Abstract
Tau P301S transgenic mice (PS19 line) are used as a model of frontotemporal lobar degeneration (FTLD)-tau. Behavioral alterations in these mice begin at approximately 4 months of age. We analyzed molecular changes related to disease progression in these mice. Hyper-phosphorylated 4Rtau increased in neurons from 1 month of age in entorhinal and piriform cortices to the neocortex and other regions. A small percentage of neurons developed an abnormal tau conformation, tau truncation, and ubiquitination only at 9/10 months of age. Astrocytosis, microgliosis, and increased inflammatory cytokine and immune mediator expression also occurred at this late stage; hippocampi were the most markedly affected. Altered mitochondrial function, increased reactive oxygen species production, and limited protein oxidative damage were observed in advanced disease. Tau oligomers were only present in P301S mice, they were found in somatosensory cortex and hippocampi at the age of 3 months, and they increased across time in the somatosensory cortex and were higher and sustained in hippocampi. Age-related modifications in lipid composition occurred in both P301S and wild-type mice with regional and phenotypic differences; however, changes of total lipids did not seem to have pathogenic implications. Apoptosis only occurred in restricted regions in late disease. The complex tau pathology, mitochondrial alterations, oxidative stress damage, glial reactions, neuroinflammation, and cell death in P301S mice likely parallel those in FTLD-tau. Thus, therapies should focus first on abnormal tau rather than secondary events that appear late in the course of FTLD-tau.
Introduction
Tauopathies are currently considered to be groups of neurodegenerative diseases presenting in adults with aggregates of abnormal tau in neurons and glial cells. Sporadic forms are represented by progressive supranuclear palsy, corficobasal degeneration, argyrophilic grain disease, Pick disease, and glial globular tauopathies, among others (1). Familial forms are commonly linked to mutations in MAPT, the tau gene located at chromosome 17; these can be properly designated as “tauopathies,” both in phenotypic and genetic terms ( 2 , 3 ). The characteristics of tau aggregates depend on the composition of tau, including tau mutations that variably affect tau splicing in familial tauopathies; predominance of 3Rtau or 4Rtau because of altered splicing for unknown reasons in sporadic tauopathies; grades or stages of tau phosphorylation, abnormal conformation, truncation and ubiquitination, and tau nitration in particular cells; and cytoskeletal characteristics of neurons, oligodendrocytes, and astrocyte subtypes (4). Familial tauopathies are usually manifested as frontotemporal lobar degeneration (FTLD-tau) leading to dementia and parkinsonism, and they are linked to a variety of MAPT mutations in exons 1, 9, 10, 11, 12, and 13 ( http://www.molgen.ua.ac.be/FTDMutations ).
A major limitation in the study of the pathogenesis of many human neurodegenerative diseases is the availability of cases only at terminal stages of the disease; accessible early stages of the corresponding disease are extremely rare. This does not occur in highly prevalent and common diseases such as Alzheimer disease (AD) and Parkinson disease, in which adequate brain tissue suitable for morphologic and biochemical studies is available in various brain banks; this has permitted important advances in the understanding of AD and Parkinson disease pathogenesis and progression. A similar approach is extremely difficult in sporadic tauopathies (except argyrophilic grain disease) and practically impossible in the majority of cases of FTLD-tau because of the limited numbers of affected individuals.
Several nontransgenic and transgenic animal models of tauopathies have been generated, mainly in mice but also in rats, Drosophila, zebrafish, and Caenorhabditis elegans, among other species (5). These models have been developed to gain understanding of the effects of tau not only on tau phosphorylation and aggregation and their effects on cytoskeletal stability and axonal transport but also to learn about putative alterations of metabolic pathways the defects of which underlie cellular malfunction and eventual neuronal death in tauopathies. Transgenic mouse lines bearing the MAPT P301S mutation have been generated; all of them mimic human FTLD (6–11). These mice have also been useful to demonstrate tau seeding and propagation of abnormal tau ( 12–14 ). Increased inflammatory responses, increased oxidative stress, abnormal mitochondrial function, and abnormal autophagy have been described in P301S transgenic mice ( 7 , 15–19). Unfortunately, most studies are focused on particular molecular aspects, and the reconstruction of a comprehensive scenario across time from very valuable but partial data is difficult. Therefore, studies are needed to gain understanding about the appearance of altered metabolic functions and their interactions and effects in the pathogenesis of FTLD-tau during disease progression.
Materials and Methods
Animals
The experiments were carried out in male heterozygous transgenic mice expressing human P301S tau (line PS19) and wild-type (WT) littermates aged 1 to 10 months. The generation of mice expressing the human mutated form P301S tau has been described (7). Transgenic mice were identified by genotyping from genomic DNA isolated from tail clips using the polymerase chain reaction (PCR) conditions proposed by Jackson Laboratory (Bar Harbor, ME). Animals were maintained under standard animal housing conditions in a 12-hour dark-light cycle with free access to food and water. All animal procedures were carried out following the guidelines of the European Communities Council Directive 2010/63/EU and with the approval of the local ethical committee (University of Barcelona).
Tau P301S mice and their corresponding WT littermates were evaluated at 1, 3, 6, 9, and 10 months of age. The number of animals studied varied among groups: WT 1 month (n = 10); WT 3 months (n = 10); WT 6 months (n = 8); WT 9 months (n = 24); WT 10 months (n = 9); P301S tau 1 month (n = 10); P301S tau 3 months (n = 8); P301S tau 6 months (n = 8); P301Stau9 months (n= 16); and P301S tau 10 months (n= 10).
Behavioral Evaluation
Emotional Evaluation: Elevated Plus Maze and Tail Suspension Tests
The Elevated Plus Maze assesses the anxiety levels of the animals and consists of placing the mice in a cross-shaped maze elevated 40 cm above the ground. Two arms of the maze are open laterally so that the animal can perceive the elevation of the maze. Animals presenting lower anxiety levels exhibited higher open arms exploration scores. The Tail Suspension Test suspends mice by the tail for 5 minutes at 30 cm from the ground. The immobility time is recorded, and it is considered a measure of despair behavior.
Cognitive Performance: Two-Object Recognition Test
This paradigm consists of placing the animals for 9 minutes in a V-maze containing 2 identical objects at the end of the arms. Twenty-four hours after the training session, the animals are again placed in the V-maze where 1 of the 2 familiar objects is replaced by a novel object. The time that the animals spend exploring the 2 objects is recorded in both the training and test sessions, and the object Recognition Index (RI) is calculated as the difference between the time spent exploring the novel (T N ) and the familiar object (T F ) during the test session, divided by the total time spent exploring the 2 objects [RI = (T N − T F )/(T N + T F )]. Animals exhibiting memory impairments revealed a lower object RI.
Nociception: Tail Immersion Test
The spontaneous nociceptive threshold was determined using water maintained at 48°C ± 0.5°C as the painful stimulus. The mice were immobilized in a cylinder, and their tails were immersed in heated water. The latency to a rapid flick of the tail was taken as the end point, and a cutoff was set up at 15 seconds to prevent tissue damage.
Motor Coordination: Rotarod Test
The animals were placed on the roller lane of a Rotarod increasing rotation speed, and the time latency to fall was recorded.
Sample Collection
At the end of the behavioral testing, the animals were killed by cervical dislocation and their brains and spinal cords were then rapidly removed and processed for study. The left cerebral hemisphere, the brainstem, cerebellum, and part of the spinal cord were dissected on ice, immediately frozen, and stored at −80°C until use for biochemical studies. The right hemisphere, brainstem, cerebellum, and spinal cord were fixed in 4% paraformaldehyde and processed for immunohisto-chemistry. Dissected frozen sections included cerebral cortex, hippocampus, brainstem, cerebellum, and lumbar spinal cord. In addition, the spinal ganglia and samples of the intestine were fixed in paraffin to study the myenteric plexuses.
Morphologic Studies, Immunohistochemistry, Double Labeling Immunofluorescence, and Confocal Microscopy
Tissue samples were embedded in paraffin, and 4-µm-thick coronal sections were obtained with a sliding microtome. Dewaxed sections were stained with hematoxylin and eosin and Nissl stain and processed for immunohistochemistry; the latter treated with citrate buffer (20 minutes) to enhance antigenicity. Endogenous peroxidases were blocked by incubation in 10% methanol-1% H 2 O 2 solution for 15 minutes, followed by 3% normal horse serum solution, and then incubated at 4°C overnight with one of the primary antibodies indicated, that is, AT8, phospho-specific tauThrl 81, phosphor-specific tauSer422, 3Rau, 4Rtau, AT8, PHF, Alz50, tau-C3, ubiquitin, glial fibrillary acidic protein (GFAP), and IBA1 ( Table 1 ). Sections were subsequently rinsed and incubated with biotinylated secondary antibody (Dako, Glostrup, Denmark), followed by EnVision+ system peroxidase (Dako) and, finally, with the chromogen di-aminobenzidine and H 2 O 2 . Some sections were incubated without the primary antibodies; no immunostaining was detected in these sections. Sections were lightly counterstained with hematoxylin. After staining, the sections were dehydrated and coverslipped for microscopic assessment.
Antibodies Used for Immunohistochemistry, Immunofluorescence, and Western Blotting
For Nissl staining, sections were stained with 10% cresyl violet at 56°C for 1 hour, dehydrated, and coverslipped for observation under a Nikon Eclipse E800 microscope (Nikon Imaging Inc., Tokyo, Japan). The neuronal density in CA1 and dentate gyrus layers was calculated in 3 representative pictures taken from the hippocampus of each animal using the Adobe Photoshop CS4 software. Five animals per group were used for quantifications.
For double labeling immunofluorescence, dewaxed sections were treated with citrate buffer for 20 minutes to enhance antigenicity and then stained with a saturated solution of Sudan black B for 30 minutes (Merck Millipore, Darmstadt, Germany) to block lipofuscin autofluorescence. Immediately afterward, the sections were rinsed in 70% ethanol and washed in distilled water. After blocking endogenous peroxidases with 10% fetal bovine serum for 90 minutes, 1 series of sections of the lumbar spinal cord of mice aged 9 months was incubated at 4°C overnight with antibodies tauThr181 and Alz50 or with tauThr181 and tau C3, and then incubated with Alexa488 or Alexa546 fluorescence secondary antibodies against the corresponding host species (1:400; Molecular Probes, Eugene, OR). Nuclei were stained with DRAQ5 (1:2000; Biostatus, Leicestershire, UK). After washing, the sections were mounted in Immuno-Fluore mounting medium (ICN Biomedicals, Irvine, CA), sealed, and dried overnight. Sections were examined with a Leica TCS-SL confocal microscope (Leica, Wetzlar, Germany). Another series of sections of the cerebral cortex corresponding to mice aged 3, 6, and 9 months was incubated with AT8 and anti-GFAP or anti-IBA1 to look at the relationship between tau-positive neurons and glial cells. After washing, the sections were incubated with Alexa488 or Alexa546 fluorescence secondary antibodies against the corresponding host species. The sections were washed and mounted in Immuno-Fluore Mounting medium, sealed, dried overnight, and examined with a Nikon Eclipse E800 microscope (Tokyo, Japan). ApopTag Plus Peroxidase in situ Apoptosis Detection Kit (S 7101; Millipore, Temecula, CA) was used following the protocol of the supplier to detect cells with DNA fragmentation and apoptotic morphology.
RNA Purification
The purification of RNA from mouse somatosensory cortex, hippocampus, and spinal cord (n = 5−8 samples per group) was carried out with RNeasy Lipid Tissue Mini Kit (Qiagen, Hilden, Germany) following the protocol provided by the manufacturer. Quality of isolated RNA was first measured with Bioanalyzer Assay (Agilent, Santa Clara, CA). The concentration of each sample was obtained from A260 measurements with Nanodrop 1000 (Thermo Scientific, Wilmington, DE). RNA integrity was tested using the Agilent 2100 Bio Analyzer (Agilent Technologies, Palo Alto, CA).
Retrotranscription Reaction
The retrotranscriptase reaction was carried out using a High-Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA) following the protocol provided by the supplier. Parallel reactions for an RNA sample were run in the absence of MultiScribe Reverse Transcriptase to assess the degree of contaminating genomic DNA.
TaqMan PCR
TaqMan quantitative RT-PCR assays for each gene were performed in duplicate on cDNA samples in 384-well optical plates using an ABI Prism 7900 Sequence Detection system (Applied Biosystems, Life Technologies, Waltham, MA). For each 10-µL TaqMan reaction, 4.5 µL cDNA was mixed with 0.5 µL 20× TaqMan Gene Expression Assays and 5 µL of 2× TaqMan Universal PCR Master Mix (Applied Biosystems). Parallel assays for each sample were carried out using probes for alanyl-tRNA synthase (Aars) and X-prolyl aminopeptidase (aminopeptidase P) 1 (XPNPEP1). The reactions were carried out using the following parameters: 50°C for 2 minutes, 95°C for 10 minutes, and 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Finally, all TaqMan PCR data were captured using the Sequence Detection Software (SDS version 1.9; Applied Biosystems). The identification numbers and names of all TaqMan probes used are shown in Table 2 . Probes included members of the complement system, colony-stimulating factor receptors, toll-like receptors, proinflammatory cytokines, interleukin-6, members of the tumor necrosis factor family, interleukin-10 and receptors, and transforming growth factor-β family. The selection of these particular probes was caused by the fact that similar studies are being carried out in different models of neurodegenerative diseases; the use of the same probes by the same hands permits a comparison of inflammatory responses in the different models. Samples were analyzed with the double-delta cycle threshold (ΔΔCT) method. The ΔCT values represent normalized target gene levels with respect to the internal control. The ΔΔCT values were calculated as the ΔCT of each test sample minus the mean ΔCT of the calibrator samples for each target gene. The fold change was determined using the equation 2 (−ΔΔCT) .
TaqMan Probes Used for Study of the Expression of Cytokines and Mediators of the Immune Response
Western Blotting
Frozen brain tissue samples were lysed in lysis buffer: 100 mmol/L Tris pH 7, 100 mmol/LNaCl, 10 mmol/L EDTA, 0.5% NP-40, and 0.5% sodium deoxycholate plus protease and phosphatase inhibitors (Roche Molecular Systems, Pleasanton, CA). After centrifugation at 14,000g at 4°C for 20 minutes (UI-tracentrifuge Beckman with 70Ti rotor; Beckman, Fullerton, CA), supernatants were quantified for protein concentration (BCA; Pierce, Life Technologies, Darmstadt, Germany), mixed with SDS-PAGE sample buffer, boiled, and subjected to 8% to 15% SDS-PAGE. Gels were electrophoretically transferred onto nitrocellulose membranes (200 mA per membrane, 20 minutes). Nonspecific bindings were blocked by incubation in 5% albumin in Tris-buffered saline (TBS) containing 0.2% Tween at room temperature for 1 hour. After washing, the membranes were incubated at 4°C overnight in TBS containing 5% albumin and 0.2% Tween with one of the specific primary antibodies. Samples of the brainstem at the ages of 1, 3, and 6 months (n = 6 for every age) were incubated with rabbit polyclonal antibodies anti-tau-5, anti-3Rtau, anti-4Rtau, and anti-tauThr181 in TBS containing 5% albumin and 0.2% Tween. Frozen samples of the somatosensory cortex and hippocampus of P301S tau mice and corresponding WT littermates aged 3 (n = 3−5), 6 (n = 5), and 9/10 (n = 5) months were incubated with antibodies anti-IBA1, anti-neuroketal, and anti-superoxide dismutase 2 (SOD-2); anti-β-actin was blotted for the control of protein loading. The characteristics of the antibodies are summarized in Table 1 . Membranes were then incubated for 1 hour with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody (1:1000; Dako), and the immune complexes were revealed with a chemiluminescence reagent (ECL; Amersham, GE Healthcare, Buckinghamshire, UK).
Densitometries were carried out with TotalLab software (Bionova, Madrid, Spain), and values were normalized using β-actin. Normalized values were expressed as the fold change from values obtained in control samples.
Tau Oligomers
Frozen samples of the somatosensory cortex and the hippocampus of P301S tau and corresponding WT littermates aged 3 months (n = 4–5), 6 months (n = 5), and 9/10 months (n = 5) were homogenized in lysis buffer: 100 mmol/L Tris (pH 7.0), 100 mmol/L NaCl, 10 mmol/L EDTA, 0.5% NP-40, and 0.5% sodium deoxycholate plus protease and phosphatase inhibitors (Roche Molecular Systems). After centrifugation at 14,000g for 20 minutes at 4°C (Ultracentrifuge Beckman with 70Ti rotor), supernatants were quantified for protein concentration (BCA; Thermo Scientific), mixed with SDS-PAGE sample buffer, boiled, and separated to 8% SDS-PAGE gels. The proteins were transferred to nitrocellulose membranes (200 mA per membrane for 90 minutes). The membranes were blocked with 5% nonfat milk in TBS containing 0.2% Tween for 1 hour at room temperature. After washing, the membranes were incubated at 4°C overnight with the primary antibody anti-tau-5 (1:1000; Thermo-Fisher) in TBS containing 5% albumin and 0.2% Tween. Membranes were then incubated for 1 hour with the HRP-conjugated secondary anti-mouse antibody (1:2000; Dako), and the immune complexes were visualized with a chemiluminescence reagent (ECL; Amersham, GE). Control of protein loading was checked by estimation of β-actin expression levels in the same membranes.
Dot Blot Assay
Frozen samples of the hippocampus and cortex of P301S tau and corresponding WT littermates aged 3 months (n = 3–5), 6 months (n = 5), and 9/10 months (n = 5) were homogenized in TBS buffer with a cocktail of protease inhibitors (Roche Molecular Systems) and ultracentrifuged at 100,000g for 1 hour at 4°C. Five micrograms of each soluble sample was applied to nitrocellulose membranes using a 48-well Dot Blot Manifold (Cleaver Scientific, Rugby, UK). The membranes were blocked with 10% nonfat milk in TBS at room temperature for 1 hour. The membranes were washed 3 times for 5 minutes each with TBS-T and incubated for 2 hours at room temperature with ABN454/anti-Tau (T-22) oligomeric antibody (Merck, Millipore) used at different dilutions following the instructions of the supplier. After washing, the membranes were incubated for 1 hour with the appropriate HRP-conjugated secondary antibody (1:2000; Dako), and the immune complexes were visualized with a chemiluminescence reagent (ECL; Amersham).
Mitochondria Isolation
Brains were kept in chilled Celsior preservation solution (Osmolality 320 mOsm/L with 100 mmol/L sodium, 15 mmol/ L potassium, 13 mmol/L magnesium, 0.25 mmol/L calcium, 20 mmol/L glutamic acid, 30 mmol/L histidine, 60 mmol/L mannitol, 80 mmol/L lactibionate, and 3 mmol/L glutathione) until dissection of the neocortex for mitochondria isolation. The protocol of mitochondria isolation is based on different centrifugation strengths that sequentially separate a mitochondria-rich fraction from larger membrane organelles. The isolation was carried out on ice. Brain tissue was homogenized with a dounce-homogenizer in 10 vol/wt (µL/mg) of isolation buffer composed of (in mmol/L) 250 sucrose, 0.5 Na 2 EDTA, and 10 HEPES pH 7.4 at 0°C. The homogenate was centrifuged at 600g for 109 minutes. The supernatant was centrifuged at 10,000g for 10 minutes, and the pellet was washed with
1 mL isolation buffer and centrifuged at 1000g for 10 minutes. This second supernatant was centrifuged for 10 minutes at 10,000g, yielding the final pellet containing isolated mitochondria (IM). The pellet was resuspended in respiration media (Miro5; see below). Mitochondria protein con centration was determined with the Bradford reagent using bovine serum albumin as standard.
High-Resolution Respirometry With Reactive Oxygen Species Detection
High-resolution oxygraph (Oroboros 02K; Oroboros Instruments, Innsbruck, Austria) was used to measure oxygen consumption. Each oxygraph has two 2-mL chambers with polarographic (Clark-type) oxygen sensors (20). The respiration medium Miro5 used was composed of 0.5 mmol/L EGTA, 3 MgCl 2 -6H 2 O, 60 K-lactobionate, 20 mmol/L taurine, 10 mmol/L KH 2 PO 4 , 20 mmol/L HEPES, 110 mmol/L sucrose, and 1 g/L bovine serum albumin. Before starting each experiment, the medium was allowed to equilibrate with atmospheric oxygen. The assays were run at 37°C. At that time, O 2 concentration was calibrated with the software calculations for O 2 saturation considering 100 kPa barometric pressure and the solubility factor of the respiration medium as 0.92. This value was 190 umol/L O 2 at 37°C.
The oxygraph Oroboros O2K has additional amphoteric and volumetric ports that were used to couple purpose-built fluorimetric systems with light emission of 530 nm and detection sensors of 590 nm, respectively ( 21 , 22 ). DatLab5 software (Oroboros Instruments) was used to calculate the real-time fluorophore signal.
After the protocols described ( 21 , 22 ), net reactive oxygen species (ROS) production was measured simultaneously with respirational flux by means of the detection of the fluorescence emitted by Amplex Ultrared (AMPR; Life Technologies). Before the mitochondria, the following solution was added to the chamber (in final concentration): 25 U/mL SOD (Sigma, Madrid, Spain), 2.5 U/mL HRP (Sigma), and 25 µmol/L AMPR. The superoxide radical (O 2 ) released from mitochondria is reduced to H 2 O 2 by both endogenous and exogenous SOD. The H 2 O 2 is coupled to HRP, which, in turn, reacts with AMPR to form Amplex Ultrox Red, a fluorescent product with excitation and emission wavelengths of 530 and 590 nm, respectively. For calibration of the ROS signal, H 2 O 2 (330 nmol/L) was added before the addition of the mitochondria sample and the initiation of measurements.
During the experiments and before any addition to the chambers, the oxygen flux and the fluorimetric signals were allowed to stabilize. The assay was started with the addition of mitochondria to the chambers after which the substrates and inhibitors were added based on prior studies ( 21 , 22 ) in the following order: 1) C-I substrates: glutamate (10 mrnol/L), malate (5 mrnol/L), and pyruvate (10 mmol/L) (Leak CT); 2) CTI substrate succinate (10 mmol/L) (Leak C-I,II); 3) adenosine diphosphate (1.25 mmol/L) (OXP); 4) OXP inhibitor oligomycin (7.5 µmol/L) (Leak Oli); 5) the uncoupler carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) (1.5 µmol/L) (ETS); and 6) antimycin A (5 µmol/L) (Leak AmA).
DatLab5 software (Oroboros Instruments) was used to calculate the real-time oxygen fluxes, which were represented as O 2 consumed per second and milligram of mitochondria added. Values of released H 2 O 2 per second were expressed with respect to the uptake of oxygen atoms and per milligram of mitochondria in the chamber.
Mass Spectrometry Analysis
Nonenzymatic Protein Modifications
Markers of protein oxidation (the protein carbonyl glutamic semialdehyde [GSA]), glycoxidation (Ne-[carboxyethyl]-lysine, [CEL]), lipoxidation (Ne-[carboxymethyl]-lysine [CML] and Ne-malondialdehyde-lysine [MDAL]), and succination (S-[2-succinyl]-cysteine [2-SC]) were determined as trifluoroacetic acid methyl ester derivatives in acid hydrolyzed delipidated and reduced brain protein samples by gas chromatography (GC)/mass spectrometry ( 23 ) using an HP6890 Series II gas chromatograph (Agilent, Barcelona, Spain) with an MSD5973A Series detector and a 7683 Series automatic injector, a HP-5MS column (30-m × 0.25-mm × 0.25-µm), and the described temperature program (23). Quantification was performed by internal and external standardization using standard curves constructed from mixtures of deuterated and nondeuterated standards. Analyses were carried out by selected ion-monitoring GC/mass spectrometry. The ions used were as follows: lysine and [ 2 H g ]lysine, m/z 180 and 187, respectively; 5-hydroxy-2-aminovaleric acid and [ 2 H 5 ]5-hydroxy-2-aminovaleric acid (stable derivatives of GSA), m/z 294 and 298, respectively; CML and [ 2 H 4 ]CML, m/z 392 and 396, respectively; CEL and [ 2 H 4 ]CEL, m/z 379 and 383, MDAL and [ 2 H 8 ]MDAL, m/z 474 and 482, respectively; and 2-SC and [ 2 H 2 ]SC, m/z 284 and 286, respectively. The amounts of product were expressed as µmol/L of GSA, CML, CEL, MDAL, or 2-SC per mole of lysine.
Fatty Acid Profile and Global Fatty Acid Indexes
Fatty acids from brain lipids were analyzed as methyl ester derivatives with GC, as previously described (23). Separation was performed with a DBWAX capillary column (30 m × 0.25 mm × 0.20 µm) in a GC System 7890A with a Series Injector 7683B and an FID detector (Agilent Technologies). Identification of fatty acid methyl esters was made by comparison with authentic standards (Larodan Fine Chemicals, Malmö, Spain). Results are expressed as mol%.
The following fatty acyl indices were also calculated: saturated fatty acids; unsaturated fatty acids (UFAs); mo-nounsaturated fatty acids (MUFAs); polyunsaturated fatty acids (PUFAs) from n-3 and n-6 series (PUFAn-3 and PUFAn-6); and average chain length (ACL) = [(Σ%Totall4 × 14) + (Σ% Totall6 × 16) + (Σ%Totall8 × 18) + (Σ%Total20− 20) + (Σ% Total22 × 22) + (Σ% Total24 × 24)]/100. The density of double bonds in the membrane was calculated by the Double Bond Index (DBI) = [(1 × Σmol% monoenoic) + (2 × Σmol % dienoic) + (3 × Σmol% trienoic) + (4 × Σmol% tetraenoic) + (5 × Σmol% pentaenoic) + (6 × Σmol% hexaenoic)]. The membrane susceptibility to peroxidation was calculated by the Peroxidizability Index (PI) = [(0.025 × Σmol% monoenoic) + (1 × Σmol% dienoic) + (2 × Σmol% trienoic) + (4 × Σmol% tetraenoic) + (6 × Σmol% pentaenoic) + (8 × Σmol% hexaenoic)]. Finally, anti-inflammatory index (AI) was = [(20:3n-6) + (20:5n-3) + (22:6n-3)]/(20:4n-6).
Elongase (ElovlX) and desaturase (DxD) activities were estimated from specific product/substrate ratios (24): D9D (n-7) = 16:ln-9/16:0; D9D(n-9) = 18:ln-9/18:0; D8D(n-6) = 20:3n-6/20:2n-6; D5D(n-6) = 20:4n-6/20:3n-6; D6D(n-3) = 24:6n-3/24:5n-3;Elovl3(n-9)=20:ln-9/18:ln-9;Elovl6=18:0/ 16:0; Elovll-3-7 = 20:0/18:0; Elovll-3-7 = 22:0/20:0; Elovl 5 (n-6)a = 20:2n-6/18:2n-6; Elovl5(n-6)b = 20:3n-6/18:3n-6; EI0VI2-5 (n-6) = 22:4n-6/20:4n-6; Elovl 2-5(n-3) = 22:5n-3/ 20:5n-3, and Elovl 2(n-3) = 24:5n-3/22:5n-3. Finally, the n-6 pathway, the n-3 pathway, and peroxisomal P-oxidation were estimated according to the ratios 22:5n-6/18:2n-6, 22:6n-3/ 18:3n-3, and 22:6n-3/24:6n-3, respectively.
Statistical Analysis
Data were analyzed with 2-way analysis of variance with genotype and age as between factors, followed by Tukey post hoc or Student t -test when required. The survival curves were compared with the Gehan-Breslow-Wilcoxon test. In all the experiments, the significance level was set at p < 0.05.
Results
General Health of P301S Mice
Tau P301S mice exhibited lower body weight from 3 months of age compared with WT animals, which was more evident from the age of 7 months onward ( Fig. 1A ). The analysis of the survival curves revealed significant increased mortality of tau transgenic mice #### compared with WT littermates ( Fig. 1B ). Most of the Tau P301S mice deaths occurred between 8 and 10 months of age after paralysis of the hind limbs and rapid general feebleness, which impeded regular drinking and feeding. Paralysis of the hind limbs was considered the end point criterion for killing the animals for ethical reasons.
(A) Body weights of tau P301S mice (TAU, black columns) are significantly reduced from 3 months of age versus wild-type (WT) littermates (white columns). (B) Survival curves reveal increased mortality of P301S mice (Tau) versus WT littermates. (C) P301S mice from 6 to 10 months of age exhibit reduced despair behavior in the Tail Suspension Test as revealed by the reduction in the immobility time with respect to WT littermates. (D) P301S mice from 9 months significantly explore the open arms for longer than controls in the Elevated Plus Maze, revealing reduced anxiety levels. (E) Transgenic mice at 9 and 10 months explore the objects for longer times versus WT littermates in the 2-Object Recognition Test. (F) P301S mice show memory impairment at 10 months of age in the 2-Object Recognition Test. (G) The latency to flick the tail in response to a thermal stimulus is reduced in P301S mice aged 9 and 10 months in the tail immersion test, indicating a decrease in the nociceptive threshold. (H) Surviving P301S mice at the age of 9 and 10 months show better motor coordination skills than WT littermates, as revealed by the increased latency to fall from the Rotarod. Data are expressed as the mean values ± SE. * p < 0.05, ** p < 0.01, *** p < 0.001 compared with WT mice, g, grams; sec, seconds.
P301S Mice Behavior, Memory, Nociceptive Threshold, and Locomotion
Transgenic mice exposed to the Tail Suspension Test exhibited a significant reduction in immobility time from 6 months onward (p < 0.05), which is considered as reduced despair behavior and indicative of disinhibition ( Fig. 1C ). P301S mice aged 9 and 10 months exhibited reduced anxiety levels on the Elevated Plus Maze as revealed by the significant increase in the exploration time of the open arms when compared with wild-type animals (9 months, p < 0.05; 10 months, p < 0.001) ( Fig. 1D ). Moreover, P301S mice from 9 months of age exhibited increased total exploration time during the training session in the 2-Obj ect Recognition Test ( Fig. 1E ; 9 months, p < 0.001 ; 10 months, p < 0.05), supporting disinhibited behavior in front of novel or stressful events (tail suspension or elevated arms). Transgenic mice aged 10 months showed memory impairment as revealed by a significant reduction in the recognition index in the 2-Object Recognition Test compared with corresponding WT littermates ( Fig. 1F ; p < 0.01). The latency to flick the tail in response to a thermal stimulus was significantly reduced in P301S mice aged 9 and 10 months (p < 0.05), thus revealing a reduction in nociceptive thresholds ( Fig. 1G ). Curiously, surviving P301S mice at 9 and 10 months had better motor coordination skills than WT littermates, as revealed by the significant increase in the latency to fall on the Rotarod Test ( Fig. 1H ; p < 0.05 in both cases). As mentioned before, the final stage of P301S mice was characterized by severe motor impairment of the hind limbs followed by rapidly progressing generalized weakness during 3 or 4 days. Details of the statistical analysis are found in Table, Supplemental Digital Content 1 ( Supplementary Data ).
Tau Pathology in P301S Transgenic Mice
Immunohistochemistry revealed hyperphosphorylated tau deposition, as depicted with the AT8 antibody, in neurons of the entorhinal and piriform cortex and more sparsely in neurons of the upper layers of the somatosensory cortex in transgenic mice aged 1 month. In P301S animals aged 3 months, hyperphosphorylated tau increased in these regions and extended to neurons of the deep layers of the somatosensory cortex and cingulate cortex, hypothalamus, septal nuclei, granule cell layer of the cerebellum and brainstem, mainly pontine nuclei, tegmentum, and raphe nuclei ( Fig. 2A ).
Representative images of hyperphosphorylated tau deposition in P301S transgenic mice at the ages of 1, 3, 6, and 9 months. (A) Transgenic mice aged 1 month (a, b) and 3 months (c-i) ; (a, c) entorhinal cortex; (b, d) upper layers of the somatosensory cortex; (e) cingulate cortex; (f) hypothalamus; (g) septal nuclei; (h) granule cell layer of the cerebellum; (i) pontine nuclei. (B) Transgenic mice aged 6 months; (a) upper layers of the somatosensory cortex; (b) cingulate cortex; (c) basal forebrain; (d) hypothalamus; (e) septum; (f ) cerebellum; (g) CA2 region of the hippocampus; (h) pontine nuclei; (i) ventral spinal cord at the cervical level. (C) Transgenic mice aged 9 months; (a) thalamus; (b) CA1 region of the hippocampus; (c) dentate gyrus; (d) dorsal spinal cord; (e-g) ventral spinal cord; (h, i) dorsal ganglia. Paraffin sections, slightly counterstained with hematoxylin. (A, B,Ca-d, i) AT8 antibody; (Ce) 4Rtau; (Cf ) tau-C3; (Ci) Alz50. Scale bar = 25 µm.
The numbers of neurons with hyperphosphorylated tau deposition increased in transgenic mice aged 6 months in all these regions and extended to the amygdala, nuclei of the basal forebrain, striatum, CA2 region of the hippocampus, other nuclei of the brainstem, and the ventral spinal cord ( Fig. 2B ).
P301S transgenic mice aged 9/10 months had widespread hyperphosphorylated tau accumulation in the entorhinal and piriform cortices, upper and inner layers of the somatosensory cortex, cingulate cortex, amygdala, nuclei of the basal forebrain, striatum, hypothalamus, thalamus, dentate gyrus, CA1 region of the hippocampus, granular and molecular layer of the cerebellum, brainstem, ventral and dorsal spinal cord, and spinal ganglia ( Fig. 2C ). Neurons of the myenteric plexuses also contained hyperphosphorylated tau (data not shown).
Hyperphosphorylated tau deposition was restricted to neurons; astrocytes and oligodendrocytes were spared. Tau deposition was massively 4Rtau ( Fig. 2Ce ).
In addition to the regional spread of hyperphosphorylated tau deposition, the characteristics of abnormal tau were also modified with disease progression. Tau hyperphosphorylation, as revealed with the phospho-specific antibodies ant-tauThr181, anti-tauSer422, and AT8, was observed at the first stages of tau deposition. However, conformational modifications of tau as revealed with the antibody Alz50, which recognizes modifications at amino acids 5-15, and truncated tau at aspartic acid 421, as identified by the antibody tau-C3, were only observed in a restricted number of neurons in transgenic mice aged 9/10 months ( Fig. 2Cf , Cg, Cl). Double labeling immunofluorescence and confocal microscopy carried out in the ventral spinal cord of transgenic mice aged 9/10 months revealed that only approximately 30% of tauThrl 81 -immunoreactive neurons were stained with Alz50 antibodies, and between 20% and 30% of tauThrl81 positive neurons were stained with tau-C3 (Figure, Supplemental Digital Content 2, Supplementary Data ). Neurons with hyperphosphorylated tau and abnormal tau conformation and truncated tau were almost absent in mice aged 3 and 6 months in all regions.
Western blots of total homogenates of the brainstem revealed the same high levels of 3Rtau in WT and transgenic mice at the age of 1 month, decreasing with age in a similar way in both groups. However, higher expression levels of Tau-5, 4Rtau, and phosphorylated tau at Thrl 81 in P301S were found in transgenic mice when compared with WT as early as at the age of 1 month, thus indicating that Western blotting was more sensitive than immunohistochemistry to abnormal tau expression levels in structures in which hyperphosphorylated tau was not noted until the age of 3 months. Densitometry study of the 4Rtau bands indicated about a 6-fold increase in 4Rtau in P301S animals in comparison with WT littermates. Therefore, increased tau expression was caused by mutant tau in transgenic mice, as anticipated in this model ( Fig. 3 ).
Western blotting of total homogenates of the brainstem of wild-type (WT) and P301S transgenic mice at the ages of 1, 3, and 6 months (n = 6 for every group) showing similar high expression levels of 3Rtau in WT and transgenic animals aged 1 month, decreasing thereafter in a similar way in the 2 groups. In contrast, Tau-5 and 4Rtauare markedly increased in P301S mice at all ages. Note the 2 bands of Tau-5 and 4Rtau in transgenic mice; the lower corresponds to the baseline levels of tau and the upper and thicker band to the mutant tau; only the lower band is present in WT animals. Hyperphosphorylated tau, as revealed with the antibody P-tauThr181, is dramatically increased as early as 1 month of age in the P301S mice. Beta-actin was used as a control of protein loading.
Ubiquitin immunohistochemistry revealed a few neurons containing ubiquitinated granules and tangle-like inclusions in the brainstem and granular neurons in the dentate gyrus and CA3 and CA1 regions of the hippocampus; fibers in the hilus of the dentate gyrus were also stained with anti-ubiquitin antibodies ( Fig. 4 A–D).
(A–H) Ubiquitin immunohistochemistry (A–D) and in ApopTag Plus peroxidase in situ detection of apoptosis (E–H) in P301S transgenic mice aged 3 months (A) and 9 months (B–H) . Ubiquitin-immunoreactive dots in the hilus of the hippocampus and in scattered neurons with tangles in the dentate gyrus in P301S aged 9 months (B) contrast with the lack of similar immunostaining in mice aged 3 months (A) . Ubiquitin-immunoreactive tangles are also in the brainstem in P301S mice aged 9 months (C, D) . Apoptotic cells (extremely condensed nuclei), which are stained brown with ApopTag are found in the CA1 region (E) , dentate gyrus (F, C) , and brainstem (H) . Paraffin sections lightly stained with hematoxylin; (A–D) scale bar in (D) = 25 µm; (E–H) scale bar in (H) = 10 µm.
Neuronal Viability in Hippocampus of P301S Transgenic Mice
Nissl staining revealed no decrease in the neuronal density in the CA1 region of the hippocampus and in dentate gyrus in P301S transgenic mice compared with WT littermates at 1,3,6, and 9/10 months (data not shown). However, in situ detection of apoptosis showed a few intensely stained condensed nuclei in the dentate gyrus, CA3 and CA1 regions of the hippocampus, and scattered positive nuclei in the brainstem ( Fig. 4 E–H). Apoptosis was rare in the somatosensory cortex, amygdala, diencephalic nuclei, and spinal cord even at the age of 9/10 months.
Astrocyte and Microglia Reactions in the Brain and Spinal Cord
P301S mice exhibited an age-dependent increase in the number of astrocytes between the ages of 6 and 9/10 months mainly in the piriform cortex ( Fig. 5 Aa–Af), hippocampus ( Fig. 5 Ba–Bf), brainstem, and spinal cord when compared with those in WT animals. In addition, the morphology of astrocytes changed in P301S mice with age, showing astrocytes with enlarged bodies, robust branches, and increased GFAP expression in mice aged 9/10 months ( Fig. 5A , Bc, Bf).
Astroglial and microglial reactivity in the piriform cortex (A) and hippocampus (B) of wild-type (WT) (a, d, g, j) and P301S (Tau) mice (b, c, e, f, h, i, k, I) aged 1 and 9 months. Representative images of glial fibrillary acidic protein (GFAP) immunohistochemistry showing an increase in the number of astrocytes and density of GFAP protein in P301S mice. Higher-magnification images reveal a shift in the astrocyte morphology toward a reactive phenotype in P301S mice at the age of 9 months (A, B, f ) when compared with 1 -month-old transgenic mice (A, B, c) . Similarly, the density of microglial IBA1 -positive cells increases in P301S mice aged 9 months. The morphology of microglial cells has a reactive phenotype in transgenic mice aged 9 months (A, B, i) compared with transgenic mice aged 1 month (A, B , I) . Paraffin sections with slight hematoxylin counterstaining. Scale bars = (A) 200 µm; (c, f, i, l) 50 µm; (B) 500 µm; (c, f, i, I) 100 µm. (C) Double labeling immunofluorescence showing no direct relationship between neurons with hyperphosphorylated tau deposition (AT8-positive cells, green, white arrows) and astrocytes ( a , GFAP-positive cells, red) and microglia ( b , IBA1-positive cells, red). Scale bar = 50 µm. (D) Quantification of astrocytes (GFAP-positive cells) and microglia (IBA1 -positive cells) in selected regions of the brain and spinal cord in P301S transgenic mice (TAU) and WT littermates at 1, 3, 6, and 9/10 months of age. Values represent the number of cells per mm 2 as follows: 1 : <30; 2: 30–70; 3: 70–120; 4:120–150; 5:150–200; 6: >200. Major differences in the numbers of astrocytes and microglia are seen between the ages of 6 and 9/10 months in the piriform cortex, hippocampus, brainstem, and spinal cord. The increase in microglia is particularly important in the hippocampus and spinal cord. In contrast, the somatosensory cortex, amygdala, striatum, thalamus, and hypothalamus do not show significant modifications in astrocyte and microglia cell numbers. (E) Western blotting to IBA1 in the somatosensory cortex of WT and P301S transgenic (TAU) mice showing no significant increments in IBA1 expression levels in animals aged 3, 6, and 9/10 months. (F) Western blotting to IBA1 in the hippocampi of WT and P301S transgenic (TAU) mice showing a significant increase in IBA1 expression levels in P301S transgenic mice aged 9/10 months when compared with age-matched WT littermates with transgenic mice aged 3 and 6 months. Beta-actin was used as a control of protein loading. ** p < 0.01 comparing P301S with age-matched WT; # p < 0.05 comparing with mice aged 9/10 months.
Similarly, the density of microglial cells was increased with age in the piriform cortex ( Fig. 5 Ag–Al), brainstem, and spinal cord and, remarkably, in the hippocampus ( Fig. 5 Bg–Bl) of P301S mice. The morphology of microglia also shifted to a reactive phenotype characterized by enlarged bodies and short thick ramifications in transgenic mice at the age of 9/ 10 months ( Fig. 5A , Bi, Bl).
Although the glial reactions in P301S mice partially overlapped with the distribution of neurons with hyper-phosphorylated tau, double labeling immunofluorescence revealed no direct relationship between tau hyperphosphorylation in neurons and astrocytic and microglial responses ( Fig. 5C ).
The numbers of astrocytes (GFAP-positive cells) and microglia (IBA1-positive cells) expressed in selected regions of the brain and spinal cord of P301S and wild-type mice are summarized in Figure 5D. An increase in the number of astrocytes was observed in the piriform cortex, hippocampus, brainstem, and spinal cord between 6 and 9/10 months of age in transgenic mice. A similar change was found for microglia, with more marked changes in the hippocampus. Other regions, such as the somatosensory cortex, amygdala, striatum, thalamus, and hypothalamus, did not show major glial modifications with disease progression, suggesting region specificity of astrocytes and microglia responses. This was not a mere reflection of regional differences in the baseline numbers of astrocytes and microglia concomitant to the P301S phenotype because differences were only apparent at advanced stages of the disease. In contrast to P301S mice, no significant modifications in numbers of astrocytes or microglia were recognized in any region in WT littermates.
Results of Western blotting to IBA1 in the somatosensory cortex and hippocampus at 3, 6, and 9/10 months paralleled those seen with immunohistochemistry. No major densitometric changes in the expression of IBA1 were seen in the somatosensory cortex between WT and P301S mice ( Fig. 5E ), which was in striking contrast with the significant increase in the expression of IBA1 in the hippocampus in P301S mice compared with WT at the age of 9/10 months ( Fig. 5F ).
Gene Expression of Cytokines and Mediators of the Immune Response
A broad panel of probes normalized with 2 housekeeping genes was used to study gene regulation of inflammatory responses in WT and P301S transgenic mice in the somatosensory cortex, hippocampus, and lumbar spinal cord in animals aged 1, 3, 6, and 9/10 months. Results are summarized in Tables 3 to 5 .
mRNA Expression of Selected Cytokine-Related Genes Involved in the Inflammatory Response in Wild-type and P301S Mice in Somatosensory Cortex
mRNA Expression of Selected Cytokine-Related Genes Involved in the Inflammatory Response in Wild-type and P301S Mice in Lumbar Spinal Cord
In WT animals, a region-dependent increase in the expression of several cytokines and mediators of the immune response was observed with age. C4b, Thr4, Ccl3, CxC110, and 116 mRNA expression increased with age in the somatosensory cortex; C1q11, C4b, Csflr, Cs3r, Tlr7, Ccl6, I16, Tnfrsfla, IllOrb, Tgfb1, and Tgfb2 increased although following individual pace in the hippocampus; and C4b, Csf3r, Ccl4, I16, and Il10 in the lumbar spinal cord ( Tables 3 –5).
In P301S mice, a similar increase in gene expression was observed with age in the 3 regions. Values were higher but mostly not significant in P301S versus WT mice at the ages of 1, 3, and 6 months. However, there was a dramatic increase in gene expression in P301S animals aged 9/10 months when compared with age-matched WT animals and P301S animals aged 6 months ( Tables 3 –5).
A significant increase (varying between p < 0.05 and p < 0.01) in Clqtnf7, C3arl, C4b, CsGr, Tlr4, Tlr7, Ccl3, Ccl4, Ccl6, CxCllO, Il1b, Tnf-a, Tnfrsfla, IllOrb, and Tgfb1 mRNA expression was found in the somatosensory cortex of P301S mice when compared with WT littermates ( Table 3 ).
Changes were more extended and more marked (between p < 0.01 and p < 0.001) in the hippocampus of transgenic mice. In addition to the abovementioned mRNAs, C1ql1, Csf1r, I16, I16st, Il10, and Il10ra mRNAs were markedly up-regulated (p < 0.01 and p < 0.001) in P301S mice at the age of 9/10 months ( Table 4 ).
mRNA Expression of Selected Cytokine-Related Genes Involved in the Inflammatory Response in Wild-type and P301S Mice in Hippocampus
Changes in gene expression were less marked in the lumbar spinal cord than in the hippocampus and somatosensory cortex (majority of p values between <0.05 and <0.01) and involved fewer genes: C3arl, C4b, Csflr, CsGr, Tlr7, Ccl3, Ccl4, Ccl6, CxCl10, Illb, Tnf-α, Il10, and Tgfb1 were upregulated in P301S animals aged 9/10 months when compared with WT littermates and transgenic mice aged 6 months ( Table 5 ).
Tau Oligomers and Dot Blots
Western blots incubated with the tau-5 antibody of the somatosensory cortex and hippocampus of P301S and WT mice aged 3, 6, and 9/10 months were overexposed to visualize tau bands of high molecular weight corresponding to tau oligomers. Several bands from 80 kDa to about 200 kDa were observed in transgenic mice but not in the corresponding age-matched littermates; the intensity increased with the age of P301S mice in the somatosensory cortex ( Fig. 6 , left panel). Increased olig-omeric species were also seen in the hippocampus in P301S mice when compared with WT littermates, but the intensity of high-molecular bands was equal to or even greater in P301S animals when compared with transgenic mice aged 9/10 months ( Fig. 6 , middle panel). Interestingly, the intensity of oligomeric bands was greater in the hippocampus when compared with the somatosensory cortex at any age. This was not caused by differences in protein loading, antibody concentration, or exposure time but rather by the greater number of oligomeric species in the hippocampus in comparison with the somatosensory cortex in P301S mice ( Fig. 6 , right panel).
Western blots using the antibody tau-5 of the somatosensory cortex (left panel) and hippocampus (middle panel) of P301S (TAU) and wild-type (WT) mice at the ages of 3, 6, and 9/10 months. Tau bands of molecular weight between 80 kDa and about 200 kDa are seen in the somatosensory cortex and hippocampus in transgenic mice, whereas no bands are seen in corresponding WT littermates. The intensity of the bands increases with disease progression in the somatosensory cortex, but it is already pronounced in the hippocampus in P301S mice at the age of 3 months. The intensity of oligomeric bands is greater in the hippocampus when compared with the somatosensory cortex at any age (right panel). This is not caused by differences in protein loading, antibody concentration, or exposure time but rather by the greater number of oligomeric species in the hippocampus in comparison with the somatosensory cortex in P301S mice. Films are overexposed and the bands of nonoligomeric tau bµmed out to facilitate visualization of oligomeric species. Beta-actin was used as a control of protein loading.
The presence of oligomeric species was also analyzed using dot blots with the ABN454/anti-Tau (T-22) oligomeric antibody. In our hands, this antibody stains neurofibrillary tangles in cases with AD and human tauopathies but failed to stain hyperphosphorylated tau deposits in P301S mice (data not shown). Moreover, dot blots showed nonspecific signals (elevated background) in both WT and P301S animals (data not shown). Therefore, the antibody was not considered useful to study tau oligomers in mice.
Respiration and ROS Production Analysis
High-resolution respirometry linked to the measurement of H 2 O 2 increase in the chamber as an indirect indicator of the release of superoxide radical from IM was performed with samples from 3- or 9/10-month-old WT and P301S cortices. In the nonphosphorylating state using complex I substrate glutamate, pyruvate, and lactate (Leak I state), increased respiration by IM was found in mice aged 3 months compared with mice aged 9/10 months both in WT (p < 0.05) and P301S transgenic mice (p < 0.05) ( Fig. 7A ). This difference accounted for a decreased C-II/C-I ratio (values with substrate succinate/ values of complex I substrates) in WT (p < 0.01) and P301S mice (p < 0.01) aged 3 months with respect to animals aged 9/10 months ( Fig. 7B ). Respiration was also higher in IM in WT (p < 0.05) and P301S (p < 0.01) mice aged 3 months after inhibition of complex III with antimycin A (Leak AmA) with respect to animals aged 9/10 months ( Fig. 7A ). The only significant difference between P301S and WT mice was observed after calculating the ratios of LeakN (Leak C-I,II/OXP [adenosine diphosphatase]; p < 0.05) and LeakO (Leak OU [OXP plus oligomycin]/OXP; p < 0.05) at 9/10 months, indicating that proton leak increases in aged P301S ( Fig. 7B ). The highest rates of ROS production per oxygen uptake were recorded at the leak states. Isolated mitochondria from P301S mice at 9/10 months of age produced more ROS than WT at the Leak C-I,II state (p < 0.05) ( Fig. 7C ).
Respiration and reactive oxygen species (ROS) production measurements from isolated mitochondria (IM) of 3- and 9/10-month-old wild-type (WT) and P301S cortices. (A) Respiration in different leak, oxidative phosphorylation, and uncoupled states. O 2 flux per second is represented per mass of IM added to the chamber. Substrates and inhibitors of the respiratory system were added to establish the different respiratory states as follows: Leak CI (glutamate, malate, pyruvate); Leak CI, II (succinate); adenosine diphosphate (OXP); Leak Omy (Oligomycin); ETS (FCCP); AmA (Antimycin A). IM from both WT and P301S mice aged 3 months consume more O 2 than 9/10-month animals at the nonphosphorylating state linked to CI activity and after inhibition of CIIl with AmA. (B) Respiration ratios (Log2 scale) calculated with the respiration values plotted in A reveal decreased CII/CI ratio in both WT and P301S mice aged 3 months with respect to animals aged 9/10 months. Proton leak (Leak N and Leak O) is increased in P301S 9/10 months IM compared with the corresponding WT. CI/CII, (Leak Cl,ll-Leak Cl)/Leak CI; Leak N, Leakl,ll/OXP; Leak O, Leak Omy/OXP; UCR, ETS/OXP; FCR, ETS/Leak Omy. (C) ROS production is represented as H 2 O 2 production per oxygen uptake at the same states plotted in A (Log2 scale). IM from P301S mice at 9/10 months of age produce more ROS than WT at the Leak C-I,ll state. In all graphs, mean ± SEM is represented. * p < 0.05 compared with WT animals aged 9/10 months.
Oxidative Stress
Expression levels of the mitochondrial antioxidant enzyme SOD2 were assessed in the somatosensory cortex and hippocampus in P301S transgenic mice and WT littermates at the ages of 3, 6, and 9/10 months. The SOD2 protein levels in the somatosensory cortex were significantly lower in transgenic mice at the age of 3 months (p < 0.05), and a similar tendency was observed in hippocampus when compared with WT animals. However, no significant differences between genotypes were seen at the age of 6 and 9/10 months in any of the brain structures evaluated. An age effect in SOD2 levels was observed in the hippocampus of WT animals because a significant increase at 3 months was recorded when compared with 9/10-month-old samples ( Fig. 8C , D ).
Expression levels of neuroketal (A, B) and superoxide dismutase 2 (SOD-2) (C, D) in the somatosensory cortex (A, C) and hippocampus (B, D) in P301S transgenic mice and wild-type (WT) littermates at the ages of 3, 6, and 9/10 months. No modifications in the expression levels of neuroketal are seen in WT and P301S mice with age. A significant reduction in SOD-2 levels is observed in the somatosensory cortex and similar tendency in the hippocampus of transgenic mice aged 3 months versus WT littermates. Later on, no differences in SOD-2 levels caused by genotype are seen in the somatosensory cortex and hippocampus in mice aged 6 and 9/10 months. WT mice aged 3 months exhibited increased SOD-2 levels in the hippocampus with respect to 9/10-month-old WT animals. Data are represented as the mean ± SE. * p < 0.05 compared with WT animals. # p < 0.05 compared with animals aged 9/10 months.
The study of nonenzymatic protein modifications revealed a significant increase only in the levels of GSA in the somatosensory cortex and hippocampus and CEL in the hippocampus in WT mice aged 9/10 months when compared with mice aged 1 month. No modifications with age were observed in the spinal cord. No changes in the steady-state levels of GSA, CEL, CML, MDAL, and SC were found in P301S transgenic mice at any age ( Table 6 ). In line with these findings, neuroketal levels were similar in WT and transgenic mice and did not change with age in any of these groups in the somatosensory cortex and hippocampus ( Fig. 8 A, B).
Nonenzymatic Protein Modifications for Wild-type and P301S (TAU) Mice Aged in the Somatosensory Cortex, Hippocampus, and Lumbar Spinal Cord
Fatty Acid Profile
Fatty acid composition is modified with age in WT mice, although with regional variations among the somatosensory cortex, hippocampus, and lumbar spinal cord.
Decreased levels of 14:0, 16:0, 16:ln-7, 18:0, 18:2n-6, 18:3n-6, 18:4n-3, 20:2n-6, 22:5n-6, and 24:5n-3 together with increased levels of 18:ln-9, 20:ln-9, 22:4n-6, 22:5n-3, 22:6n-3, and 24:6n-3 were found in the somatosensory cortex when comparing mice aged 9/10 months with 1-month-old mice. Decreased 14:0, 16:0, 18:2n-6, 18:4n-3, 20:0, 20:2n-6, 20:4n-6, and 22:5n-6 accompanied by 18:ln-9, 20:ln-9, 22:4n-6, 22:5n-3, and 24:6n-3 (similar to that seen in the somatosensory cortex except 22:6n-3) was found in the hippocampus in the same period. Reduced levels of 16:0, 16:ln-7,18:2n-6,18:3n-6,18:4n-3, 20:2n-6, 20:3n-6, 20:4n-6, 22:5n-6, 22:5n-3, and 24:5n-3 occurred in the lumbar spinal cord; however, only 20:ln-9, 22:4n-6, and 24:6n-3 were increased in the lumbar spinal cord when comparing mice aged 9/10 months with WT mice aged 1 month.
Modifications in P301S mice were also subject to regional particularities. Although transient changes in fatty acid composition were seen at 3 and 6 months in the somatosensory cortex and hippocampus, major differences between P301S and WT mice at the age of 9 months included a significant decrease in 18:3n-6, 18:3n-3, and 18:4n-3 and an increase in 22:6n-3 and 24:6n-3 in the somatosensory cortex; decrease in 20:3n-6 and 22:5n-6 and an increase in 18:ln-9, 18:3-n3, 18:4n-3, 20:5:n-3, 22:4n-6, and 22:5n-3 in the hippocampus. In contrast to the somatosensory cortex and hippocampus, only the levels of 20:3n-6, 20:4n-6, and 22:5n-3 were decreased and 22:0 levels increased in P301S mice at the age of 9/10 months (Table, Supplemental Digital Content 3, part A, Supplementary Data , Table, Supplemental Digital Content 4, part A, Supplementary Data , and Table, Supplemental Digital Content 5, part A, Supplementary Data ).
Global fatty acid indexes were also subject to regional and phenotypic variations. Increased UFA, MUFA, and PUFAn-3 together with increased DBI, PI, and AI were found in the somatosensory cortex of WT and P301S mice. In contrast, decreased PUFA, PUFAn-3, and PUFAn-6 and reduced PI and AI (accompanied by increased UFA and MUFA indexes) were observed in the spinal cord of WT with age; yet only reduced PUF An- 6 was noted in P301S mice aged 9/10 months.
Very divergent values between WT and P301S mice were identified in the hippocampi. Increased PUFAn-3 and decreased PUFAn-6 together with increased AI occurred in WT mice, but reduced total PUFA and PUFAn-6 accompanied by increased MUFA and reduced DBI and PI were identified in the hippocampi of P301S mice aged 9/10 months (Table, Supplemental Digital Content 3, part B, Supplementary Data , Table, Supplemental Digital Content 4, part B, Supplementary Data , and Table, Supplemental Digital Content 5, part B, Supplementary Data ).
For estimated desaturase (DxD) activities, increased D5D (n-6) and D6D(n-3)24 in the somatosensory cortex; D9D(n-9), D8D(n-6), and D6D(n-3)24 in the hippocampus; and D5D(n-6) and D6D(n-3)24 in the lumbar spinal cord together with decreased D6D(n-3)18 and D8D(n-6) in the spinal cord respectively, were observed in WT mice aged 9/10 months when compared with mice aged 1 month. Increased D9D(n-7) and D9D(n-9) and decreased D5D(n-6) in the hippocampus and increased D8D(n-6) activity in the somatosensory cortex occurred in P301S transgenic mice at the age of 9/10 months (Table, Supplemental Digital Content 3, part C, Supplementary Data , Table, Supplemental Digital Content 4, part C, Supplementary Data , and Table, Supplemental Digital Content 5, part C, Supplementary Data ).
Elongase activity varied from 1 region to another in WT animals, but they were barely modified in P301S animals. Increased Elovl3, Elovl2-5(n-6), and Elovl2(n-3) and reduced Elovll-3-7 activity in the somatosensory cortex; reduced Elovll-3-7 and Elovl5 and increased Elovl3, Elovl6, Elovll-3-7, and Elovl2-5(n-6) in the hippocampus; and reduced Elovl2-5(n-3) but increased Elovl3, Elovl6, and Elovll-3-7 and reduced Elovl2-5(n-3) in the spinal cord were noted in WT mice aged 9/10 months when compared with 1-month-old mice. In contrast, only reduced Elovl2-5(n-3) in the somatosensory cortex and increased Elovll-3-7 and reduced Elovl2-5(n-6) in the spinal cord, with no modifications in elongase activity in the hippocampus, were observed in P301S transgenic mice (Table, Supplemental Digital Content 3, part D, Supplementary Data , Table, Supplemental Digital Content 4, part D, Supplementary Data , and Table, Supplemental Digital Content 5, part D, Supplementary Data ).
Finally, peroxisomal β-oxidation decreased with age mainly involving the n-6 pathway in the somatosensory cortex, hippocampus, and lumbar spinal cord in WT animals aged 9/ 10 months when compared with WT aged 1 month. The n-6 pathway was also reduced in P301S animals aged 9/10 months in the somatosensory cortex and hippocampus, but the lumbar spinal cord was not affected (Table, Supplemental Digital Content 3, part E, Supplementary Data , Table, Supplemental Digital Content 4, part E, Supplementary Data , and Table, Supplemental Digital Content 5, part E, Supplementary Data ).
Discussion
Clinical and Pathologic Phenotype
We found a reduction in body weight and reduced survival in P301S transgenic mice versus their WT littermates. Increased mortality is currently associated with motor deficits, and it is probably linked, at least in part, to reduced water and food intake. Whether alterations in the neurons of the myenteric plexus here observed can lead to nutritional deficiencies must also be considered.
Behavioral alterations start between the fifth and sixth months of age in the P301S transgenic mouse model generated under the control of the murine thy1.2 (mThy1.2) promoter ( 6–10 ). Abnormal behavior starts earlier at about 4 months in another model generated under the control of the MoPrP promoter (PS19 line) ( 7 , 25 ). In both lines, clinical manifestations are similar and are characterized by disinhibition, anxiety, increased locomotion, impaired memory, altered nociception, and terminal severe motor decline ( 6 , 25 ). Our behavioral studies in the PS19 line match those already described, thereby permitting comparisons with other parameters and implications of the present observations on previous data reported in these models.
Abnormal hyperphosphorylated tau deposition is restricted to neurons and is first noted in the entorhinal and piriform cortices and more sparsely in the somatosensory cortex at the first month. However, Western blotting of total homogenates reveals a marked increase in tau and hyperphosphorylated tau in the brainstem at the same age, indicating that tau hyperphosphorylation occurs at very early stages of the disease. Hyperphosphorylated tau increases in the same regions and extends to other regions with disease progression. Considering these data, major clinical signs appearing at later stages might be preceded by a preclinical period as occurs in the majority of human neurodegenerative diseases. On the other hand, major clinical signs are preceded by subtle changes such as increased ultrasound emission by P301S pups ( 10 ) and enhanced prepulse inhibition, a determination used as a marker of sensorimotor gating, which occurs in parallel with prime neuropathologic changes in associated brain regions in mThyl.2 P301S transgenic line (25). Thus, more precise methods than those currently used in behavioral studies might unveil hidden disease manifestations at preclinical or prodromal stages.
Characteristics of the Tauopathy
The PS19 line produces about 6 times more 4Rtau than WT littermates (7), values that are also present in this study. P301S transgenic mice driven by the mThyl.2 promoter produce smaller amounts of tau (about 3 times) (6). This may explain minor differences between the 2 models with a relatively delayed onset of clinical disease in the line with less 4Rtau burden (5). 3Rtau is very abundant in WT and P301S mice at 1 month of age, as occurs in normal development, but only hyperphosphorylated tau is found at this age in P301S mice. This suggests that hyperphosphorylated tau is 4Rtau as expected for tau mutations involving exon 10 (3).
Another important point is the modification in the composition of hyperphosphorylated tau deposits with the progression of the disease. The vast majority of involved neurons throughout the disease contain only hyperphosphorylated tau. Only a small percentage of neurons containing hyperphosphorylated tau (−30%) in P301S mice aged 9/10 months contained abnormal tau conformation, as revealed with the antibody Alz50, which recognizes modifications at amino acids 5 to 15, and only approximately 20% contain tau truncated at aspartic acid 421 as identified by the tau-C3 antibody. The percentage of neurons containing altered conformation of tau and truncated tau is negligible in hyperphosphorylated tau-containing neurons in animals aged 3 and 6 months. Several studies have shown that phosphorylation of tau precedes altered tau conformation, and that tau truncation occurs only after complex tau phosphorylation and altered tau conformation in neurofibrillary tangles ( 26–31 ). Moreover, tau truncation in neurons at aspartic acid 421 is needed to develop the cascade of NFT formation, including recruitment of full-length tau ( 32–34 ). Therefore, neurons in P301S transgenic mice have the characteristics of pretangle neurons. It is at advanced stages that truncated tau at aspartic acid 421 recruits tau and eventually blocks the ubiquitin-proteasome system, thus leading to neurofibrillary tangle formation (4).
The regional distribution of hyperphosphorylated tau deposition in neurons, as seen by immunohistochemistry, extends progressively from the piriform and entorhinal cortices and upper layers of the somatosensory cortex to the basal fore-brain, septal nuclei, striatum, hypothalamus, granule cell layer of the cerebellum, brainstem, ventral spinal cord, dorsal spinal cord, spinal ganglia, and myenteric plexus in approximately that order. This is grossly consistent with the seeding hypothesis of abnormal tau spreading from 1 neuron to another in a prion-like manner. In fact, tau seeding and propagation of abnormal tau have been demonstrated using P301S transgenic models ( 12–14 ). However, the reason for the relatively late involvement of the hippocampus proper and dentate gyrus in this model is elusive, as is the preservation of particular cell types such as Purkinje cells. Such neuron vulnerability probably depends on different causes, one of them being the burden and phosphorylation/dephosphorylation tµmover of tau in particular cell types; another may be the characteristics of the promoter used in the generation of the different models.
Therefore, it is not clear why the amygdala is apparently spared in P301S mice generated using the mThy1.2 promoter, whereas it is severely affected in P301S transgenic mice generated under the control of the MoPrP promoter. These observations suggest that, in some way, seeding is also modulated by specific cell vulnerability. If true, it would be important to understand the mechanisms by which a neuron can act as a transfer of seeding without being affected itself by the abnormal protein.
Astrocyte and Microglia Responses and Cytokines and Mediators of the Immune Response Are Upregulated in P301S Mice Only at Advanced Stages of the Tauopathy
Neuroinflammation seems to play a role in sporadic ( 35–37 ) and familial tauopathies ( 38–40 ), but all the studies have been carried out at clinical or at terminal stages of the diseases. The present findings related to inflammatory responses in advanced stages of P310S transgenic mice are similar to those found in a human case of FTLD with P301S mutation (39).
Inflammatory mediators and microglial activation are detected after tau aggregation in neurons in the mThyl.2 P301S line (39). However, it has been reported that microglial activation precedes neurofibrillary tangle formation in the PS 19 line ( 7 ) and that anti-inflammatory treatment reduces neurodegeneration in this line (41). The present findings show regional differences in the astrocyte and microglia responses both appearing at advanced stages of the disease, months after the appearance of hyperphosphorylated tau deposition in neurons. Major changes occur between months 6 and 9/10 at preterminal stages of the disease. These morphologic changes match the modulation of gene expression of various cytokines and mediators of the immune response; all of them are upregulated in P301S mice aged 9/10 months when compared with mice aged 1,3, and 6 months.
To consider this aspect in more detail, it is worth stressing that upregulation of selected cytokines and mediators occur in WT mice with age. This observation is in line with previous observations in mice and humans ( 42 , 43 ). Regional differences in baseline gene expression here observed have also been demonstrated in mice and humans ( 42–44 ).
The shift in gene regulation observed in P301S mice at the age of 9/10 months is not a mere acceleration of brain aging, as many other cytokines and mediators are recruited and upregulated in transgenic mice, and the increments are double or triple those seen in corresponding WT littermates at the age of 9/10 months. In addition, inflammatory responses are region dependent; the most dramatic change occurs in the hippocampus, followed by the somatosensory cortex and then by the spinal cord, the 3 regions comprehensively tested for gene regulation of cytokine and related mediators. The intensity of the molecular response matches the regional differences in the microglial response, that is, the microglial reaction is more marked in the hippocampus than in the somatosensory cortex, which in tµm is more severe than in the spinal cord. Importantly, gene regulation compromises proinflammatory and antiinflammatory cytokines, chemokines, toll receptors, members of the complement system, and colony-stimulating factors and involves 21, 15, and 12 of the 22 genes tested in the hippocampus, somatosensory cortex, and spinal cord, respectively. Regional differences in gene expression have also been reported in human neurodegenerative diseases and corresponding animal models ( 42–44 ).
An important point is the relationship between hyper-phosphorylated tau deposition and the inflammatory response in P301S transgenic mice. Contrary to the seminal descriptions in the PS19 line (7), microglia reaction and gene regulation of cytokines and mediators of the immune response occur at advanced stages of the disease, months before hyper-phosphorylated tau deposition occurs in neurons. Therefore, the mere deposition of abnormal tau in neurons does not trigger inflammatory responses. Rather, inflammatory responses occur in parallel with the occurrence of altered conformation of tau and truncated tau in neurons, as revealed with Alz50 and tau-C3 antibodies, and with increased levels of soluble oligomers as revealed with anti-tau-5 antibodies in P301S mice when compared with age-matched littermates. Curiously, expression levels of oligomeric tau species are higher and appear earlier in the hippocampus when compared with the somatosensory cortex. This is an interesting point because the major shift in the inflammatory response occurs in humans with increased levels of β-amyloid oligomers in brain, soluble β-amyloid species in APP/PS1 transgenic mice, and extracellular abnormal PrP deposition in PRNP-humanized mice intrace-rebrally injected with Creutzfeldt-Jakob disease MM1 subtype brain homogenates (murine Creutzfeldt-Jakob disease) (42).
In the same line, the intensity and extent of the inflammatory response are greater in the hippocampus than in the somatosensory cortex in P301S animals, and this response correlates with the higher levels of oligomeric species (and truncated tau) in the hippocampus in relation to the somatosensory cortex.
Cell Death
It is worth stressing that no evidence of cell death was found in P301S mice until the age of 9/10 months and, when present, it was localized in very discrete areas involving scattered neurons in the brainstem and the dentate gyrus, CA3, and CA1 regions of the hippocampus. Apoptotic cell death as revealed with the method of in situ labeling is very rare in the somatosensory cortex, amygdala, diencephalic nuclei, and spinal cord. The regional distribution and localization of apoptosis are similar to that seen for neurons with truncated tau and ubiquitin deposits.
Respiration in IM Is Slightly Altered in P301S Transgenic Mice Only at Advanced Stages of the Disease
Mitochondrial dysfunction and oxidative stress have been reported to play a major role in the pathogenesis of an unrelated model of P301S tauopathy in mice (16). The generation of this line has not been reported but, curiously, tau deposition is delayed several months, appearing at about the 10th month, and oxidative stress and mitochondrial dysfunction precede tau pathology in this transgenic mouse model of P301S tauopathy (16). Differences between this model and previous well-characterized lines of P301S tauopathy and with any other tauopathy in mice expressing P301L, G272V, N279K, R406W, V337M, K369I, and combined tau mutations ( 5 ) render that line suboptimal for comparative studies.
Results of the present study using high-resolution respirometry demonstrate that subtle functional alterations in IM only occur in our P301S transgenic mouse line at the age of 9/10 months. The only significant difference between P301S and WT mice was observed after calculating the ratios of LeakN (Leak C-I,II/OXP [adenosine diphosphatase]) and LeakO (Leak Oli [OXP plus oligomycin]/OXP). Increased proton leak can be linked to increased oxidative stress and increased ROS production, although the exact mechanism remains obscure.
Oxidative Stress and ROS Production Occur at Advanced Stages of the Disease in P301S Mice
Results of the present study indicate that no major modifications in neuroketal-modified protein content occur during disease progression, and this occurs in parallel with preserved expression levels of SOD2 even at advanced stages of the disease. Moreover, although increased levels of GSA in the somatosensory cortex and hippocampus and CEL levels in the hippocampus occur with aging in WT and transgenic mice, no increased concentration of GSA, CEL, CML, or MDAL is identified in P301S mice aged 9/10 months. Moreover, no differences in ROS production have been observed between WT and P301S mice aged 3 and 6 months after assessing net ROS production measured simultaneously with respiration flux by means of the detection of the fluorescence emitted by Amplex Ultrared. However, P301S mice aged 9/10 months produced 38.4% more ROS than WT at the Leak C-I,II state. Together, the present findings using 3 different methodological approaches suggest that oxidative stress occurs at very advanced stages of the disease and seems not to play a major pathogenic role at early and middle stages of this murine tauopathy. This finding is in accordance with the observed increased oxidative stress and GFAP oxidative damage at terminal stages of human FTLD linked to P301L mutations in the tau gene (45).
Lipids
The brain contains a large amount of lipids and the largest diversity of lipid classes and lipid molecular species. Our findings verify that, in WT mouse brain, fatty acids are saturated monounsaturated or polyunsaturated hydrocarbon chains that normally vary from 14 to 24 carbons in length, with an average chain length strictly maintained at about 18 carbon atoms. Furthermore, the available data demonstrate the existence of cross-regional differences in mouse brain fatty acid composition. Thus, despite maintaining an interregional stable average chain length (18 carbon atoms), there are cross-regional differences with respect to the ratio satu-rated:unsaturated ( 45 –55:45–55), global content of DHA and AA (between 15% and 25%), and the distribution of MUFA and PUFA. The higher the presence of MUFA for a given region, the lower the PUFA content. Assuming that the age range analyzed in this study is a period of brain maturation, it is plausible to postulate that the described changes in the different regions represent adaptation to age to achieve in the adult state a particular fatty acid composition according to the specific activities and needs developed by the different regions. Similar regional variations are also observed in P301S mice. These interregional differences can be ascribed to physiologic changes in the activity of the enzymes (desaturases, elongases, and peroxisomal beta-oxidation) belonging to the respective biosynthetic pathways. Interestingly, the pressure over the homeostatic system induced by the pathologic condition seems to intensify (although modestly) changes induced by age, particularly for somatosensory cortex and hippocampus, probably expressing adaptive responses to cope with oxidative stress to ensure cell survival.
Lipid modifications in P301S mice when compared with WT littermates offer little to increase our understanding of possible implications of lipid composition as a pathogenic factor in the development of associated lesions linked to the tauopathy. Regional DHA levels are not dramatically different in P301S and WT mice. However,increased PUFAn-3 and decreased PUFAn-6 together with increased anti-AI occur in the hippocampus in WT when compared with P301S mice. In contrast, reduced total PUFA and PUFAn-6, accompanied by increased MUFA, and reduced PI occur in the hippocampus of P301S mice aged 9/10 months. Taken together, total lipid profiles, although indicative of regional particularities in both WT and P301S mice, seem to not be particularly contributory to the pathology, including neuroinflammatory responses, in P301S mice.
Concluding Comments
The present findings show that P301S transgenic mouse (line) replicates clinical and neuropathologic findings of FTLD-tau cases with only neuronal involvement, thus appearing to be a suitable model for the study of possible cumulative associated pathology with disease progression in human FTLD-tau cases. Clinical disease appears at a certain time before tau oligomers are produced in brain and hyperphosphorylated tau accumulates in neurons; however, alterations in mitochondrial function, ROS production, oxidative damage of proteins, astrogliosis, microgliosis, and increased expression of cytokines and mediators of the immune response occur at advanced stages of the disease with marked regional differences. Apo-ptosis is only observed in certain regions at advanced stages. Modifications in lipid composition occur with age with marked regional differences in WT and P301S animals, but the pathogenic role of these changes seems to be limited regarding disease progression. These observations center on abnormal tau production, increased tau oligomers, and neuronal deposition of hyperphosphorylated tau as the primary events in P301S transgenic mice, which are followed by late alterations in several molecular pathways covering astroglia and microglia responses, altered respiration, increased ROS production and oxidative damage, neuroinflammation, and selected neuronal death.
Acknowledgement
The authors thank T. Yohannan for editorial help.
References
Supporting Information
Author notes
This study was funded by the Spanish Mnistry of Health, Institute Carlos III: FIS grants PI14/00757 and PI14/00328.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site ( http://jnen.oxfordjournals.org/ ).
