Introduction
Tauopathies, including Alzheimer’s disease (AD) are a heterogeneous group of progressive neurodegenerative disorders characterized by severe cognitive impairment, memory loss, and aberrant behavioral patterns [
1,
2]. The neuropathological hallmarks of tauopathies include cerebral atrophy, hyperphosphorylation and aggregation of tau filaments into intracellular neurofibrillary tangles (NFTs), and chronic neuroinflammation. In tauopathies, post-translational modifications (PTM) of tau lead to the loss of its physiological function and subsequent progressing assembly of tau into insoluble aggregates [
3]. Although the mechanism of tau aggregation is not fully understood, several factors, such as phosphorylation, acetylation, truncation, oxidative stress have been suggested to be able to facilitate tau fibrillization [
4].
Growing evidence suggests critical involvement of lipid dyshomeostasis in pathogenesis of tauopathies [
5,
6]. Lipids are the dominant structural component of the brain and each lipid class has specific functions. Metabolism of brain lipids is closely linked to brain energy homeostasis. Under healthy conditions, lipids are important components of cellular membrane bilayer and participate in cellular transport, energy storage, regulation of growth and differentiation of cells. Lipids provide a hydrophobic matrix protein-lipid/protein- protein interactions and act as essential signaling molecules and a key modulators of signal transduction. Under pathological conditions, deregulation of brain lipid composition results in disrupted blood- brain barrier, dysfunction in endocytosis, exocytosis and autophagocytosis, altered myelination, unbalanced energy metabolism and enhanced inflammation [
7‐
10].
The role of tau protein in the regulation of lipid metabolism is much less characterized and not well understood. Recent studies have shown that membrane lipids like phosphatidylcholine (PC), cholesterol, and sphingolipid interact with tau protein and actively regulate its fibrillization [
6,
11]. Phosphatidylglycerol (PG) synthesized in mitochondria is involved in activation of protein kinase C (PKC). It is possible to assume that increased activity of PKC may be involved in tau hyperphosphorylation and tangle formation [
12,
13].
Lipid metabolism in the brain is also closely linked to oxidative stress and neuroinflammation. Lipid accumulation in neuroglial cells affected normal neuronal activity and induced activation of microglia cells, followed by increased expression of proinflammatory mediators such as TNF-α and IL-6 [
14]. Accumulation of lipids and impaired lipid metabolism are associated with production of lipotoxic metabolites that may further trigger progression of neurofibrillary pathology.
Both metabolites and lipids reflect the physiological and pathological status of the organism. In this study, we compared samples of transgenic (Tg) SHR-24 rat model expressing human truncated tau protein with age-matched control SHR rats. The SHR-24 model was created to study the pathophysiological effects of 3R truncated tau protein (aa151-391/3R). SHR-24 rats express a tau fragment truncated at the N- and C-terminals, containing three microtubule-binding domains and a proline-rich region on an SHR background. Expression of the transgene is driven by the mouse Thy1 promoter. This animal model has disease onset already at eight months (sensorimotor deficits and abnormal reflexes) of age and survives up to 15 months of age and is characterized by the presence of progressive neurofibrillary pathology in the brainstem and during the pre-terminal and terminal stages also in the frontal cortex [
15]. Moreover, activated of microglia and astrocytes are found in the areas of tau pathology [
16]. To monitor disease progression, we analyzed brain tissue (medulla oblongata, pons), CSF and plasma samples from rats aged 4, 6, 8, 10, 12, and 14 months. For phenotypic profiling, we employed targeted metabolomic and lipidomic analysis, followed by characterization of tauopathy markers. Here, we showed that lipid changes have been initiated in the early stages of tau pathology before the formation of high-molecular-weight tau aggregates. Our results indicate that lipids not only accumulate as a result of aberrant metabolic processes caused by pathological tau protein but may also contribute to aberrant protein aggregation and progression of neurofibrillary pathology. Even our multi-omics approach reveals and quantifies specific and sensitive metabolites and lipids suitable for new therapeutic interventions and playing a role in cellular pathways for early disease detection and progression of tau pathology.
Methods
Chemicals
Acetonitrile, methanol, isopropanol, water, methyl-tert butyl ether and ammonium acetate (LC-MS grade) were purchased from Sigma Aldrich (St. Louis, MO, USA), internal standard mixture SPLASH® LIPIDOMIX® was obtained from Avanti Polar Lipids (Alabaster, Alabama, USA) and internal standards: Creatine-d3 (methyl-d3) from CDN isotopes (Pointe-Claire, QC, Canada), L-Leucine-5,5,5-d3 and Butyryl-L-carnitine-(N-methyl-d3) hydrochloride, both from Sigma Aldrich (St. Louis, MO, USA).
Animals
All experiments were performed on Tg SHR-24 rats expressing human truncated tau aa151-391/3R. As controls, we used age-matched non-Tg animals (SHR). Animals included in this study were 4, 6, 8, 10, 12, and 14 months old. All animals were bred in the in-house animal facility of the Institute of Neuroimmunology of the Slovak Academy of Sciences in Bratislava. All animals were housed under standard laboratory conditions with free access to water and food and were kept under diurnal lighting conditions (12 h light/dark cycles with light starting at 7 am). All experiments were performed according to the institutional animal care guidelines and following international standards (Animal Research: Reporting of In Vivo Experiments guidelines) and approved by the State Veterinary and Food Administration of the Slovak Republic (Ro-933/19–221/3a) and by the Ethics Committee of the Institute of Neuroimmunology, Slovak Academy of Sciences. Efforts were made to minimize the number of animals utilized and limit discomfort, pain, or any other suffering of the experimental animals in this study.
Human tissue specimens
Human brain tissue samples were obtained from the Queen Square Brain Bank for Neurological disorders (London, UK). All samples were obtained with informed patient consent and approval from the local ethical committees.
Collection of samples
For the biochemical, metabolomics, and lipidomics experiments, animals were deeply anesthetized with tiletamine-zolazepam/xylazine anesthesia (4/2 mg/kg). From every rat, we collected plasma, CSF, and brain tissue. CSF samples were collected from cisterna magna in an amount ranging from 80 to 120 µL. Anesthetized animals were fixed in a head holder, and a midline incision in the skin was made up to the head area to permit easy access to the cisterna magna. After the centrifugation (4,000xg, 10 min, 4 °C) samples were immediately flash-frozen in liquid nitrogen and stored at -80 °C until further use. Approximately 4 ml of blood was collected from the heart to vacuum blood collection tubes with EDTA and centrifuged (4,000xg, 10 min, 4 °C). Collected plasma samples were immediately frozen in liquid nitrogen and stored at -80 °C. Brain tissue was cut sagittally and divided into different anatomical regions – pons and medulla oblongata. Every brain region, the left and right hemispheres, was transferred into separate tubes, flash-frozen immediately in liquid nitrogen, and stored at -80 °C.
Sample preparation
Firstly, each tissue was homogenized using homogenizer FastPrep-24 in 300 µL of methanol spiked with an internal standard mixture SPLASH® LIPIDOMIX® (2.5%). After that, 1 ml of methyl-tert butyl ether was added, and the sample was vortexed for 10s. Then, 300 µL of water was added to form a two-phase system, vortexed, and left vibrating for 15 min at room temperature (RT). The samples were subsequently left to equilibrate for 15 min at 4 °C then centrifuged at 10,000xg for 10 min at 4 °C. At last, 400 µL of upper phase was collected for lipid analysis, and a mixture of upper and lower phase (1:1 v/v) for analysis of metabolites. Collected extracts were left to freeze at -80 °C for 20 min and freeze-dried overnight. Freeze-dried samples for lipid analysis were reconstituted in 100 µL of acetonitrile/isopropanol/water (2:2:1 v/v/v), centrifuged at 10,000xg for 10 min at RT and 80 µL of supernatant was collected and transferred into LC-MS grade glass vials. For the preparation of quality control (QC) samples, 5 µL of each supernatant was mixed to form a pooled sample. Freeze-dried samples for metabolomic analysis were reconstituted in a 100 µL mixture of water/methanol (1:4 v/v). The methanolic part of the mixture contained isotopically labeled internal standards: leucine (2.0 µM), butyryl carnitine (0.2 µM), and creatinine (2.0 µM). After centrifugation (15,000xg, 15 min, 4 °C) supernatants were transferred into vials, and a QC sample was prepared the same as in the sample preparation procedure for lipid analysis. CSF and plasma samples for lipid analysis were prepared by adding 80 µL of isopropanol and 2 µL of internal standard mixture SPLASH® LIPIDOMIX® (2.5%) to 20 µL of each sample, vortexed and kept overnight at -80 °C for deproteinization. After centrifugation (15,000xg, 10 min, RT) 80 µL of supernatant was collected and transferred into a glass vial which was used for the subsequent LC-MS analysis. For analysis of metabolites, CSF and plasma samples were prepared by adding 20 µL of sample to 80 µL of the methanolic solution containing selected labeled internal standards, vortexed, and kept overnight at -80 °C for deproteinization. Subsequently, mixtures were vortexed and centrifuged (15,000xg, 5 min, 4 °C). The supernatant was transferred into vials and analyzed. The pooled QC samples for each sample type were prepared by collecting 5 µL of each sample supernatant and pooled into one.
LC-MS analysis
Targeted lipidomic analysis was carried out by a pseudotargeted approach adopted from [
17] using a liquid chromatography-tandem mass spectrometry system consisting of ExionLC™ for UHPLC and QTRAP® 6500 + for MS/MS (Sciex, USA). A reversed-phase ACQUITY BEH C8 column (2.1 × 100 mm, 1.7 μm, Waters) was used for chromatographic separation. Mobile phase A consisted of acetonitrile/water (60:40 v/v), and mobile phase B consisted of isopropanol/acetonitrile (90:10 v/v) both containing 10 mM ammonium acetate. The column temperature was maintained at 55 °C and the sample manager temperature was set at 15 °C. Gradient starting conditions were 32%B (0–1.5 min) with a gradual increase to 85%B (1.5–15.5 min), and a further steep rise to 97%B (15.5–15.6 min). 97%B was maintained until 18.0 min followed by a re-equilibration step at 32%B (18–20 min). The total run time was 20 min at a flow rate of 0.35 mL/min. Mass spectra were acquired simultaneously in both positive and negative modes in the scheduled multiple-reaction monitoring (scheduled MRM) mode. The settings for the mass spectrometer were as follows: capillary voltages were set to + 5500 V/-4500 V, the pressure of curtain gas to 40 psi, and the source temperature was 500 °C. Samples were measured in randomized order with continuous monitoring of QC samples. Confirmation of correct identification of lipids was evaluated using lipid elution pattern plots, generated by R script and the same workflow as in Drotleff, 2020 [
18]. Targeted metabolomic analysis adopted from Yuan, 2012 [
19] and performed using the same LC-MS instrumentation as in the lipidomic analysis. Identification of metabolites and optimization of their measurement parameters were based on standard compounds as was already described in our previous work [
20]. The aminopropyl column (Luna 3 μm NH
2, 2 × 100 mm, Phenomenex) was used for the chromatographic separation. Mobile phase A consisted of 20 mM ammonium acetate (pH 9.75), and acetonitrile was used as mobile phase B. The column temperature was maintained at 35 °C. Gradient starting conditions were as follows: 95% − 10%B (0–7 min), 10% B (7–13 min), 95%B (13–17 min). The total analysis time was 17 min with a flow rate of 0.3 mL/min. Mass spectra were acquired in both positive and negative modes in the scheduled MRM mode. The settings for the mass spectrometer were as follows: capillary voltages were set to + 5500 V/-4500 V, the pressure of curtain gas to 40 psi, and the source temperature was 400 °C. The data were acquired using Analyst software (v1.6.2) and processed using MultiQuant v3.0 (both Sciex, USA) for both metabolomics and lipidomics. All samples were measured in randomized order with continuous monitoring of QC samples.
Neurofilament-light-chain and tau quantification
Concentrations of neurofilament light chain (NFL) in plasma and tau proteins (total-tau) in CSF and cell culture media were measured using single-molecule array (Simoa) digital ELISA, using an HD-1 Analyzer. For the analysis, Simoa™ NF-Light Advantage Kit, and Simoa™ Mouse TAU Discovery Kit were used. Briefly, frozen blinded samples were melted, centrifuged for 10 min at 25 °C for 20,000xg, and supernatants were transferred into a Simoa sample plate together with calibrators for NFL and Mouse TAU kits (parts of kits). Each measurement was done in duplicate according to the manufacturer’s recommendations. Concentrations were calculated using Simoa™ HD-1 instrument software. The concentration of human tau released into cell culture media (proline-rich domain) was measured by in-house developed ELISA assay using the DC116 (aa 188–224) and DC242 (aa156-180) antibodies (Axon Neuroscience R&D Services SE, Bratislava, Slovakia).
Cultivation and LPS stimulation of BV2 cells
Mouse microglial BV2 cells (C57BL/6, purchased from ICLC, Modena, Italy) were cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM, PAA Laboratories GmbH, Colbe, Germany) containing 10% fetal calf serum (FCS, Invitrogen, Carlsbad, California, USA), and 2 mM L-glutamine (PAA laboratories GmbH, Colbe, Germany) at 37 °C and 5% CO2. The medium was changed twice a week. One day before treating the BV2 cells with recombinant tau protein (aa 151–391/3R), the cell culture medium was replaced with a serum-free DMEM medium with L-glutamine. To stimulate the cells, tau protein was added for a final molarity of 500 nM for 24 h.
Isolation and cultivation of primary rat glial culture
Rat mixed glial culture was prepared from cerebral cortices of 1–2 day old Sprague Dawley rats (n = 8/isolation). The animals were euthanized by CO2, and the cerebral cortices were dissected, stripped of the meninges, and mechanically dissociated by repeated pipetting followed by passage through a 20 μm nylon mesh (BD Falcon, Franklin Lakes, USA). Cells were plated on 6-well plates pre-coated with poly-L-lysine (10 µg/mL, Sigma-Aldrich, St. Louis, MO) and cultivated in DMEM medium containing 10% FCS and 2 mM L-glutamine at 37 °C, 5% CO2 in a water-saturated atmosphere.
Development of a multi-component cell model system
The multi-component model was composed of rat primary microglia, astroglial cells, and human neuroblastoma cell line SH-SY5Y with inducible expression of truncated tau protein (previously described by [
21]). Firstly, the expression of truncated tau protein was induced by cultivating cells in a medium without doxycycline (Sigma-Aldrich, St. Louis, MO) for three days before cell seeding into co-culture. Secondly, SH-SY5Y cells were co-cultivated with primary glial culture in 6-well plates in density 1.10
3 cells/cm
2 for a further 5 days in DMEM medium containing 10% FCS, 2 mM L-glutamine and N2-supplement (Life Technologies Invitrogen, Carlsbad, CA) at 37 °C and 5% CO
2. The medium was changed twice a week. 24 h before the experiment, we cultivated the cells in a medium with 2% fatty acid-free BSA (Life Technologies Invitrogen, Carlsbad, CA).
Isolation of lipid droplet-associated mitochondria (LDM)
The brainstem from Tg and control rats (n = 4) was harvested and washed with phosphate buffer saline (PBS, 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4) to remove blood contamination. The brain tissue was suspended in sucrose-HEPES-EGTA buffer supplemented with 2% BSA (250 mM sucrose, 5 mM HEPES, 2 mM EGTA, 2% fatty acid-free BSA, pH 7.2). The tissue suspension was homogenized and the homogenate was transferred into a 15 mL falcon tube. Homogenate was centrifuged in a swinging bucket rotor at 900xg for 10 min at 4 °C. The post-nuclear supernatant was collected and layered with Buffer B (20 mM HEPES, 100 mM KCl, 2 mM MgCl2, pH 7.4). The second centrifugation was carried out in a swinging bucket rotor at 2,000xg for 40 min at 4 °C to allow the fat layer to separate. The layer that contains LDs appears as a translucent band on the top of the gradient. This layer was resuspended in HEPES buffer without BSA and centrifuged at 10,400xg for 10 min at 4 °C in a fixed rotor. The LDM pellet was resuspended in mitochondrial resuspension buffer (250 mM mannitol, 5 mM HEPES, 0.5 mM EGTA, pH 7.4).
ATP assay
The assay was performed using a Luminescent ATP detection assay kit from Promega (Wisconsin, USA). 50 µL of isolated lipid droplet mitochondria (LDM) were suspended in 50 µL of 20 mM Tris pH 7.5 buffer and combined with 100 µL of CellTiter-Glo® 2.0 Reagent in an ELISA standard plate. The plate was placed on an orbital shaker for 5 min, and luminescence was recorded immediately thereafter using Luminometer, Fluoroscan Ascent FL (ThermoScientific, US).
AlamarBlue assay
AlamarBlue™ solution was added to the cell culture in dilution 1/10. Cells were incubated at 37 °C and 5% CO2 for 4 h. 100 µl aliquots of each sample were transferred to replicate wells of a black 96-well plate. The fluorescence was measured at 590 nm, using an excitation wavelength of 530–560 nm.
Immunocytochemistry and BODIPY staining
BV2 and primary rat glial cells were fixed for 15 min in 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) and washed with PBS. Cells were blocked for 60 min in 5% BSA in PBS. Cells were washed with PBS and incubated with primary antibody rabbit polyclonal anti-IBA-1 (FUJIFILM Wako Chemicals, USA) overnight. After washing, the cells were stained with BODIPY 493/503 (1:1000, Thermo Fisher) for 15 min at RT. Cells were mounted and examined using an LSM 710 confocal microscope (Zeiss, Jena, Germany).
Immunohistochemistry
Animals were anesthetized by tiletamine-zolazepam/xylazine anesthesia and perfused intracardially with PBS with heparin. The brain was removed and embedded in a cryostat embedding medium (Leica, Wetzlar, Germany) and frozen above the surface of liquid nitrogen. 12-µm-thick brain sections were cut on a cryomicrotome, fixed onto poly-L-lysine coated slides, and left to dry at RT for 1 h. Sections were fixed for 15 min in 4% paraformaldehyde and blocked for 60 min in a blocking solution (DAKO, Ontario, Canada). Sections were incubated overnight in a polyclonal rabbit anti-IBA-1 primary antibody (1:1000, Wako), polyclonal rabbit anti-GLUT3 primary antibody (1:100, Invitrogen, Massachusetts, USA), and monoclonal mouse anti-NeuN primary antibody (1:1000, Abcam, Waltham, MA, USA). After washing, the sections were incubated in goat anti-rabbit or anti/mouse AlexaFluor546/488 secondary antibodies (1:2000, Invitrogen, Massachusetts, USA) for 1 h at RT. Sections were washed with PBS and incubated with BODIPY 493/503 in PBS (1:1000, Invitrogen, Massachusetts, USA) for 15 min at RT. For tau staining, sections were incubated in monoclonal mouse anti-rat pSer202/pThr205 primary antibody (1:1000, Invitrogen, Massachusetts, USA). After washing, the sections were incubated for 1 h in a secondary antibody goat anti-mouse AlexaFluor488 (1:1000, Invitrogen, Massachusetts, USA). Brain slices were mounted (Vector Laboratories, Burlingame, CA, USA) and examined using an LSM 710 confocal microscope. Oil Red O (Abcam, Waltham, MA, USA) was used for the staining of neutral lipids. The slides were incubated in propylene glycol for 5 min and then in heated Oil Red O solution for 10 min. The sections were differentiated in 85% propylene glycol for 1 min and washed twice in water. The brain slices were mounted in an aqueous mounting medium.
Measurement of membrane fluidity
Membrane fluidity was measured by lipophilic pyrene probe - pyrenedecanoic acid (PDA, Abcam, Waltham, MA, USA), which undergoes dimerization after the interaction and exhibits changes in its fluorescent properties. BV2 cells were treated with tau protein (final concentration 1 μm) for 24 h. The cells were incubated with PDA/Pluronic F127 solution for 1 h at 25 °C in the dark with gentle agitation. The unincorporated PDA was removed by washing cells twice with cultivation media. The live-cell fluorescence microscopy was performed. Changes in cell membrane lipid order between the monomer gel/liquid-ordered phase (fluorescent at 430 ∼ 470 nm) and the excimer liquid phase (fluorescent at 480 ∼ 550 nm) were measured by confocal microscopy.
Preparation of sarkosyl-insoluble PHF-tau from tg rat brain
Purified sarkosyl-insoluble PHF-tau from Tg rat brains was isolated according to a previously published protocol [
22]. For isolation of sarkosyl-insoluble tau, we used brain tissue from 4, 6, 8, 10, 12, and 14-month-old Tg animals (
n = 6). Briefly, 20 mg of the medulla and pons from a Tg rat brain enriched in PHF-tau were dissected, cleaned of blood vessels and meninges, and used as a starting material. The brain tissue was homogenized on ice in 10 volumes of ice-cold SL buffer (20 mM Tris- HCl, pH 7.4, 800 mM NaCl, 1 mM EGTA, 1mM EDTA, 0.5% β-mercaptoethanol, 10% sucrose, 1x protease inhibitors Complete, EDTA-free). The homogenate was centrifuged for 20 min at 20,000xg. Solid sarkosyl (Sigma-Aldrich, St. Louis, MO, USA) was added to the supernatant to achieve a 1% concentration. The samples were stirred for 1 h at RT. The samples were spun for 1 h at RT in a Beckman ultracentrifuge using SW 120 Ti rotor (Beckman Coulter, Inc., Brea, CA, USA) at 100,000xg. After sarkosyl extraction and ultracentrifugation, pellets containing PHF-tau were dissolved in a 1 x SDS sample loading buffer and boiled for 5 min. Samples were determined by following western blot analysis.
Biochemical Western blot analysis
Samples were separated on 12% SDS-polyacrylamide gels and transferred to a nitrocellulose membrane in 10 mM N-cyclohexyl-3-aminopropanesulfonic acid (CAPS, pH 11, Roth, Karlsruhe, Germany). The membranes were blocked in 5% milk in Tris-buffered saline with 0.1% Tween 20 (Sigma-Aldrich, St. Louis, MO, USA) (TBS-T, 137 mM NaCl, 20 mM Tris-base, pH 7.4, 0.1% Tween 20) for 1 h and incubated with primary antibody overnight at 4 °C. As phospho-dependent anti-tau antibodies, we used: mouse monoclonal anti-rat pThr181 (1:1000, Invitrogen Life Technologies, Carlsbad, CA), mouse monoclonal anti-rat pSer202/pThr205 (AT8, 1:1000, Invitrogen Life Technologies, Carlsbad, CA), mouse monoclonal anti-rat pThr231 (1:1000, Invitrogen Life Technologies, Carlsbad, CA), mouse monoclonal anti-rat pThr181 (1:1000, Invitrogen Life Technologies, Carlsbad, CA), rabbit polyclonal anti-rat pSer199 (1:1000, Invitrogen Life Technologies, Carlsbad, CA), rabbit polyclonal anti-rat pSer262 (1:1000, Invitrogen Life Technologies, Carlsbad, CA), mouse monoclonal anti-rat DC217 (1:200, Axon Neuroscience R&D SE, Bratislava, Slovakia) and polyclonal anti-rat pThr212 (1:1000, Invitrogen Life Technologies, Carlsbad, CA). For total tau, we used a mouse monoclonal anti-rat DC25 antibody (recognizing epitope 368–376, 1:200, Axon Neuroscience R&D SE, Bratislava, Slovakia), GFAP (1:1000, Abcam, Cambridge, UK), IBA-1 (1:1000, FUJIFILM Wako, USA), AQP4 (1:500, Abcam, Cambridge, UK). Membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies in TBS-T (1:3000, Dako, Glostrup, Denmark) for 1 h at RT. Immunoreactive proteins were detected by chemiluminescence (SuperSignal West Pico Chemiluminescent Substrate, Thermo Scientific, Pittsburgh, USA) and the signals were digitized by Image Reader LAS-3000 (FUJIFILM, Bratislava, Slovakia). The signal was semi-quantified by ImageJ software.
Quantitative analysis of immunocytochemical and immunohistological data
Relative staining patterns and intensity of projections from immunohistochemical and immunocytochemical stainings were visualized by confocal microscopy and evaluated. ImageJ (public domain ImageJ software) was used for the evaluation and quantification. We quantified 10 slices from each sample. For semiquantitative analysis, the color pictures were converted to grayscale 8-bit TIFF file format, and regions of interest were analyzed. The grayscale 8-bit images were converted to 1-bit images, on which the number of immunolabeled structures localized in the area of interest was measured. The average intensity/pixel values of each area were then calculated, and the average intensity/pixel values representing the background intensity were subtracted from those of the immunolabeled areas.
Cytokine determination by Bio-plex assay
100 mg of the brainstem from SHR-24 and control animals (8-month old) was homogenized in 1 ml of ice-cold PBS with protease inhibitors complete EDTA-free (Roche, Mannheim, Germany). The protein concentration of brain extracts was determined by Bio-Rad protein assay (Bio-Rad Laboratories GmbH, Germany). 600 µg of total proteins were loaded per well. The plasma and CSF were diluted according to the recommendation of the manufacturer. The level of cytokines was measured by Bio-Plex Pro rat cytokine 23-plex assay (Bio-Rad Laboratories, California, USA) according to the manufacturer’s protocol. The Bio-PlexTM 200 system with Bio-Plex ManagerTM software version 6.1 (Bio-Rad Laboratories, Inc., CA, USA) was used to acquire and analyze the data.
Data analysis
Lipidomics and metabolomics data were processed and statistically evaluated in R software (
www.r-project.org, 2019, v 3.5.0) using the Metabol package [
23]. At first, the quality control-based locally estimated smoothing signal correction (LOESS) was applied to each dataset. The instrumental signal stability was monitored by internal standards across experimental and QC samples. The coefficient of variation (CV), based on QC samples, was calculated, and compounds (lipids or metabolites) with a CV higher than 30% were excluded from further processing. In the case of the brain, the correction of datasets to sample weight was carried out. The data were transformed by natural logarithm (ln), and the mean centering was applied. Bonferroni correction of p-values was applied to prevent false positivity of the t-test. Univariate and multivariate statistical methods were used to evaluate and visualize the results. Cytoscape software v3.8.2 was used to create metabolic and lipid maps according to the results of statistical analyses [
24]. For metabolomics data, an Enrichment pathway analysis was performed using the web-based platform MetaboAnalyst (v5.0) [
25] using our in-house database, which contained up to 53 metabolic pathways (Additional file Table
S1). The lipid ontology (LION) enrichment analysis web application was used for the enrichment analysis of lipidomic data [
26]. Data from western blot analysis, immunochemistry, ELISA, and FACS were statistically analysed using GraphPad Prism software v.8.0.1 (Inc., La Jolla, CA, USA).
Discussion
The pathological aggregation of tau protein is a hallmark of neurodegenerative diseases collectively referred to as tauopathies. Growing evidence suggests the significant involvement of deregulated lipid metabolism in the pathogenesis of neurodegenerative diseases, including tauopathies. While the relationship between lipid metabolism and tauopathy is not as well understood as that between lipid metabolism and Aβ pathology, recent studies have indicated a potential link between lipid metabolism and abnormal tau aggregation [
37]. In the present study, targeted lipidomic and metabolomic analysis was employed to reveal metabolic alterations induced by the cascade involving tau pathology.
Using a transgenic model with progressive age-dependent neurodegeneration and a well-established multi-component cell model system, we demonstrated that misfolding of tau is associated with marked metabolic changes. The presence of extracellular and aggregated tau was associated with increased production of lipids participating in protein fibrillization, membrane reorganization, and inflammation. Consequently, the membrane fluidity in glial cells was affected, leading to the accumulation of lipid particles in mitochondria, thus affecting the cell bioenergetics of brain cells.
Several independent studies characterized the neurolipidome of the AD human brain and identified changes in the abundance of a number of lipid species [
38‐
40]. In the current study, dysregulation of lipid metabolism was prominent in CSF and brain tissue (pons, medulla) of tau Tg rats compared to age-matched controls. Comprehensive lipidomics analysis established that LPC, LPE, PC, and PG were increased in brain tissue and CSF of Tg rats compared with controls. We were also interested in understanding the lipid dysregulation associated with disease progression and analyzed the transgenic and control animals in longitudinal samples (six-time points). Interestingly, significant changes have been found in the early steps of tau aggregation before the formation of sarkosyl-insoluble high-molecular-weight aggregates. The altered levels of several classes of GPL metabolism (LPC and LPE) were increased with time in Tg animals compared with controls. These observations suggest that lysophospholipid (lysoPL) species (LPC and LPE) are particularly associated with disease progression. GPLs normally are components of cellular or vesicle membranes. Membrane lipids act as essential signaling molecules and key modulators of signal transduction and vesicle trafficking [
41].
LysoPLs are intermediate metabolites in well-regulated phospholipid biosynthetic pathways produced by phospholipase-mediated hydrolysis [
42]. Aß aggregation has been shown to induce PLA2 up-regulation and activation [
43], and increased levels of lysoPLs have been found in brain ischemia [
44], traumatic brain injury [
45], amyloid transgenic mice [
46]. LPE and LPC bind to proteins such as receptors, enzymes (kinases, phosphatases), and act as neurotrophic factors that promote neuronal differentiation. The decrements observed in our Tg rats could indicate a potential role in dysfunctional synaptic transmission early in the disease process. The phospholipid degradation to lysoforms could also play an important role in the early stages of tau accumulation. We found significantly increased amounts of LPC and LPE in 6-months old animals before the development of high-molecular-weight tau aggregates. This suggests that LPC and LPE classes could be crucial for the early steps of tau aggregation and can promote further protein fibrillization. Previously, LPC has been shown to directly enhance the formation of Aß(1–42) fibrils [
47,
48]. Except for protein fibrillization, the accumulation of lysoPLs in brain tissue elicits neurotoxic effects, through the release of proinflammatory molecules [
49], astrocyte and microglial activation [
50], and alterations in membrane organization/fluidity [
51]. Our previous work showed that abnormal truncated tau induced inflammatory response in late stages of pathology [
52] and therefore the gradual increase of LPC and LPE in brain tissue of Tg rats could be result of neuroinflammation.
During the early stages of tauopathy, we showed significant upregulation of pathological hyperphosphorylated forms of tau protein and increased levels of total tau in the CSF of Tg rats. We, therefore, anticipate that lipid deregulation during the early stages of tauopathy could be caused by increased secretion of extracellular tau. This was accompanied by the activation of glial cells as indicated by the elevation of several cytokines in the brain tissue and increased reactivity of astrocytes. This finding aligns with the fact that neuroinflammation is a fundamental response of the CNS to chronic neurodegenerative processes [
53]. To better understand how extracellular tau modulates lipid metabolism, we used a multi-component in vitro cell model system. The model was developed by co-cultivation of human neuroblastoma cells overexpressing pathological truncated tau and primary rat glial cells (microglia, astrocytes). We showed that human neuroblastoma cells overexpressing truncated tau actively release monomeric total tau and proline-rich domain of tau into the media. Interestingly, levels of secreted proline-rich domain of tau significantly increased in time. A lipidomic study revealed increased production of lipids participating in protein aggregation and membrane reorganization as well as the formation of Soluble truncated tau protein released by neuron-like cells induced the formation of LDs in microglia. Our results showed that this process depended on the amount of secreted tau. Tau-mediated accumulation of LDs was also shown in the Tg rat model and on human brain tissue from patients with tauopathies. LDs were found adjacent or partially colocalized with mitochondria, and this was associated with decreased ATP activity. The formation of LDs could be caused by the accumulation of lipidic by-products of inflammatory response that cannot be processed due to dysfunctional mitochondria. Little is known about the dynamics of LD composition, which depends on the cell type and its current metabolic state [
54]. Previous reports showed that immune cells such as microglia accumulate large amounts of lipids in LDs [
55,
56]. LD-accumulating microglia are defective in phagocytosis and secrete high levels of proinflammatory cytokines. LDs have been previously proposed to serve as a lipid buffering system to prevent acylcarnitine accumulation and lipotoxicity in mitochondria [
57]. The lipid dyshomeostasis within microglia is likely to alter neuron-glia interplay and the metabolic integrity of neurons, leading to accelerated neurodegeneration.
Previous evidence has demonstrated that full-length or truncated tau is secreted by neurons and actively released into extracellular space. Although the physiological function of extracellular tau remains unclear, it is suggested its role in trans-cellular propagation. Extracellular tau can spread through the brain tissue and be internalized by healthy neurons affecting their function and activate glial cells during processes [
58,
59]. Tau secretion involves multiple pathways including vesicle-free direct translocation across the plasma membrane (PM). It was shown that during pathological and physiological conditions, the majority of tau is secreted in a vesicle-free form via translocation across PM. Tau filaments were found at the PM of AD brains. Membrane lipids such as PC and sphingolipids were associated with tau filaments isolated from AD patients. It was confirmed that MTBD of tau is involved in the binding of tau to lipid membrane and the N-terminal fragment of tau containing proline-rich domain promotes tau aggregation. Based on the data from transgenic animals and in vitro model systems, we showed that the early lipid changes before the development of neurofibrillary pathology could be induced by overexpression and increased secretion of extracellular tau. Our data showed that in the early stage of tauopathy, extracellular tau caused dysregulation of lipid homeostasis, which could accelerate further neurodegenerative and neuroinflammatory processes. The subsequent abrogation of lipid homeostasis could lead to tauopathy-like molecular phenotypes.
During the further development of neurofibrillary pathology, we uncovered the deregulation of SM metabolism. A significant increase in the levels of several SM, Cer, and hexosyl-ceramide (HexCer) species was detected. Increased levels of SM in CSF of Tg rats could be caused by membrane breakdown, demyelination, and progressive decline of neuronal cells in the brain tissue during the development of neurofibrillary pathology. SM metabolism is tightly associated with the metabolism of ceramides (Cer) [
60]. We observed elevated levels of two specific ceramides (Cer 40:2 (d18:1/22:1), Cer 42:2 (d18:1/24:1)) and three hexosylceramides (HexCer 40:2 (d18:1/22:1), HexCer 42:2 (d18:1/24:1), HexCer 42:3 (d18:2/24:1)) species. Several studies connect aberrant Cer production to Aß pathology (58,59). Aß oligomers cause activation of SMase enzyme, leading to Cer accumulation and subsequent cell death [
61]. In AD, levels of SM in CSF correlate with the levels of Aß and tau protein, and elevated levels of very-long-chain Cer species were detected in APOE ɛ4 carriers [
62].
With age and pathology progression several phosphatidylglycerol (PG) species have been gradually elevated in the brainstem of Tg rats. PG is known mainly for its function as a pulmonary surfactant, however in mammalian cells, PG is synthetized in mitochondria and involved in processes such as RNA transcription, and activation of protein kinase (PKC) [
63]. In a 5XFAD mouse model of AD, increased activity of PKC was found in astrocytes adjacent to Aß plaques [
64]. Interestingly, PKC requires phosphatidylserine (PS), localized on the cytoplasmic leaflet for its activation [
65]. Alongside PG accumulation we also observed elevated levels of PS species in brain tissue, however to a lesser extent. This is in agreement with studies on human AD brain tissues [
38]. PKC can promote tau protein phosphorylation [
66]. Therefore, pathological PKC activation by PG and PS lipids may be involved or further promote tau hyperphosphorylation and progressive formation of neurofibrillary tangles in Tg animals.
PCs and PEs, as the two most abundant membrane phospholipids across all types of cells, were found to be significantly increased in the brain tissue of 8-month-old rats. These lipid species could convoy inflammatory processes associated with widespread neurofibrillary pathology. Both elevated and decreased levels of PCs were found in the APP mouse model for late-onset AD [
67]. However, no significant alterations [
68], or even decreased levels [
69] in post-mortem AD brain tissue and human plasma of AD patients [
39].
The findings from the metabolic study showed that significant changes occur in the early stages of diseases before the formation of sarkosyl-insoluble high-molecular-weight tau aggregates. We found that tau accumulation leads to disruption in metabolic pathways associated with mitochondria, such as fatty acid oxidation (FAO), elongation, tricarboxylic acid cycle (TCA), creatine metabolism, purine catabolism, and arginine metabolism. These results are consistent with the observation that soluble tau plays a role in mitochondria distribution and oxidative stress, and long-term mitochondrial stress can trigger further tau dimerization [
70,
71].
We found changes in creatine and phosphocreatine in 4- and 6-month-old Tg rats compared to controls. Although we found no changes in brain creatine (Cr) nor phosphocreatine (PCr) in brain tissue, we observed their increased levels in the CSF. The Cr/PCr system involving brain-specific Cr kinase (CK) enzyme, functions as an energy buffer in the ATP-mediated cellular energetics [
72]. Disturbances in Cr/PCr balance are indicative of impaired energy metabolism in mitochondria and oxidative stress during the early stages of pathology in Tg rats. Brain-specific CK activity has been shown to decrease in AD brain [
31].
Our analysis also showed that the transition of tau protein from monomeric to high-molecular-weight oligomers allows the transition the brain metabolism from glycolysis to FA uptake and beta-oxidation. In conditions of low glucose availability, astrocytes can switch their metabolism to fatty acid oxidation (FAO). This metabolic flexibility allows astrocytes to adapt to energy demands and maintain their functions under changing circumstances [
73,
74]. In physiological state, brain energy metabolism is prevented from using FAs as a source of energy for their cytotoxicity and potential capacity to stimulate neurodegeneration [
75,
76]. We showed that expression of the neuron-specific glucose transporter GLUT3 was decreased in Tg rats, and this positively correlated with the amount of neurofibrillary pathology in brain tissue. In the case of impaired glucose uptake, of FAO in astrocytes serves as an alternative source of cellular energy. This may have been due to increased levels of ACs with different chain lenghts and modification. Taken together, these results suggest that after GLUT3 decrease, an increase in ACs contents in the brain tissue of Tg rats is observed as a result of the energy metabolism switch from glucose to FAO. Our results showed increased levels of hydroxylated and non-hydroxylated long-chain ACs in the brain tissue of Tg rats. Their amount increased from 6 to 12-month-old Tg animals compared to controls. The increase coincided with the appearance of tau aggregates in brain tissue. These findings could also be due to aggregation of amyloid-β in brain tissues of APP/PS1 transgenic mice models [
77]. On the other hand, the amount of short-chain ACs decreased in 4 and 6-month-old Tg rats. Interestingly, their levels increase from 8 to 10-month-old Tg animals. Levels of short-chain ACs were probably supported by FAO of long-chain ACs or possibly by amino acid catabolism at this age [
78]. Our result also showed that the brain levels of ACs returned to control levels, except for C14:1-OH, and C16:1-OH, which remained elevated up to 14 months of age. This suggests that the brain affected by tauopathy can, to some extent, stabilize its metabolic status.
Except for the mentioned ACs, elevated levels of free carnitine were found in brain tissue, CSF, and plasma. However, this upregulation was limited to 8-10-month-old tau Tg animals, which may be associated with higher demand for lipid catabolism. Free carnitine is reported to have neuroprotective effects [
79,
80] and provides the transport of FAs into mitochondria by the carnitine shuttle system.
Our study provides a broad overview of lipidomic and metabolomic changes associated with tau pathology in the brain tissue of SHR24 rats, alongside extracted mitochondria. This study is not exempt of limitations. One of the limitations is that the analysis conducted does not extend to the investigation of cell-specific effects. Recognizing this gap, we emphasize the necessity for future research employing advanced techniques for cell-specific insights. Secondly, the animal model used, SHR-24, predominantly exhibits pathology in the brainstem, with a minor presence of NFTs in the frontal cortex. Consequently, some findings may reflect region-specific effects. Thirdly, while modern omics technologies can target a large number of individual metabolites, they remain limited to specific subsets of the entire metabolome. Therefore, additional metabolites could be important in neurodegenerative diseases such as tauopathies.