Structural and functional damage to neuronal nuclei caused by extracellular tau oligomers

Abstract

INTRODUCTION

Neuronal nuclei are normally smoothly surfaced. In Alzheimer's disease (AD) and other tauopathies, though, they often develop invaginations. We investigated mechanisms and functional consequences of neuronal nuclear invagination in tauopathies.

METHODS

Nuclear invagination was assayed by immunofluorescence in the brain, and in cultured neurons before and after extracellular tau oligomer (xcTauO) exposure. Nucleocytoplasmic transport was assayed in cultured neurons. Gene expression was investigated using nanoString nCounter technology and quantitative reverse transcription polymerase chain reaction.

RESULTS

Invaginated nuclei were twice as abundant in human AD as in cognitively normal adults, and were increased in mouse neurodegeneration models. In cultured neurons, nuclear invagination was induced by xcTauOs by an intracellular tau‐dependent mechanism. xcTauOs impaired nucleocytoplasmic transport, increased histone H3 trimethylation at lysine 9, and altered gene expression, especially by increasing tau mRNA.

DISCUSSION

xcTauOs may be a primary cause of nuclear invagination in vivo, and by extension, impair nucleocytoplasmic transport and induce pathogenic gene expression changes.

Highlights

• Extracellular tau oligomers (xcTauOs) cause neuronal nuclei to invaginate.

• xcTauOs alter nucleocytoplasmic transport, chromatin structure, and gene expression.

• The most upregulated gene is MAPT, which encodes tau.

• xcTauOs may thus drive a positive feedback loop for production of toxic tau.

1. BACKGROUND

Oligomeric and filamentous tau are key pathogenic factors in tauopathies, such as Alzheimer's disease (AD), progressive supranuclear palsy (PSP), and some forms of frontotemporal dementia due to tau mutations (FTD‐tau). ¹ , ² Tauopathy spreads from neuron to neuron by cycles of tau aggregate release into the extracellular space and subsequent neuronal uptake. ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ While the mechanisms of this prion‐like spread of pathogenic tau have been extensively studied, the cell biological responses of neurons to aggregated tau uptake has attracted much less attention.

The nuclear lamina, a meshwork of intermediate filaments formed by lamins A, B1, B2, and C, that underlies the inner nuclear membrane, plays critical roles in maintenance of normal nuclear structure and function. ¹¹ , ¹² Deformation of the nuclear membrane, including nuclear lamina invagination, has been implicated in multiple neurodegenerative diseases, such as AD, FTD, and Parkinson's disease, and can be provoked by dysfunctional tau. ¹³ , ¹⁴ , ¹⁵ For example, expression of pathogenically mutated human tau in Drosophila neurons caused nuclear lamina invagination, ¹⁵ and intracellular tau oligomerization led to disassembly of lamin B2, a major component of the nuclear lamina. ¹⁶

With this background in mind, we investigated whether extracellular tau oligomers (xcTauOs) affect neuronal nuclei. We present the novel findings that xcTauOs rapidly cause striking nuclear invagination, and that it occurs by a mechanism that depends on intracellular tau, and is associated with altered nucleocytoplasmic transport, chromatin structure, and gene transcription. Altogether, our results implicate xcTauOs as seminal factors in the conversion of healthy neurons into diseased neurons in AD and other tauopathies.

2. METHODS

2.1. Human brain tissue

Paraffin‐embedded, 5 to 6 μm thick human brain autopsy sections (temporal lobe with hippocampus) were obtained from the archives of the University of Virginia Department of Pathology. Frozen human cortical brain autopsy samples were provided by Dr. Heather A. Ferris of the University of Virginia School of Medicine. Pertinent information about each donor is shown in Table S1 in supporting information. Institutional approval for use of archival autopsy tissue was obtained from the University of Virginia Biorepository and Tissue Research Facility.

RESEARCH IN CONTEXT

1. Systematic review: Tauopathies, including Alzheimer's disease (AD), are linked to accumulation of oligomeric and filamentous tau. Despite extensive research on transcellular propagation of tau pathology, there are few reports of functional effects of tau oligomers on neurons. We investigated the impact of extracellular tau oligomers (xcTauOs) on the structure and functions of neuronal nuclei.

2. Interpretation: Invaginated nuclei are common in human AD neurons and mouse AD model neurons in vivo, and are induced in cultured mouse neurons by xcTauOs. This effect requires intracellular tau, and is accompanied by disrupted nucleocytoplasmic transport, altered chromatin organization, and differential gene expression patterns, particularly upregulation of tau mRNA.

3. Future directions: Mechanistic studies will determine how xcTauOs alter the structure and function of neuronal nuclei. Such studies can reveal new targets for the prevention of AD and other tauopathies, and for developing drugs directed at those targets.

2.2. Mouse brain tissue

CVN mice were originally obtained from Drs Michael Vitek and Carol Colton of Duke University, and were maintained as a breeding colony. PS19 mouse brain sections were provided by Dr. John Lukens of the University of Virginia Department of Neuroscience. Animals were maintained, bred, and euthanized in compliance with all policies of the Animal Care and Use Committee of the University of Virginia. Fifty micrometer floating sections of brain tissue were cut after trans‐cardiac perfusion of 4% paraformaldehyde in phosphate buffered saline (PBS) of mice that had been deeply anesthetized intraperitoneally with ketamine/xylazine (280/80 mg/kg).

hTau mice were originally obtained from the Jackson Laboratory and maintained as a breeding colony. Mice were euthanized by carbon dioxide inhalation. The brains were then harvested and saved at −80°C for future use. For all experiments, only male mice were used.

2.3. Cultured mouse neurons

Brain cortices were collected from E17/18 wild type (WT) C57/Bl6 or tau knockout mice ¹⁷ and processed as described earlier. ¹⁸ Briefly, brain tissue was cut into small pieces in ice‐cold Hank's balanced salt solution, and then was digested at 37°C with 0.25% trypsin (Gibco, 15090‐046) and 500 units DNase (Worthington, LK003170) for 20 minutes. Digestion was stopped by adding an equal volume of fetal bovine serum. The resulting single cell suspension was washed three times with 5 mL of Neurobasal Plus medium (Gibco, A3582901) containing B27 Plus supplement (Gibco, A3582801), 2 mM GlutaMAX (Gibco, 35050061), 6 mg/mL glucose (Sigma‐Aldrich, 16301), and 10 μg/mL gentamicin (Gibco, 15750078), mechanically dissociated with a fire polished Pasteur pipet and diluted into supplemented Neurobasal Plus medium. Sixty‐five thousand cells/cm² were plated on 50 μg/mL poly‐D‐Lysine (Sigma‐Aldrich, P0899) coated plates. Neurons were maintained in a tri‐gas incubator in an atmosphere of 5% each of O₂ and CO₂. Fifty percent media changes were done every 3 to 4 days until day 12 or 13 when experimental treatments began.

2.4. Recombinant tau

All six human tau isoforms with a C‐terminal his‐tag were produced by expression in BL21(DE3) Escherichia coli cells. Protein expression was induced with 0.5 mM Isopropyl β‐D‐1‐thiogalactopyranoside (Sigma‐Aldrich, I6758) for 2 hours at 37°C. Cells were pelleted by centrifugation, resuspended, and sonicated with Misonix Sonicator 3000 20 times for 30 seconds each at 60% power. Next, the E. coli lysate was centrifuged in Sorvall RC6 Plus 6 centrifuge at 18,083 × g_max for 25 minutes with an SLA 1500 rotor. Tau was then purified batchwise from the supernatant using TALON Metal Affinity Resin (TaKaRa, 635502) according to the vendor's instructions. Finally, each purified tau isoform was concentrated, and buffer exchanged into 10 mM HEPES, pH 7.6, using Amicon Ultra‐4 Centrifugal Filters (Millipore, UFC801024).

2.5. Tau oligomers

Oligomeric tau was prepared as described earlier. ¹⁹ Each purified tau isoform was adjusted to 8 μM in the oligomerization buffer: 10 mM HEPES (pH 7.6), 100 mM NaCl, 0.1 mM ethylenediaminetetraacetic acid, and 5 mM dithiothreitol (DTT). The protein was then allowed to oligomerize in the presence of 300 μM arachidonic acid (Cayman Chemicals, 90010) for 18 hours at room temperature in the dark. Oligomerization was verified on western blots. Primary mouse neurons at 12 to 13 days in vitro were exposed to oligomeric or monomeric tau at 30 to 500 nM total tau for the indicated time periods. Oligomerization buffer containing 300 μM arachadonic acid was used as vehicle control for cell treatments.

2.6. Atomic force microscopy

Five microliters of monomers and oligomers of 1 μM total tau was deposited onto freshly cleaved mica (Sigma‐Aldrich, AFM‐71855‐15‐10) and allowed to absorb. Samples were rinsed with 100 μL of distilled H₂O three times and then were air dried. Imaging was acquired by an Integrated Bruker Innova atomic force microscopy (AFM) in tapping mode with a silicon cantilever (Bruker, FESPA‐V2). The data were visualized by Gwyddion software.

2.7. Electron microscopy

Monomers and oligomers of 4 μM total tau were adsorbed to Formvar‐coated electron microscopy (EM) grids overnight in a hydrated petri dish. Grids were washed with three drops of water, then two drops of 2% uranyl acetate (Electron Microscopy Sciences, 22400) in water. After 1 minute of incubation, the grids were blotted and air dried. The samples were imaged in a Tecnai F20 Twin transmission electron microscope (FEI, now Thermo Fisher Scientific) operating at 120 kV. The digital micrographs were recorded on a TVIPS XF416 camera (Teitz Video and Image Processing Systems).

2.8. Lentiviruses

Human 2N4R tau cDNA was amplified by polymerase chain reaction (PCR) and transferred to the FSW lentiviral vector between the BamHI and HpaI sites. Primers were described previously. ²⁰ The expression plasmids encoding tau and the packaging vectors, (pSPAX2 and pMD2.G; Addgene plasmids 12260 and 12259, respectively) were transfected using Lipofectamine 3000 (Thermo Fisher, L3000001) into HEK293T cells grown in 15 cm Petri dishes to ≈ 80% confluence in Dulbecco's Modified Eagle's medium (Gibco 11965‐092) supplemented with 10% fetal bovine serum (VWR, 89510‐186). Each transfection used 15 μg total DNA at a 50%/37.5%/12.5% ratio of expression vector/pSPAX2/pMD2.G. Lentivirus‐conditioned medium was collected 24 and 48 hours after the start of transfection. Lentiviral particles were concentrated in a Beckman Coulter Optima LE‐80K ultracentrifuge for 2 hours at 95,152 × g_max at 4°C in an SW28 rotor, resuspended in 400 μL Neurobasal medium, and stored at −80°C in small aliquots. Cultured neurons were transduced in Neurobasal/B27 medium and incubated for 48 to 72 hours before assays were performed.

2.9. Brain homogenates

Cerebral hemispheres were dissected from freshly euthanized mice, suspended in three volumes of Neurobasal plus medium at 4° C, and homogenized on ice using 25 pulses of 30 seconds each of a probe sonicator (Misonix Sonicator 3000) at 30% power. Lysates were centrifuged at 21,000 × g_max for 15 minutes with a TLA 120.2 rotor (Beckman Optima TLX Ultracentrifuge) to remove cellular debris and large, insoluble material. Supernatants were then passed through a 0.22 um filter, aliquoted, and stored at −80°C until further use. Total protein concentration was determined by Pierce BCA Protein Assay Kit (Thermo Scientific, 23225).

2.10. Antibodies

All antibodies used in this study are listed in Table S2 in supporting information. It is important to note that the rabbit polyclonal anti‐lamin B1/B2 antibody is sold by Abcam as anti‐lamin‐B1, but it apparently recognizes lamin‐B2 as well. Both lamins have a molecular weight of ≈ 66,000, but as shown in Figure S1 in supporting information, the Abcam antibody labeled bands of ≈ 66 kD and ≈ 50 kD in both mouse and human brain. We examined both bands by mass spectroscopy, which detected lamins B1 and B2 in the upper band, but the lower band contained only lamin B2, which increased in content as a function of post mortem interval (PMI) in mouse brain. The antibody was produced by injection of a 51‐mer lamin B1 peptide that is 60% identical and 67% similar to the corresponding region of lamin B2. It is thus likely that the antibody recognizes lamin B1 and lamin B2, both of which are constituents of the inner nuclear lamina in neurons and many other cell types.

2.11. Immunofluorescence microscopy

All micrographs were acquired using an inverted Nikon Eclipse Ti microscope equipped with a Yokogawa CSU‐X1 spinning disk confocal head with 405 nm, 488 nm, 561 nm, and 640 nm lasers, and 10X and 20X dry, and 40X and 60X oil immersion Nikon Plan Apo objectives. All brain and cultured neuron immunofluorescence micrographs are shown as maximum projections of Z‐stacks produced using the Fiji derivative of ImageJ. Primary and secondary antibodies are listed in Table S2.

Primary mouse neurons growing on coverslips were rinsed once with PBS, and with one exception, were fixed and permeabilized in methanol for 5 minutes at −20°C. The exception was that cells stained with anti‐Ran were fixed with 2% paraformaldehyde for 10 minutes at room temperature and permeabilized with 0.2% Triton X‐100 for 10 minutes. After washing three times with PBS, cultured neurons were incubated with Intercept (PBS) Blocking Buffer (LI‐COR, 927‐70001) /0.1% Tween 20 for 1 hour followed by the indicated primary and secondary antibodies for 30 minutes each. All antibodies were diluted in Intercept (PBS) Blocking Buffer/0.1% Tween20. After each antibody incubation step, the cells were rinsed three times for 5 minutes each with PBS. In the last wash, the coverslips were incubated with 5 μg/mL DAPI (Sigma‐Aldrich, D9542). The coverslips were then mounted onto slides using Fluoromount‐G (Southern Biotech, 0100‐01). To quantify nuclear invagination in primary neurons, lamin B1/B2‐positive pixels were assigned as either nuclear boundary or invaginated using a Fiji thresholding algorithm. The ratio between the area of invaginated lamin B1/B2 over that of the total lamin B1/B2 staining was reported. At least 100 neurons per condition were evaluated for each experiment.

For Video S1 in supporting information, primary tau^−/− neurons were labeled with mouse anti‐pan tau (Tau5) and chicken anti‐MAP2, followed by Alexa Fluor 488 goat anti‐mouse immunoglobulin G and Alexa Fluor 568 goat anti‐chicken immunoglobulin Y. Cells were imaged using the 60X objective, and 26 confocal planes separated from each other by 300 nm were captured. Fiji was then used to convert his Z‐stack of images into an avi video file.

Human paraffin embedded brain sections were first deparaffinized and rehydrated by sequential incubation in Xylenes (2 × 5 minutes), 1:1 xylenes:100% ethanol (3 minutes), 100% ethanol (3 minutes), 95% ethanol (3 minutes), 70% ethanol (3 minutes), 50% ethanol (3 minutes), and distilled water (3 minutes). Antigen retrieval was achieved by microwaving the sections in citrate buffer pH 6.0 (Vector Laboratories, H‐3300) for 15 minutes. After cooling to room temperature, sections were rinsed in PBS and blocked in PBS/5% bovine serum albumin (BSA)/0.1% Triton X‐100 for 1 hour at room temperature and incubated with the indicated primary antibodies diluted in PBS/5% BSA overnight at 4°C. Sections were then washed four times for 5 minutes each in PBS and incubated with the indicated secondary antibodies for 2 hours at room temperature. After washing four times for 5 minutes each in PBS, sections were incubated with DAPI for 10 minutes and then with autofluorescence eliminator (Millipore, 2160) for 5 minutes, followed by three washes with ethanol. Finally, the tissue sections were mounted under #1 thickness coverslips using Fluoromount‐G. Approximately 50 MAP2‐positive nuclei were evaluated in each case. A nucleus was considered invaginated if the total length of its invaginations was at least 10% of the longest axis of the nucleus.

Free‐floating, 50 μm mouse brain sections were rinsed in PBS for 5 minutes and blocked with PBS/5% normal goat serum (Southern Biotech, 0060‐01) for 2 hours at room temperature. Sections were then incubated with the indicated primary antibodies diluted into PBS/2% normal goat serum/0.05% Tween 20 overnight at 4°C. Sections were then washed with PBS/0.05% Tween 20 three times for 10 minutes each and incubated with the indicated secondary antibodies for 2 hours at room temperature. After three washes with PBS, the sections were incubated with DAPI for 10 minutes and autofluorescence eliminator for 5 minutes, followed by three washes with ethanol. After the final wash, the sections were rinsed with PBS and mounted between #1 thickness coverslips and glass slides using Fluoromount‐G. To ensure thorough neuroanatomical coverage, cortical sections from anterior, middle, and posterior regions were imaged and ≈ 300 neurons from each section were evaluated. A nucleus was considered invaginated if the total length of its invaginations was at least 10% of the longest axis of the nucleus.

2.12. Protein electrophoresis and western blots

For cultured neurons, total protein was extracted using RIPA buffer (Thermo Scientific, 89900) supplemented with protease inhibitors (Thermo Scientific, 78430) and phosphatase inhibitors (Thermo Scientific, 78420). Protein samples were mixed with 1X NuPage LDS Sample Buffer (Invitrogen, NP0007) and 1x sample reducing agent (Invitrogen, NP0004). Samples were heated at 70°C for 10 minutes and separated on 4% to 12% gradient Bis‐Tris sodium dodecyl‐sulfate polyacrylamide gel electrophoresis (SDS‐PAGE) gels (Invitrogen, NP0323) and transferred to 0.2 μm nitrocellulose membrane (Bio‐Rad Laboratories, 1620112) for 1 hour at 100 V.

For human brain tissue, 50 to 100 mg of tissue was blended with one scoop of zirconium oxide beads (Next Advance, ZOB05) in 300 μL buffer containing 65 mM Tris‐HCl, pH 6.8, 2.1% SDS, 5% β‐mercaptoethanol, and 0.1% Triton X‐100 for 10 minutes. The lysate was then sonicated for 5 minutes and centrifuged at 14,000 × g for 15 minutes with an accuSpin Micro 17R Centrifuge (Fisher Scientific) at 4°C. Samples were diluted 5‐fold into 1x Laemmli sample buffer (Bio‐Rad Laboratories, 1610747) with 5% β‐mercaptoethanol. The samples were then heated at 95°C for 10 minutes, separated on 4% to 12% gradient Bis‐Tris SDS‐PAGE gels (Invitrogen, NP0323), and transferred to 0.2 μm nitrocellulose membrane (Bio‐Rad Laboratories, 1620112) for 8 hours at 25 V.

Primary and secondary antibodies for western blotting are listed in Table S2. All antibodies were diluted in Intercept (TBS) Blocking Buffer/0.1% Tween‐20. After protein transfer, nitrocellulose membranes were blocked with Intercept (TBS) Blocking Buffer (LI‐COR, 927‐60001) for 1 hour at room temperature and incubated with the indicated primary antibodies overnight at 4°C. Membranes were then washed three times for 5 minutes each in TBST (TBS/0.1% Tween 20) before being incubated with the indicated secondary antibodies at room temperature for 1 hour. Finally, the membranes were washed three times in TBST, and were imaged using a ChemiDoc MP imager and analyzed using Image Lab software (Bio‐Rad).

2.13. xcTauO endocytosis assay

A previously described protocol was adapted to label oligomeric tau. ²¹ EZ‐Link Sulfo‐NHS‐Biotin (2 mg, Thermo Scientific; A39258) was reconstituted in water to 2 mM. TauOs were incubated with the biotin at 1:1 molar ratio (total tau:biotin) for 30 minutes at room temperature or 2 hours on ice. To remove unbound biotin and enable buffer exchange into PBS, the solution was centrifuged four times at 10,000 × g for 15 minutes each in Amicon Ultra‐0.5 Centrifugal Filters (Millipore, UFC501096), with addition of PBS to the solution remaining above the filter at each step. Cultured mouse neurons were exposed to 300 nM biotinylated xcTauOs. After 1 to 3 hours of incubation, biotin‐tagged xcTauOs that remained bound to the outer surface of the plasma membrane were reduced and washed with 50 mM DTT for 1 hour at 37°C. Cells were then washed three times with PBS for 15 minutes and lysed under native conditions with a hypotonic buffer (10 mM Tris HCl, pH 7.4; 1.5 mM MgCl2; 10 mM KCl). Protein samples were mixed with 1X NuPage LDS Sample Buffer (Invitrogen, NP0007) without reducing agents and analyzed by western blotting using Streptavidin‐IRDye 800CW (LI‐COR; 926‐32230) to detect biotinylated tau that had been endocytosed.

2.14. Dextran exclusion assay

A previously described protocol ²² was adapted to primary WT mouse neurons that grew on glass coverslips and were exposed to xcTauOs (250 nM total tau) or vehicle for 24 hours, and then washed once with prewarmed PBS and ice‐cold PBS, respectively. Next, neurons were incubated in permeabilization buffer (20 mM HEPES pH 7.5, 110 mM KOAc, 5 mM MgCl₂, and 0.25 M sucrose) for 5 minutes on ice, followed by a 7 minute incubation with permeabilization buffer containing 20 μg/mL of digitonin (Sigma‐Aldrich, CHR103). Neurons were then washed four times with diffusion assay solution (20 mM HEPES pH 7.5, 110 mM KOAc, 5 mM NaCl, 2 mM MgCl₂, 0.25 M sucrose) on ice and incubated with 1 mg/mL of fluorescein‐dextran (Invitrogen, D1823) for 10 minutes at room temperature. Hoechst 33342 was used as the nuclear counterstain (NucBlue Live ReadyProbe Reagent, Invitrogen, R37605). The glass coverslips were sealed onto slides with clear nail polish and imaged using a 20 × 0.8 NA Plan Fluorite objective on an EVOS M5000 cell imaging system (Thermo Fisher Scientific). Images were quantitatively analyzed using Fiji software: regions of interest (ROIs) defining nuclei region were hand drawn based on Hoechst staining, and the mean pixel intensity of nucleus‐associated fluorescein–dextran was standardized against the mean pixel intensity of the background.

2.15. Specificity validation of an antibody to histone H3 trimethylated at lysine 9 (H3K9Me3)

Specificity of anti‐H3K9Me3 was tested against two types of substrates: recombinant histone H3 (New England BioLabs, M2507S), which lacked methylation, and native histones that contain the modification and were purified from HEK293T cells using a Histone Extraction Kit (Abcam ab113476). Protein concentration was determined using the Pierce BCA assay. Samples were diluted into 1X NuPage LDS Sample Buffer (Invitrogen, NP0007), supplemented 1x sample reducing agent (Invitrogen, NP0004). Then, 2.5 μg of native mixed histones and 625 ng of pure recombinant histone H3 were loaded on the gel.

2.16. Gene expression analysis and quantitative reverse transcription PCR

The nanoString nCounter system was used for quantitatively analyzing gene expression using the Mouse Neuropathology panel of 760 mRNA probes. Total RNA was isolated from cultured mouse neurons treated with xcTauOs (250 nM total tau) or vehicle for 6 hours using the mirVana isolation kit (Invitrogen, AM1560). RNA integrity was evaluated by capillary electrophoresis using an Agilent Bioanalyzer. One hundred nanograms of RNA with a RNA integrity number > 7 was used per sample, and three biological replicates were used for each condition. Data analysis was done using the ROSALIND platform considering a P‐value threshold of 0.05 for differential gene expression; batch effects were considered in the analysis.

Semi‐quantitative real time PCR was used to quantify MAPT mRNA levels in cultured mouse neurons after xcTauO or vehicle treatment for 6 hours. RNA was isolated using Trizol (Invitrogen, 15596026) and MAPT mRNA levels were quantified using One‐Step quantitative reverse transcription (qRT)‐PCR (Invitrogen, 11746100) in an Applied Biosystem StepOne Plus Real‐Time PCR instrument as follows: 3 minutes at 50°C; 5 minutes at 95°C; 40 cycles of 15 seconds at 95°C, 30 seconds at 60°C, and 1 minute at 40°C; and 72°C to 90°C melting analysis. In all cases, Actb and Rpl19 were used as reference genes, and four independent biological replicates with two technical replicates each were quantified. Primer sequences were obtained from Origene (MP208179, MP200232, and MP212857). Primer efficiency was calculated and incorporated in the ΔΔCt method analysis. ²³

2.17. Statistical analysis

Data are presented as mean values of the number of independently conducted experiments indicated in the legend of each figure. Error bars represent the standard error of the mean (SEM). Statistical analysis was performed using Prism 9 software (GraphPad). Statistical tests used for all figures are indicated in the corresponding legends.

3. RESULTS

3.1. Nuclear lamina invagination in human AD and transgenic mouse brains

We first evaluated the morphology of neuronal nuclei in cortices (temporal lobe with hippocampus) from AD and age‐matched cognitively normal individuals (Table S1 shows clinical characterizations). Invaginated neuronal nuclei were identified by surrounding MAP2‐positive cytoplasm and deep infoldings of the nuclear lamina, as determined by anti‐lamin B1/B2 immunofluorescence. Approximately 70% of neuronal nuclei were invaginated in AD brains, whereas only ≈ 30% showed such deformation in cognitively normal brains (Figure 1).

FIGURE 1.

Nuclear lamina invaginations in human cortical (temporal lobe with hippocampus) Alzheimer's disease (AD) neurons in vivo. Left, Immunofluorescence localization of the lamina proteins, lamin B1/B2, in MAP2‐positive neurons, and quantitation of invaginated lamina. Quantitative data were obtained from four AD and four control cases (Table S1 in supporting information), and for each case at least 50 neurons were scored. Significance was determined using unpaired t test, and error bars represent standard error of the mean (SEM). Right, Quantitative western blotting of lamin B1/B2 in prefrontal cortices of human AD and non‐demented control brains. For quantitation, lamin B1/B2 signals were normalized against histone H3 for five AD and four normal cases (Table S1). Significance was determined using unpaired t test, and error bars represent SEM.

We next studied nuclear morphology in two transgenic mouse lines: PS19, which overexpress human P301S tau, and accumulate tangles by 6 months and neuron loss within 9 months; ²⁴ and CVN, which overexpress human APP with the Swedish, Dutch, and Iowa mutations, are knocked out for nitric oxide synthase 2, and develop plaques, tangles, and neuron loss within 12 months. ²⁵ To ensure thorough neuroanatomical coverage, we sampled coronal sections from anterior, middle, and posterior brain regions for PS19 mice (Figure S2 in supporting information), and lateral, middle, and medial sagittal sections for CVN mice (Figure S3 in supporting information). Necab1, which is highly expressed only in layer IV neurons, ²⁶ was used to differentiate neurons in inner cortical layers to those in outer layers (Figure S2A). Compared to age‐matched WT mice, 6‐month‐old PS19 mice had higher levels of invaginated neuronal nuclei, especially in deep cortical layers. Elevated levels of such nuclei were also found in deep cortical layers of lateral brain regions in CVN mice (Figure S3). Altogether, these findings indicate that invaginated neuronal nuclei are abundant in vivo in the human AD brain, and in the brains of transgenic mouse models of AD and a pure tauopathy.

3.2. Human xcTauOs induce nuclear invagination and endogenous tau aggregates in cultured mouse neurons

Previous studies established that pathogenically mutated intracellular tau can provoke neuronal nuclear deformation, ¹³ , ¹⁵ but possible roles for xcTauOs ²⁷ in this process have not been reported. We therefore investigated effects of xcTauOs on neuronal nuclei. All six human central nervous system (CNS) tau isoforms were expressed in bacteria and purified (Figure S4 in supporting information), and oligomerized by addition of arachidonic acid (ARA; Figure 2). ¹⁹ The number of tau subunits in oligomers with apparent molecular weights of ≈ 110 to 130 kDa and ≈ 200 kDa seen on our western blots is unknown, but TauOs can migrate anomalously in SDS‐PAGE, as ≈ 180 kDa oligomers made from 2N4R tau by a similar method are actually dimers as determined by mass spectrometry. ²⁸ Tau monomers and oligomers were also examined by AFM and negative staining EM (Figure S4). Monomeric tau appeared mainly as small fusiform or globular structures with long axes of ≈ 5 to 8 nm by AFM and ≈ 10 to 15 nm by EM, whereas oligomers were much larger and of diverse morphologies. A few oligomer‐like structures were also observed in the tau monomer preprations.

FIGURE 2.

Extracellular tau oligomers (xcTauOs) induce lamina invagination in cultured primary wild type (WT) mouse cortical neurons. A, Criteria for classifying nuclei as invaginated. For each nucleus, each lamin B1/B2‐positive pixel was assigned as either on the boundary or in an invagination. The ratio between the area of invaginated lamin B1 over that of the total laminB1 staining was reported. B, Tau5 western blot of all six recombinant human central nervous system (CNS) tau isoforms oligomerized with arachidonic acid (ARA), and recombinant human 2N4R tau (Figure S4 in supporting information) with (oligomers) and without (monomers) ARA treatment. C, Immunofluorescence localization of lamin B1/B2 in (MAP2‐positive) neurons treated for 1 hour with xcTauOs or tau monomers (250 nM total 2N4R tau in both cases), or vehicle. D, Immunofluorescence localization of lamin B1/B2 in (MAP2‐positive) neurons treated for 1 hour with xcTauOs (250 nM total tau; an equimolar mix of all six human tau CNS isoforms), or vehicle. Significance was determined using Student t test; error bars represent standard error of the mean; n = 3 biological replicates, and at least 100 nuclei were scored per condition for (C) and (D).

Nuclear invagination was induced in neurons exposed to xcTauOs made from 2N4R tau or an equimolar mixture of all six isoforms, but not to tau monomer or vehicle controls (Figure 2). Compared to controls, xcTauOs increased the extent of neuronal  nuclear invagination by 63% to 95%. Dose‐response and kinetic analyses established that xcTauOs made from 250 nM total 2N4R tau caused maximum nuclear invagination within just 1 hour of treatment (Figure S5 in supporting information). Because oligomers represented an average of ≈ 30% of the total tau, and may have been mainly dimers, the total oligomer concentration was actually ≈ 30 to 40 nM. 

FIGURE 3.

TauOs derived from hTau mouse brain induce lamina invagination and extracellular tau oligomers (xcTauOs) cause endogenous tau aggregation. A, Tau5 western blot of hTau^+/− mouse brain extracts containing partially oligomeric or entirely monomeric tau. B, Lamina invagination in primary mouse cortical neurons after a 4 hour exposure to hTau^+/− mouse brain extracts containing tau oligomers (diluted to 0.7% v/v). C, A 1 hour treatment of primary mouse cortical neurons with xcTauOs made from recombinant human 2N4R tau (250 nM total tau) causes aggregation of endogenous, perinuclear mouse tau (white arrows). Significance was determined using Student t test; error bars represent standard error of the mean; n = 3 biological replicates for (B) and (C).

FIGURE 4.

The extracellular tau oligomer (xcTauO) effect requires intracellular tau. A, xcTauOs do not induce lamina invagination in primary cortical neurons derived from tau^−/− mice. Cultures were treated with xcTauOs (250 nM total tau) or vehicle for 1 hour. B, Lentiviral expression of human 2N4R tau in tau^−/− neurons restores nuclear lamina sensitivity to a 1 hour exposure to xcTauOs (250 nM total tau). Significance was determined using Student t test; error bars represent standard error of the mean; n = 3 biological replicates for (A) and (B).

To evaluate the in vivo relevance of xcTauOs assembled from recombinant tau, we also prepared and tested soluble brain extracts from 21‐month‐old hTau^+/− transgenic mice that express approximately equimolar levels of the six CNS tau isoforms encoded by human genomic tau DNA in the absence of full‐length mouse tau. ²⁹ Brain extracts from comparably aged littermates that were null for human tau (hTau^−/−) were used as controls. Western blots revealed that all hTau^+/− brains contained detectable tau, but only some harbored TauOs, typically with a size range of ≈ 100 to 160 kDa (Figure 3A). hTau^+/− brain extracts containing TauOs plus tau monomers, but not monomers alone, potently induced neuronal nuclear invagination (Figure 3B). The results described so far in this section indicate that xcTauOs made from recombinant tau can reliably substitute for TauOs made in vivo in the brain, and that xcTauOs made from recombinant 2N4R tau are as effective as those made from a cocktail of all six recombinant CNS tau isoforms. Accordingly, all other experiments described in this report relied on xcTauOs made from recombinant 2N4R human tau.

In a prior study, we reported that xcTauOs made from recombinant human tau induce accumulation of intracellular tau aggregates in cultured mouse neurons, but we did not confirm the presence of endogenous mouse tau in those aggregates. ³⁰ As shown in Figure 3C, conspicuous, perinuclear aggregates labeled with an antibody that recognizes mouse, but not human, tau were abundant in xcTauO‐treated cultured neurons.

3.3. The effect of xcTauOs on nuclear lamina architecture depends on intracellular tau

A recent report that intracellular TauOs cause translational stress ¹⁶ prompted us to test whether the mouse tau aggregates, formed in cultured mouse neurons after exposure to human xcTauOs (Figure 3C), signal a requirement of intracellular tau for neuronal nuclear invagination induced by the xcTauOs. Accordingly, neurons derived from tau^−/− mice ¹⁷ were exposed to xcTauOs. As shown in Figure 4, neuronal nuclei in tau^−/− neurons did not invaginate after xcTauO exposure, unless human (2N4R) tau was expressed in the cells by lentiviral transduction.

The failure of nuclei to invaginate in tau^−/− neurons exposed to xcTauOs might be due to an inability of the neurons to either endocytose xcTauOs ²¹ , ³¹ , ³² or to respond to them once the oligomers have entered the cells. We therefore used two approaches to determine whether tau^−/− neurons endocytose xcTauOs. First, we stained fixed cultures of primary tau^−/− neuron cultures exposed to xcTauOs with antibodies to tau and to the neuron‐specific protein, MAP2. Next, we used confocal microscopy to capture a series of optical sections that extended through the vertical height of the culture and stitched the images together to create a movie that scrolls through the Z‐dimension of the culture. As shown in Video S1, most of the xcTauOs appeared to be bound to outer surfaces of neurons, but many were clearly located within the cytoplasm.

The other assay involved exposing primary tau^−/– neuron cultures to biotinylated xcTauOs, washing away unbound oligomers and reducing the cell surface–associated biotin with DTT. The cells were then collected for western blotting with fluorescently tagged streptavidin, which detected only the xcTauOs that had been endocytosed. Figure S6 in supporting information illustrates that tau^−/− neurons internalized biotinylated xcTauOs as well as WT neurons. Based on these immunofluorescence and endocytosis assays, we conclude that xcTauO uptake by tau^−/− neurons is not impaired by the lack of endogenous tau, but rather that endogenous tau is somehow required for nuclear invagination induced by endocytosed xcTauOs.

3.4. xcTauOs disrupt nucleocytoplasmic transport in cultured neurons

The nuclear lamina provides a structural framework for nuclear pore complexes (NPCs), which regulate exchange of molecules larger than ≈ 40 kDa globular proteins between the nucleus and the cytoplasm. Nucleocytoplasmic transport is negatively impacted by mutant lamins, ³³ , ³⁴ which prompted us to assess whether it is also functionally compromised by xcTauOs. We first tested this possibility by a dextran exclusion assay, in which digitonin‐permeabilized cultured mouse neurons were exposed to fluorescein‐tagged 70 kDa dextran. In this assay, NPCs that are structurally damaged allow greater penetration of the dextran–fluorescein into the nucleus than undamaged NPCs (Figure 5A). We found that xcTauOs caused ≈ 36% more dextran–fluorescein to enter nuclei of neurons exposed to xcTauOs compared to vehicle (Figure 5B).

FIGURE 5.

Extracellular tau oligomers (xcTauOs) disrupt nucleocytoplasmic transport in primary wild type mouse cortical neurons. A, Schematic illustration of 70 kDa dextran‐fluorescein exclusion assay. B, Neurons treated with xcTauOs (250 nM total 2N4R tau) for 24 hours had leaky nuclear pore complexes (NPCs). C, Ran mislocalization caused by xcTauOs. Cultures were treated with xcTauOs (250 nM total 2N4R tau) or vehicle for 24 hours. Ran mislocalization was measured as the ratio of the mean pixel intensity of the cytoplasmic/nuclear Ran signal. Significance was determined using Student t test; error bars represent standard error of the mean; n = 3 biological replicates for (A) and (B).

Next, we assessed the effect of xcTauOs on active nucleocytoplasmic transport in cultured mouse neurons by measuring the cytoplasmic/nuclear distribution of Ran, a small GTPase that reversibly shuttles between the nucleus and cytoplasm, but is normally highly enriched in nuclei at steady state. ³⁵ , ³⁶ As shown in Figure 5C, xcTauOs caused an ≈ 26% increase in the cytoplasmic/nuclear Ran ratio relative to a vehicle control. Together, these dextran–fluorescein and Ran results demonstrate that xcTauOs functionally impair nucleocytoplasmic transport in neurons.

3.5. xcTauOs alter histone methylation and gene expression

Because the nuclear lamina serves as a tether for heterochromatin, we hypothesized that alterations of chromatin structure and gene expression accompanies the morphological distortion of neuronal nuclei caused by xcTauOs. To test that hypothesis, we first evaluated whether xcTauOs alter cultured mouse neuron levels of trimethylated lysine 9 of histone H3 (H3K9me3), which promotes transcriptional silencing. Figure S7 in supporting information shows that on western blots, the anti‐H3K9me3 antibody used for these experiments was immunoreactive with bulk cell–derived histones, which include H3K9me3, but not with recombinant histone H3, which lacks post‐translational modifications. This antibody also revealed by western blotting that the H3K9me3/total histone 3 ratio increased by ≈ 52% after xcTauO exposure (Figure 6A).

FIGURE 6.

Extracellular tau oligomers (xcTauOs) alter gene expression in primary mouse cortical neurons. A, Western blot of H3K9Me3 for cultures treated with xcTauOs (250 nM total 2N4R tau) or vehicle for 24 hours. Significance was determined using unpaired t test; error bars represent standard error of the mean (SEM); n = 3 biological replicates. B, Volcano plot and heat map showing expression level changes detected by nanoSTRING nCounter technology for a panel of probes for 760 genes implicated in neurodegeneration. Cells were exposed to xcTauOs (250 nM total 2N4R tau) or vehicle for 6 hours; n = 3 biological replicates. C, Quantification by quantitative reverse transcription polymerase chain reaction of MAPT mRNA levels in cells treated with xcTauOs (250 nM total 2N4R tau) or vehicle for 6 hours. Significance was determined using unpaired t test; error bars represent SEM; n = 4 biological replicates.

Next, we investigated whether xcTauOs alter gene expression in cultured neurons using the nCounter platform from nanoString for direct, unamplified measurement of 760 mRNAs associated with neuropathology. We found that for WT neurons, xcTauOs significantly altered mRNA levels for 15 genes, most of which were upregulated (Figure 6B). Annotation analysis of the differentially expressed genes highlights transcription functions based on genes like Gtf2b, Srsf4, Prpf31, and Cdk7. In contrast, treatment of cultured tau^−/− neurons with xcTauOs yielded a different pattern: more genes were downregulated than upregulated, and none of genes affected in WT neurons were affected in tau^−/− neurons (Figure S8 in supporting information). Interestingly, the most upregulated gene in WT neurons after xcTauO treatment was MAPT, which encodes tau (fold change = 3.93; Figure 6B). However, MAPT elevation did not reach the significance threshold (P < 0.20), probably because of the high variation observed in this analysis. We therefore used an alternate method, qRT‐PCR, to evaluate MAPT mRNA levels in a separate cohort of WT neuron cultures. That approach detected a statistically significant 2.84‐fold increase of MAPT mRNA after xcTauO treatment of WT neurons (Figure 6C).

4. DISCUSSION

It is well established that tau can be released into the extracellular space by neurons independently of cell death and coupled to neuronal activity. ³⁷ Much effort has focused on mechanisms of pathological tau transfer from neuron to neuron, but few cell biological responses of neurons to toxic tau uptake have been reported. We now provide evidence that xcTauOs provoke rapid structural and functional changes in neuronal nuclei. Within 1 hour of cultured neuron exposure to xcTauOs, their nuclei invaginate (Figures 2, 3, 4 and S5), and by no later than 24 hours of exposure, nucleocytoplasmic transport is impaired (Figure 5), and H3K9Me3 levels rise and gene expression changes are evident (Figure 6). Intriguingly, the most prominent gene expression change is a nearly 3‐fold increase in tau mRNA levels, which might signal a positive feedback loop whereby toxic TauOs drive their own expansion fueled not only by prion‐like propagation, ⁸ , ⁹ , ¹⁰ but by a growing pool of tau monomers as well. The likely in vivo significance of these results for cultured neurons is emphasized by our finding that invaginated neuronal nuclei are increased in the human AD brain (Figure 1), consistent with a prior report, ¹⁵ and in mouse models of AD (Figure S3) and a pure tauopathy (Figure S2). Together, these results implicate xcTauOs in the rapid transformation of healthy neurons into diseased neurons by changing nuclear architecture, and by extension, nuclear function.

Various forms of oligomerized tau were found to induce nuclear invagination in cultured neurons. Oligomers made from recombinant human 2N4R tau or an equimolar mix of all six human CNS tau isoforms were equally effective at 250 nM total tau (Figure 2C, D). Because only approximately one third of the total tau was typically oligomerized (Figure 2B), however, with dimers suspected of being most prevalent, the effective concentration of active oligomers was probably in the range of 30 to 40 nM. It is also noteworthy that TauOs made from all six human tau isoforms in hTau mice potently induced nuclear invagination (Figure 3A). This robust activity of TauOs made in vivo validates our use of TauOs made from recombinant human 2N4R tau for most experiments, especially because the activity of 2N4R TauOs was indistinguishable from those made from a mixture of all six recombinant human CNS tau isoforms.

Mechanistic insight into how xcTauOs cause nuclear invagination was provided by our finding that intracellular tau is required (Figure 4). In light of tau's known prion‐like behavior, ⁸ , ⁹ , ¹⁰ it is significant that xcTauOs induced formation of perinuclear aggregates of endogenous tau in cultured mouse neurons (Figure 3C). Detailed structural analysis of these tau aggregates, and their potential roles in deforming and functionally compromising neuronal nuclei, awaits further study.

A peak level of nuclear deformation in cultured neurons was reached within 1 hour of initial exposure to xcTauOs and was stable for at least 24 hours (Figure S5). By that time point, and perhaps even earlier, a constellation of cell biological effects beyond nuclear deformation was evident. One example is impaired nucleocytoplasmic transport (Figure 5), which occurs at the NPC. That finding is consistent with earlier reports of tau tightly associated with the NPC in frontotemporal dementia ¹³ , ³⁵ and directly interacting with Nup98, an NPC structural protein, to cause defective nucleocytoplasmic transport. ¹³ , ³⁵ This collection of results from our labs and others has profound implications. For example, multiple transcription factors, including p53, mislocalize from the nucleus to the cytoplasm in AD, ³⁸ perhaps because of nucleocytoplasmic transport disruption by xcTauOs.

Dysregulation of chromatin organization is closely associated with the aberrant accumulation of toxic tau species. ³⁹ , ⁴⁰ , ⁴¹ We also found that xcTauOs increase the level of H3K9Me3, which epigenetically drives heterochromatin condensation and is elevated in the AD brain. ⁴² Given the role of H3K9Me3 in controlling chromatin organization, it is not surprising that we observed multiple changes in gene expression after cultured neuron exposure to xcTauOs (Figure 6). We first explored fluctuations in mRNA levels using the nanoString nCounter platform with a neuropathology panel that directly quantifies mRNAs encoded by 760 genes. Differential gene expression analysis after 6 hours of xcTauO exposure revealed 14 significantly upregulated mRNAs and 1 that was significantly downregulated. Several of those upregulated genes, like general transcription factor IIB (Gtf2b), Ser/Arg‐splicing factor 4 (Srsaf4), and cyclin dependent kinase 7 (Cdk7), are associated with transcription and splicing annotations. Consistent with our finding that nuclear invagination by xcTauOs requires intracellular tau (Figure 4; see also Figure S6 and Video S1), none of the genes whose transcription levels were affected by xcTauOs in WT neurons was altered in tau^−/− neurons (Figure S8 in supporting information).

Remarkably, the most upregulated gene in WT neurons after xcTauO treatment was MAPT, which encodes tau itself. Its upregulation did not reach statistical significance for the nCounter experiments, though, because one of the three biological replicates yielded an outlier result. Accordingly, we also used qRT‐PCR to measure tau mRNA in four biological replicates, all of which were distinct from the samples analyzed using nCounter. That approach indicated a statistically significant ~3‐fold increase in MAPT mRNA levels because of xcTauO treatment of WT neurons. When considered together, the nCounter and qRT‐PCR results point to the likelihood that xcTauOs induce a large increase in mRNA for tau. In such a scenario, xcTauOs might trigger a positive feedback loop that stimulates production of excess tau mRNA, and by extension, more tau protein and toxic TauOs.

AUTHOR CONTRIBUTIONS

Xuehan Sun co‐conceived, co‐designed, and performed the bulk of the research, and co‐analyzed the data. Yu Shi and Subhi Saibaba performed and co‐analyzed the immunohistochemistry data. Joseph A. Thompson collected the AFM data. Michael D. Purdy and Kelly Dryden collected the EM data. Guillermo Eastman and Ana K. Oliveira performed and co‐analyzed the nCounter data. James W. Mandell provided annotated human brain samples for immunohistochemistry and assisted with interpretation of the corresponding micrographs. John R. Lukens provided PS19 brain sections, and along with Andrés Norambuena, provided ongoing counsel about experimental design and data interpretation. Evelyn Pardo designed and performed the xcTauO endocytosis assay. George S. Bloom co‐conceived, co‐designed, and supervised the research. Xuehan Sun, Guillermo Eastman, and George S. Bloom wrote the manuscript. All authors participated in editing the manuscript and approved the submitted version.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to report. Author disclosures are available in the supporting information

CONSENT STATEMENT

Institutional approval for use of archival autopsy tissue was obtained from the University of Virginia Biorepository and Tissue Research Facility. Additional institutional review board approval was not required.

Supporting information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Supporting Information

Sun X, Eastman G, Shi Y, et al. Structural and functional damage to neuronal nuclei caused by extracellular tau oligomers. Alzheimer's Dement. 2024;20:1656–1670. 10.1002/alz.13535