significance (Figure 1D). These genotype differences indicate that a mature amyloid environment promotes greater accumulation and/or reduced clearance of these pathologic proteins. Sex‐stratified analyses showed similar overall trends in males and females, although reduced statistical power rendered some effects non‐significant (Figure S3).

3.2. Phospho‐pathologies are exacerbated by amyloid

Aberrant phosphorylation of asyn and tau is a defining neuropathological hallmark of synucleinopathies and tauopathies, respectively. Prior studies indicated that amyloid could exacerbate concomitant asyn and tau pathologies induced by genetic crosses or pre‐formed fibril inoculation. ¹² , ¹³ , ¹⁴ , ¹⁵ We hypothesized that AAV‐mediated expression of asyn and tau would induce neuronal phospho‐pathologies that are further exacerbated in an amyloid‐laden environment. To test this, we immunostained for phospho‐asyn (Ser129) and phospho‐tau (Ser202, Thr205; AT8) at 6 mpi (12 months of age; Figure 2A). We focused our analysis on the cortex, which is susceptible to amyloid, tau, and synuclein multi‐proteinopathies in human disease. ² , ³⁹ , ⁴⁰ , ⁴¹

FIGURE 2.

Amyloid amplifies phospho‐asyn and phospho‐tau burden following post‐plaque induction. (A) Representative immunofluorescence images of the cortex of 6‐month‐old treated mice at 6 mpi stained for phospho‐asyn (Ser129) and phospho‐tau (Ser202, Thr205; AT8 antibody). (B and C) Quantification of cortical mean intensity of (B) phospho‐asyn and (C) phospho‐tau signals across treatments in control and APP. Intensities of p‐tau were normalized to EGFP‐treated mice. (D–F) Percentage of cortical neuronal soma positive for (D) only phospho‐asyn (p‐asyn⁺), (E) only phospho‐tau (p‐tau⁺), and (F) both (p‐asyn⁺p‐tau⁺). Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Tukey post hoc test for multiple comparisons (**p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001). Violins show median (solid line) and quartiles (dotted) with individual animal data points shaped by sex. N = 6 to 16 mice/mouseline/treatment, roughly equal male:female. Scale bar = 100 µm.

As anticipated, asyn^WT and asyn^E46K expression caused marked increases in phospho‐asyn levels at 6 mpi, and this effect was strongly exacerbated by amyloid (Figure 2B). In amyloid mice, mutant asyn^E46K produced higher phospho‐asyn levels than asyn^WT (Figure 2B), despite higher total asyn protein levels in asyn^WT mice (Figure 1C), suggesting that amyloid increased susceptibility to more pathogenic forms of asyn. Tau^A152T drove more phosphotau primarily in APP mice (Figure 2C), tracking with the larger total tau increase in amyloid compared to non‐amyloid mice (Figure 1D).

To assess neuronlevel susceptibility, we quantified the percentage of neurons positive for phospho‐asyn, phosphotau, or both. Overall, phospho‐asyn and phospho‐tau neuronal positivities were higher in amyloid mice than controls, consistent with the trends in overall phospho‐levels observed. On average, ∼10% of neurons were strongly phospho‐asyn+ in controls versus ∼20% in the presence of amyloid, across asynexpressing groups (Figure 2D). Compared to asyn pathology, tau^A152T yielded lower phospho‐tau+ prevalence overall (∼3% in controls, 12% in amyloid mice; Figure 2E). Doublepositive neurons from asyn^E46K/tau^A152T co‐transduction were infrequent, although still higher in the presence of amyloid (<1% in controls; <5% in amyloid mice; Figure 2F), reflective of human mixed‐pathology dementias reporting fewer double‐positive cortical neurons compared to single‐positive. ² These data indicate that a mature amyloid milieu increases susceptibility to both phosphopathologies and that neuronal phospho‐asyn is more prone to accumulation, especially in the presence of amyloid.

Sex‐stratified phospho‐pathology analyses showed similar overall trends in males and females, although some comparisons did not reach significance after stratification (Figure S4).

3.3. Alpha‐synuclein and tau co‐induction exacerbates hyperactivity and anxiety‐like behavior in aged APP mice

Hyperactivity and altered anxiety‐like behaviors are common consequences of amyloidosis, tauopathy, and synucleinopathy in rodent models. ³¹ , ⁴² , ⁴³ , ⁴⁴ We hypothesized that combined pathologies would produce enhanced or emergent phenotypes relative to single proteinopathies and that these effects would worsen over time.

To assess behavioral changes longitudinally, we assayed open‐field activity at pre‐treatment baseline (6 months of age), at 3 mpi (9 months of age), and at 6 mpi (12 months of age; Figure 3A–C). Amyloid mice were consistently hyperactive at baseline and at subsequent time points relative to controls, which exhibited decreased activity upon reassessment in the open‐field assay, likely due to familiarity/habituation (Figure 3A). These trends were largely unaffected by single pathologies; however, the asyn^E46K/tau^A152T combination in amyloid mice produced dramatic increases in locomotion at 3 and 6 mpi and reduced center:total distance ratios relative to EGFP, an indication of anxiety‐like behavior (Figure 3A,B). AAVtau^A152T alone and asyn^E46K/tau^A152T led to similar reductions in rearing in APP at 6 mpi, which could be interpreted as motor incoordination or anxiety‐related reduced exploratory behavior (Figure 3C). Together, increased locomotion with center avoidance in the triple‐pathology mice suggests synergistic interactions exacerbate the hyperactive, anxiety‐like phenotype in aged amyloid mice, while the presence of tau^A152T alone is sufficient to suppress rearing behavior.

FIGURE 3.

Co‐induced asyn and tau pathologies exacerbate amyloid‐associated hyperactivity and anxiety‐like behaviors. (A–C) Open‐field assay behaviors were measured in post‐plaque treated mice at baseline (6 months of age), 3 mpi (9 months of age), and 6 mpi (12 months of age). Metrics for (A) total distance traveled, (B) center:total distance traveled ratio, and (C) time spent rearing were automatically recorded during each test. Points show group means with error bars (± SEM). Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Fisher's least significant difference post hoc test for multiple comparisons (*p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001). Asterisks denote significant relative EGFP at a given time point and are colored by group for overlapping data points. N = 8 to 17 mice/mouseline/treatment/time point, roughly equal male:female.

Sex‐stratified open‐field analyses showed that total distance and rearing trends were generally similar between males and females, although some comparisons did not reach significance after stratification. The reduced center:total distance phenotype in APP mice appeared most evident in females, while interpretation in males was limited by smaller subgroup size (Figure S5).

3.4. Tau^A152T drives glial activation, enhanced by a mature amyloid environment

Neuroinflammation is both a mediator and a consequence of neurodegeneration. ⁴⁵ , ⁴⁶ We hypothesized that additional pathologies would enhance inflammatory responses when overlaid onto a pre‐existing inflamed amyloid environment. To characterize glial responses, we co‐stained for activated astrocytes (GFAP), microglia (IBA1), and activated glia as identified by galectin‐3 (LGALS3) immunoreactivity (Figure S6A).

Global astrocyte and microglia coverage were markedly elevated in amyloid mice versus controls at 12 months of age (25% vs 6% area GFAP; 11% vs 4% area IBA1; p < 0.0001) but were not significantly altered by additional asyn or tau pathologies (Figure 4A,B). This suggests that high amyloid plaque load at 12 months of age dominates overall astrocyte and microglia reactivity, potentially masking interactions with asyn and tau pathologies. In contrast, LGALS3‐related glial activation increased robustly with tau^A152T expression – with or without concomitant asyn^E46K expression – in amyloid mice and, more modestly, in non‐amyloid control mice (Figure 4C). These trends in overall glial responses and LGALS3 activation were similar in males and females, although LGALS3 increases did not reach significance in female controls after stratification (Figure S6B–G).

FIGURE 4.

A mature amyloid environment enhances tau^A152T‐driven LGALS3 inflammatory responses in white‐matter fibers. (A–C) Quantification of the percentage of whole‐section area covered by thresholded positive pixels in immunofluorescent co‐stain (overview images in Figure S6) for (A) GFAP (reactive astrocytes), (B) IBA1 (microglia), and (C) LGALS3 (reactive glia) markers in mice induced at 6 months of age and analyzed 6 mpi (12 months of age). (D) Insets from zoomed 10× overviews depict spatial overlap of LGALS3 with GFAP and IBA1 in white‐matter tracts of the inferior lateral forebrain bundle system. (E) Higher‐magnification 40× images from the inset region indicated in panel (D) demonstrate clear double labeling of LGALS3 and GFAP (arrows) and LGALS3 and IBA1 (red arrowheads), as well as LGALS3 without clear co‐localization with either marker (open arrowheads). Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Tukey's post hoc test for multiple comparisons was used for combined analyses; while one‐way ANOVA and Dunnett's post hoc test for multiple comparisons were used for analyses of non‐amyloid control cohort alone (**p ≤ 0.01, ****p ≤ 0.0001). Bar graphs indicate group averages ± SEM, with individual animal data points shaped by sex. N = 6 to 17 mice/mouseline/treatment, roughly equal male:female. Scale bar = 1 mm (D); 30 µm (E).

Surprisingly, the most robust taulinked glial activation localized to whitematter fiber tracts of the lateral forebrain bundle system (Figure S6A – arrows), where increased GFAP, IBA1, and LGALS3 signals could be observed (Figure 4D). Higher‐magnification imaging demonstrated LGALS3 signal in direct overlap with both GFAP‐positive astrocytic and IBA1‐positive microglial cells and processes, while additional LGALS3‐positive staining without clear co‐localization with either marker was also observed (Figure 4E). These findings suggest that astrocytes, microglia, and other glial cells may contribute to this tau‐mediated LGALS3‐related white‐matter inflammatory response. Notably, amyloid plaque load was unchanged in either sex by AAV treatments (Figure S7), indicating that the heightened LGALS3 activation in APP mice was not a byproduct of altered amyloid burden.

3.5. Induction of asyn or tau pathologies before substantial amyloidosis results in strong total and phospho‐protein levels without amyloid dependence

Extensive evidence supports the idea that amyloid precipitates taurelated pathogenesis in AD. ⁴⁷ However, the temporality of amyloid, tau, and synuclein pathogenesis in LBD is more ambiguous, attributable to the heterogeneity of LBD‐related neuropathologies and clinical trajectories. ⁴⁰ , ⁴⁸ , ⁴⁹ Hence, we asked how early, “pre‐plaque” induction of asyn and tau pathologies might shape susceptibilities. To answer this, we injected a separate cohort of mice at 3 months of age – when only sparse plaques are present and plaque‐associated gliosis is scarce – and subjected them to the same behavioral and histological assessments (Figure 5A).

FIGURE 5.

Early pre‐plaque induction drives robust increases in asyn and tau levels, independent of APP status. (A) Experimental schema for pre‐plaque induction: 3‐month‐old APP and control mice received equal amounts of AAV encoding asyn^WT, asyn^E46K, tau^A152T, asyn^E46K/tau^A152T (co‐induction; equal viral genomes each), or AAV‐EGFP and brains were collected for histology at 6 mpi (9 months of age). (B) Representative immunofluorescence images of coronal hemibrain sections stained for total asyn (endogenous mouse + AAV‐mediated human asyn) and total tau (endogenous human + AAV‐mediated human tau). (C and D) Quantification of whole‐section mean intensity and fold change relative to EGFP (indicated above violin plots) of (C) asyn and (D) tau signals across treatments in control and APP. Intensities of tau were normalized to EGFP‐treated mice. Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Tukey post‐hoc test for multiple comparisons (*p ≤ 0.05, ***p ≤ 0.001, ****p ≤ 0.0001). Violins show median (solid line) and quartiles (dotted) with individual animal data points shaped by sex. N = 12 to 19 mice/mouseline/treatment, roughly equal male:female. Scale bar = 100 µm.

At 6 mpi (9 months of age), asyn and tau levels were strongly elevated to similar degrees in amyloid and control mice: asyn treatments averaged ∼1.8fold over endogenous (Figure 5B,C) and tau ∼1.5fold (Figure 5B,D). Cortical phospho‐asyn signal intensities and percent neuronal positivities were marginally elevated in amyloid mice compared to controls but only reached significance in asyn^WT mice (Figure 6A,B,D). As in the postplaque induction cohort, mutant asyn^E46K induced greater phospho‐asyn pathologies than asyn^WT (Figure 6B,D), demonstrating that the higher pathogenicity of mutant asyn^E46K was independent of induction age. Consistent with total protein levels, phospho‐tau signal and neuronal positivity were similarly increased in both amyloid and control mice (Figure 6A,C,E). Sex‐stratified pre‐plaque cohort total protein (Figure S8) and phospho‐pathology (Figure S9) analyses demonstrated similar trends for males and female.

FIGURE 6.

Pre‐plaque induction yields phospho‐asyn and phospho‐tau increases with minimal amyloid impact. (A) Representative immunofluorescence images of cortex of 3‐month‐old treated mice at 6 mpi stained for phospho‐asyn (Ser129) and phospho‐tau (Ser202, Thr205; AT8 antibody). (B and C) Quantification of cortical mean intensity of (B) phospho‐asyn and (C) phospho‐tau signals across treatments in control and APP. Intensities of p‐tau were normalized to EGFP‐treated mice. (D–F) Percentage of cortical neuronal soma positive for (D) only phospho‐asyn (p‐asyn⁺), (E) only phospho‐tau (p‐tau⁺), and (F) both (p‐asyn⁺p‐tau⁺). Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Tukey post hoc test for multiple comparisons (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001). Violins show median (solid line) and quartiles (dotted) with individual animal data points shaped by sex. N = 12 to 19 mice/mouseline/treatment, roughly equal male:female. Scale bar = 100 µm.

Notably, the fraction of phospho‐asyn+ neuronal soma (∼15%; Figure 6D) remained higher than phosphotau+ soma (∼8%; Figure 6E) across treatments. Doublepositive neurons after asyn^E46K/tau^A152T were again the rarest population, averaging ∼2.5% in controls and ∼3.5% in amyloid mice (Figure 6F). These findings indicate that amyloid does not confer substantially increased susceptibility to proteinopathies that are induced before plaque deposition, underscoring induction age as a key determinant of interactions with amyloid.

3.6. Pre‐plaque alpha‐synuclein and tau pathologies potentiate emerging amyloid phenotypes

Open‐field behavior in the pre‐plaque treated cohort was measured at baseline (3 months of age), 3 mpi (6 months of age), and 6 mpi (9 months of age, endpoint) to determine the impact of early asyn and tau pathologies in the context of an emerging amyloid environment. While APP mice were hyperactive relative to controls at baseline, they showed decreased locomotion at 3 and 6 mpi, except when induced with tau^A152T or asyn^E46K/tau^A152T pathologies, which potentiated or exacerbated hyperactivity, respectively (Figure 7A). These hyperactive APP groups corresponded to reduced anxiety‐like behavior (Figure 7B). Interestingly, in non‐amyloid controls, asyn and tau co‐induction was sufficient to induce a mild anxiety‐like phenotype at 6 mpi (Figure 7B), albeit more slowly and to a lesser magnitude than amyloid mice. Rearing behavior, in contrast, was strongly reduced at 3 and 6 mpi in controls and amyloid mice with tau or both asyn/tau pathologies induced (Figure 7C). This heightened tau^A152T‐driven rearing phenotype in 3‐ compared to 6‐month‐old treated mice indicates early susceptibility to tau‐mediated alterations in vertical exploration related to anxiety‐like behavior and/or motor stability. These trends hold across open‐field metrics when analysis is stratified by sex, though statistical power renders some subgroup effects non‐significant (Figure S10).

FIGURE 7.

Pre‐plaque alpha‐synuclein and tau pathologies potentiate emerging APP behavioral phenotypes. (A–C) Open‐field assay behaviors were measured in pre‐plaque treated mice at baseline (3 months of age), 3 mpi (6 months of age), and 6 mpi (9 months of age). Metrics for (A) total distance traveled, (B) center:total distance traveled ratio, and (C) time spent rearing were automatically recorded during each test. Points show group means with error bars (± SEM). Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Fisher's least significant difference post hoc test for multiple comparisons (*p ≤  0.05, **p ≤  0.01, ***p ≤  0.001, ****p ≤  0.0001). Asterisks denote significant relative EGFP at a given time point and are colored by group for overlapping data points. N = 22 to 27 mice/mouseline/treatment/time point, roughly equal male:female.

3.7. Pre‐plaque alpha‐synuclein and tau pathologies accelerate astrogliosis while tau independently drives glial activation

We next sought to determine whether introducing early asyn and tau pathologies would be sufficient to induce inflammation or whether they would alter the amyloid‐driven inflammation expected to emerge between baseline and endpoint (6 mpi, 9 months of age).

Global astrocyte and microglia coverage were markedly elevated in amyloid versus controls at 9 months of age (23% vs 7% area GFAP; 10% vs 4% area IBA1; p < 0.0001; Figures 8A,B and S11A), consistent with the mouseline‐driven differences observed in the post‐plaque cohort at 12 months of age (Figure 4A). In non‐amyloid controls, these levels were unchanged by synuclein or tau pathologies. However, in contrast to the post‐plaque cohort, astrocyte coverage in the pre‐plaque cohort increased in amyloid mice with asyn^E46K and tau^A152T induction (Figure 8A), despite the unchanged amyloid plaque burden (Figure S12), suggesting early neuronal pathologies can alter astrocyte responses to subsequent amyloid deposition.

FIGURE 8.

Pre‐plaque alpha‐synuclein and tau pathologies accelerate astrogliosis, while tau independently drives glial activation. (A–C) Quantification of the percentage of whole‐section area covered by thresholded positive pixels in immunofluorescent co‐stain (overview images in Figure S11) for (A) GFAP (reactive astrocytes), (B) IBA1 (microglia), and (C) LGALS3 (reactive glia) markers. (D) Insets from zoomed 10× overviews depict spatial overlap of LGALS3 with GFAP and IBA1 in white‐matter fiber tracts of the inferior lateral forebrain bundle system. Statistics: two‐way ANOVA (factors: mouseline, treatment) with interaction and Tukey's post hoc test for multiple comparisons was used for combined analyses, while one‐way ANOVA and Dunnett's post hoc test for multiple comparisons was used for analysis of non‐amyloid control cohort alone (**p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001). Bar graphs indicate group means ± SEM, with individual animal data points shaped by sex. N = 12 to 18 mice/mouseline/treatment, roughly equal male:female. Scale bar = 1 mm.

Tau^A152T with or without concomitant asyn^E46K pathology again elicited pronounced LGALS3‐related glial activation in the white matter of controls (Figure 8C,D), albeit stronger than in the 6‐month‐old treated cohort (Figure 4C,D), which is consistent with the higher tau burden that occurred from earlier induction. In amyloid mice, tau^A152T likewise induced a whitematter glial activation response that appeared largely additive to the global effect of amyloidosis (Figure 8C,D).

Sex‐stratified glial and amyloid plaque analyses showed highly similar trends in males and females, although some comparisons did not reach significance after stratification (Figures S11B–G and S12).

4. DISCUSSION

Our study demonstrates that the stage of amyloid pathology at the time of induction is a key determinant of how asyn and tau accumulate, phosphorylate, activate glia, and alter behavior. We leveraged state‐of‐the‐art AAV capsids (AAV.CAP‐B10) and carefully designed AAV transgenes to deliver robust and persistent asyn^WT, mutant asyn^E46K, and mutant tau^A152T payloads to the neurons of APP^NL‐G‐F/MAPT double KI mice that had the significant advantage of biologically appropriate expression of APP, triple‐mutant human Aβ, and multi‐isoform human tau.

Traditional transgenic crosses can be valuable for generating robust multi‐proteinopathies, but developmental overexpression and supraphysiologic protein levels can complicate the interpretation of when and how pathologies interact. Alternatively, asyn and tau fibril seeding models overlaid on genetic amyloid models are useful for interrogating templated aggregation and anatomical propagation following temporal and regional control of initial lesions. The present AAV and KI combination approach offers a unique experimental approach relative to previous mixed‐pathology models, in that it enables brain‐wide adult induction of proteinopathies in a defined amyloid environment without APP overexpression artifacts, making it particularly well suited to test how amyloid stage alters susceptibility to additional pathologies. Furthermore, the humanized MAPT background adds another advantage for modeling tau interactions, because exogenously introduced human tau is expressed in a species‐matched tau context, building on prior work showing that the humanization of murine Mapt enables human tau fibril propagation. ²³ Although our study did not directly test tau propagation, this design may better approximate the environment in which human tau pathology develops compared to systems relying on cross‐species tau interactions.

Using this platform, we found that a mature amyloid milieu at induction (transduction of APP mice at 6 months of age) enhanced accumulation of protein and downstream phospho‐protein pathologies by 6 mpi, whereas pre‐plaque induction (transduction of APP mice at 3 months of age) yielded strong asyn/tau burdens and phospho‐signals by 6 mpi, with little added susceptibility conferred by amyloid. Across cohorts, neuronal phospho‐asyn+ prevalence exceeded phospho‐tau+, double‐positive neurons remained rare, and triple pathology produced the most robust behavioral disruptions. Remarkably, tau^A152T elicited a distinctive dose‐dependent LGALS3‐linked white‐matter inflammatory response that occurred with and without amyloid, pointing to a potential mechanism for this variant's effect on disease risk. Together, these findings argue that (i) amyloid potentiates an environment more permissive to protein accumulation and phosphorylation, (ii) asyn, tau, and amyloid pathologies interact to drive emergent phenotypes influencing behavior, and (iii) tau^A152T‐mediated white‐matter inflammation represents a separate axis of vulnerability.

4.1. Amyloid alters age‐related shifts in susceptibility to AAV‐driven pathologies

Induction of asyn and tau expression in aged mice, after plaque deposition, revealed an amyloid related increase in protein levels and subsequent phosphorylation, with a substantially stronger impact on tau. In contrast, induction in young mice, before plaque deposition, led to comparable total and phospho‐protein levels in APP and control mice. This suggests that a mature amyloid environment is more permissive to AAV‐mediated pathological protein increases. Amyloid is known to disrupt neuronal proteostasis through proteasome impairment ⁵⁰ , ⁵¹ and autophagy/lysosomal dysfunction, ⁵² , ⁵³ which could raise the steady‐state burden of aggregation‐prone proteins, favoring their phosphorylation. It will be informative to test whether lowering neuronal proteostasis stress (e.g., enhancing proteostasis by improving autophagic flux or lowering plaque‐load with anti‐amyloid immunotherapy) rescues the post‐plaque amplification of pathology sensitivity.

4.2. E46K mutation accentuates phospho‐synucleinopathy

Mutant asyn^E46K produced higher phospho‐asyn levels and neuronal positivity than asyn^WT in both cohorts, even when total WT levels exceeded E46K. Established amyloid amplifies the effect but is not necessary for it, indicating mutant‐intrinsic pathogenicity. This contributes to the body of knowledge on the propensity of the SNCA E46K mutation to alter asyn fibril structure in ways that promote phosphorylation/aggregation and thereby increase pathogenicity. ²⁸ , ⁵⁴ , ⁵⁵

4.3. Neurons more readily accumulate phospho‐asyn than phospho‐tau, and co‐positivity is rare

By quantifying neuronal positivity for phospho‐asyn and/or phospho‐tau markers, we found that across conditions, phospho‐asyn+ neurons outnumbered phospho‐tau+, while double‐positive neurons remained the rarest class in co‐transduced mice. This pattern mirrors human mixed‐pathology LBD, where neuronal asyn inclusions far outweigh tau, and true co‐localization is limited. ² , ⁴⁸ , ⁵⁶ Mechanistically, such patterns may reflect neuronal subtype/compartment specificity or temporal differences in pathology maturation. ⁵⁷ , ⁵⁸ , ⁵⁹ Future work elucidating the underlying causes and impact of differential neuronal susceptibilities to asyn and tau pathology, as well as the influence of amyloid, will be important in treating mixed dementias.

4.4. Behavioral phenotypes point to synergistic effects

Open‐field increases in locomotion with a preference for the perimeter of the field are readouts of elevated arousal/anxiety‐like behavior in mice. ⁶⁰ The APP^NL‐G‐F single KI line shows non‐cognitive behavioral alterations that vary with age and test conditions. ⁶¹ , ⁶² , ⁶³ The MAPT‐KI‐only line (used as our controls in this study) shows normal cognition and locomotion throughout aging. ⁶⁴ Here, we show previously unreported increased ambulation and rearing in APP^NL‐G‐F/MAPT double KIs (APP) relative to MAPT KIs (control). In the post‐plaque treated cohort, single pathologies had limited impact on the 6‐month‐old APP baseline locomotion/anxiety behaviors, whereas combined asyn^E46K/tau^A152T induction markedly exacerbated these features. In contrast, in the pre‐plaque induced cohort, APP animals typically downshifted activity with repeated behavioral assessment, mirroring control behavior. Early induction of tau^A152T, however, potentiated the baseline hyperactivity observed in 3‐month‐old APP mice, and the asyn^E46K/tau^A152T combination again exacerbated hyperactivity/anxiety. That tau^A152T alone suppressed rearing is consistent with prior links between tauopathy and motor and anxiety‐like exploratory impairments, ⁶⁵ , ⁶⁶ including in other A152T models. ³¹

The concordant behavioral trends between pre‐ and post‐plaque cohorts underscore that co‐asyn/tau pathologies may synergize with mature, as well as emerging, Aβ pathologies to influence network‐level disruptions to behavior. Furthermore, the observation that the triple‐pathology behavioral exacerbations occurred despite the low frequency of phospho‐asyn/phospho‐tau double‐positive neurons argues against within‐neuron co‐aggregation as the sole driver of synergy in this model. Instead, the data are more consistent with non‐cell‐autonomous interactions, in which asyn and tau pathologies disrupt shared circuits and/or engage glial responses already altered by amyloid, ultimately amplifying functional impairment. Such a framework is also consistent with human mixed‐pathology dementia, where exacerbations of clinical features can occur despite limited neuronal co‐localization of these lesions.

4.5. Tau^A152T provokes LGALS3‐linked white‐matter inflammation

A striking and consistent finding was that tau^A152T drove an LGALS3 response in white‐matter fiber tracts, prominently of the lateral forebrain bundle system, accompanied by increases in astrocyte and microglial reactivity. Galectin‐3 (LGALS3) has many cellular roles, heavily implicated in disease‐associated microglia (DAM) states (e.g., triggering receptor expressed on myeloid cells 2 [TREM2] signaling) and other pro‐inflammatory processes across neurodegenerative injury settings, including white‐matter‐predominant contexts. ⁶⁷ , ⁶⁸ , ⁶⁹ While LGALS3 has been predominantly characterized as a marker of DAM, its roles in astrocytic proliferation, phagocytosis, and inflammation are increasingly studied. ⁷⁰ , ⁷¹ , ⁷² In support of previous work, we observed that amyloid exerted a localized impact on microglial LGALS3 upregulation. However, we showed that the dramatic tau‐mediated LGALS3 increases in the white matter, while heavily associated with microglia, also stemmed from astrocytes and, perhaps, other glial cells. Activation at white‐matter fiber tracts points to an axon‐centric tau perturbation that may be indirectly related to high distal cortical pathologies. In this context, microglial LGALS3 may reflect inflammatory/phagocytic glial activation, whereas astrocyte‐associated LGALS3 may indicate reactive remodeling of the local white‐matter environment that could influence axonal support or myelin homeostasis. Whether LGALS3 here reflects detrimental activation (e.g., TREM2‐modulating, pro‐inflammatory) and results in axon/myelin damage or impaired conduction, as opposed to reparative phagocytic programs should be a focus of future studies.

4.6. Induction timing shapes astrocyte responses to amyloid

Lastly, global microgliosis and astrogliosis were dominated by amyloid plaque load in the post‐plaque cohort and insensitive to added asyn and/or tau burden; but in the pre‐plaque cohort, early tau^A152T and asyn^E46K measurably increased astrocyte coverage in APP mice without altering subsequent amyloid plaque burden. This suggests that amyloid in this system has the strongest neuroinflammatory consequences, but, depending on the timing of their occurrence, other lesions can exacerbate or accelerate the impact of amyloid on gliosis. One interpretation is that early neuronal stressors can accelerate the astrocyte program ⁷³ that would otherwise emerge more gradually with amyloid maturation.

5. LIMITATIONS AND CONCLUSION

Several limitations should be considered when interpreting these findings. First, we did not include cognitive assays or motor‐specific behavioral tests to assess pathology interactions. Because mixed‐pathology dementias affect cognitive, motor, and behavioral systems, additional assays could have provided a more holistic picture of the functional consequences of mixed pathologies in our model. Second, although longitudinal open‐field testing was necessary to track behavioral change over time within the same cohorts, repeated exposure to the same assay introduces familiarity and habituation effects that may influence measured outcomes, particularly if memory is altered by disease or treatment. Lastly, the consequences of tau^A152T expression in our system may not be restricted to this variant of tau, as WT tau and other pathogenic tau variants were not tested in parallel.

In conclusion, our study highlights that a mature amyloid environment has the capacity to amplify the accumulation and phosphorylation of asyn and tau neuropathologies, which can interact synergistically to drive robust behavioral alterations, while tau is distinctly capable of provoking an LGALS3‐linked white‐matter inflammatory response. The timing of induction is critical to these pathologies: Young brains accumulate lesions independently of strong amyloid interactions, whereas mature amyloid brains convert similar inputs into disproportionately pathogenic outputs. These insights refine when and how amyloid intersects with asyn and tau biology, reflect the need for multi‐targeting approaches, and nominate tract‐focused LGALS3+ glial states as potential diagnostic and therapeutic targets in tau risk variant carriers.

AUTHOR CONTRIBUTIONS

Conceptualization: BER, JDF. Data curation: BER. Formal analysis: BER. Funding acquisition: JDF, RC, OAR, PJM, DWD. Investigation: BER, LN, KAH, DGN, JHB. Methodology: BER, LN, KAH, LET, DGN, NM. Project administration: BER, KAH, JDF. Resources: BER, DGN, JDF. Software: BER, DGN. Supervision: BER, KAH, JDF. Visualization: BER. Validation: BER, JDF. Writing – original draft: BER. Writing – review and editing: BER, JDF.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the Supporting Information.

CONSENT STATEMENT

No human subjects were used in this study.

Supporting information

Supporting Figures

ICMJE Disclosures Form

Supporting Information