A multifactorial model of pathology for age of onset heterogeneity in familial Alzheimer’s disease

Presenilin-1 (PSEN1) mutations cause familial Alzheimer’s disease (FAD) characterized by early age of onset (AoO). Examination of a large kindred harboring the PSEN1-E280A mutation reveals a range of AoO spanning 30 years. The pathophysiological drivers and clinical impact of AoO variation in this population are unknown. We examined brains of 23 patients focusing on generation and deposition of beta-amyloid (Aβ) and Tau pathology profile. In 14 patients distributed at the extremes of AoO, we performed whole-exome capture to identify genotype–phenotype correlations. We also studied kinome activity, proteasome activity, and protein polyubiquitination in brain tissue, associating it with Tau phosphorylation profiles. PSEN1-E280A patients showed a bimodal distribution for AoO. Besides AoO, there were no clinical differences between analyzed groups. Despite the effect of mutant PSEN1 on production of Aβ, there were no relevant differences between groups in generation and deposition of Aβ. However, differences were found in hyperphosphorylated Tau (pTau) pathology, where early onset patients showed severe pathology with diffuse aggregation pattern associated with increased activation of stress kinases. In contrast, late-onset patients showed lesser pTau pathology and a distinctive kinase activity. Furthermore, we identified new protective genetic variants affecting ubiquitin–proteasome function in early onset patients, resulting in higher ubiquitin-dependent degradation of differentially phosphorylated Tau. In PSEN1-E280A carriers, altered γ-secretase activity and resulting Aβ accumulation are prerequisites for early AoO. However, Tau hyperphosphorylation pattern, and its degradation by the proteasome, drastically influences disease onset in individuals with otherwise similar Aβ pathology, hinting toward a multifactorial model of disease for FAD. In sporadic AD (SAD), a wide range of heterogeneity, also influenced by Tau pathology, has been identified. Thus, Tau-induced heterogeneity is a common feature in both AD variants, suggesting that a multi-target therapeutic approach should be used to treat AD. Supplementary Information The online version contains supplementary material available at 10.1007/s00401-020-02249-0.


Index Item Page
Extended Methods 4 Tables 11 Table 1. Demographic and cognitive performance according to age of onset in 122 PSEN1 E280A patients 11  Table 8. pTau pathology comparison E-AOFAD vs LOSAD in PSEN1 E280A FAD 21 Table 9. pTau-related Kinases levels according Age of Onset in PSEN1 E280A FAD 22  Extended Methods

Patients and clinical data collection
Descendants of patients with confirmed PSEN1 E280A mutations were enrolled into the E280A Antioquia cohort study, an ongoing work at the University of Antioquia, Colombia. Participants were included if they were aged over 17 years. There were no other exclusion criteria for medical and neuropsychological monitoring. All participants (both carriers and non-carriers of PSEN1 E280A) or their guardians provided written informed consent for participation in the study; if the physician thought that a participant had dementia, their guardian provided written informed consent. All assessed participants with no evident dementia and examiners were masked to genetic status throughout monitoring. For genetic analyses, genomic DNA was extracted from blood by standard protocols, and PSEN1 E280A characterization was done as previously described. 19 Genomic DNA was amplified with the primers PSEN1-S 5′ AACAGCTCAGGAGAGGAATG 3′ and PSEN1-AS 5′ GATGAGACAAGTNCCNTGAA 3′. We used the restriction enzyme BsmI for restriction fragment length polymorphism analysis. The study was approved by the medical ethics board of the University of Antioquia. Follow up was conducted as previously described (1). Briefly, follow-up examinations included medical and neuropsychological assessments, which focused on registration of memory complaints and general cognitive function. Neuropsychological tests were performed by neuropsychologically trained personnel. Medical history and neuropsychological assessments were stored at the systematized information system for the neuroscience group of Antioquia (SISNE).
For cognitive assessment of differences between ages of onset we used a protocol including the CERAD (consortium to establish a registry for Alzheimer's disease) neuropsychological test battery with additional neuropsychological tests, as previously described (1). Basic demographic information was collected, including schooling time. Minimental (MMSE) testing served as baseline examination. Tests were applied according to studied cognitive domain as follows. For memory assessments, we used the memory of three phrases test, Rey-Osterrieth complex figure test (recall), list of words tests (total corrects, total intrusions, recall, intrusions recall, recognition "yes", and recognition "no"), and recall of line drawings test. We assessed language ability with the verbal fluency and naming test. To assess constructional praxis, we used the constructional praxis test and the Rey-Osterrieth complex figure test (copy). This expanded CERAD neuropsychological protocol has been validated for the Colombian population in participants over 50 years of age and it has been also established for normal parameters for participants under 50 years of age (1).
Furthermore, for neuropathological studies, Alzheimer's disease brain samples were collected from the Brain tissue bank from the University of Antioquia. PSEN1E280A FAD cases and sporadic cases or their families, signed informed consent for post-mortem brain donation and tissue use in scientific studies. All procedures were performed following ethical board approval from the University. Sporadic cases were selected based on clinical diagnosis of probable AD, lack of family history of dementia and tested as non-carriers for PSEN1 E280A mutation. Control cases were collected in the brain bank of the Bellvitge Hospital, Barcelona, Spain. They were selected based on lack of brain trauma, cognitive or neurological symptoms before death.

Morphological methods
Histopathological methods. All morphological analyses were performed on 3 μm thick de-paraffinized sections from cortices of SAD and PSEN1 E280A FAD cases (Table 3). Immunohistochemical stainings were performed following pre-treatment for antigen retrieval and probed with monoclonal anti-A antibody, anti pTau antibody and anti-A1-42 antibody (Table 16). All immunohistochemical stainings were performed on an automated Ventana HX system (Ventana-Roche Medical systems, Tucson, AZ, USA) following the manufacturer's instructions. Experimental groups were stained in one run for each antibody to provide uniform staining conditions. Primary antibodies were visualized using a standard diaminobenzidine streptavidin-biotin horseradish peroxidase method (Sigma Aldrich, Hamburg, Germany). For quantification of primary antibodies immunosignal, three representative regions (0,1349 mm2 each) were analyzed by quantifying the area immunoreactive for each antigen using the AxioVision 4.6 software (Carl Zeiss, Oberkochen, Germany) according to published methods (2).
Ultrastructural analysis. Ultrastructural analysis was performed using glutaraldehyde-fixed brain tissue from SAD, EOFAD, AOFAD and LOFAD patients as previously described (2). Temporal cortex samples were excised from paraformaldehyde fixed tissue after localizing specific areas of extracellular pTau deposits or an equivalent area from LOFAD cases. Samples were fixed with glutaraldehyde and chrome-osmium, dehydrated in ethanol, and embedded in Epon 812 (Serva Electrophoresis GmbH). After polymerization, 1-μm-thick sections were cut, stained with toluidine blue, and checked for presence of amyloid plaques. To further process them for electron microscopy, relevant specimens were cut into 60-to 80-nm-thick sections, which were contrasted with uranyl acetate and lead solution. Sections were viewed under a LEO EM 912AB electron microscope (Zeiss).
Tissue clarification and imaging. Formalin fixed 1 cm length x 1 cm width x 500 m thick, temporal cortex samples from 5 EOFAD, 5 AOFAD and 5 LOFAD cases were clarified using a CLARITY protocol as previously described (17 Samples were incubated in 87% glycerol 3 h at room temperature prior to imaging and fixed flat to the bottom of an imaging dish with 63% TDE. A Z-stack of a minimum thickness of 100 m was acquired with a Leica TCS SP5 confocal microscope (Leica microsystems, Wetzlar, Germany). 3D synaptophysin positive particle counting was performed using the 3D objects counting plugin on ImageJ 1.52p (NIH, USA).

Biochemical methods
Preparation of Soluble and Insoluble protein fractions. Soluble and insoluble fractions from brain tissue were isolated as described by Tremblay C and colleagues (3). Briefly, temporal cortex from SAD and FAD patients (~100 mg) was homogenized in 4 volumes of Tris Buffered Saline (TBS) containing a cocktail of phosphatase and protease inhibitors (Roche, Mannheim, Germany). Samples were sonicated three times for 10 s and centrifuged at 100,000 g for 20 min at 4°C to obtain a TBS-soluble fraction containing cytosolic and extracellular proteins (Soluble fraction). The pellet was sonicated using 4 volumes of lysis buffer (150 mM NaCl, 10 mM NaH2PO4, 1% Triton X-100, 0.5% SDS, and 0.5% deoxycholate) with protease and phosphatase inhibitors. The homogenate was centrifuged at 100,000 g for 20 min at 4°C. The pellet was homogenized in 200 µl of 90% formic acid and sonicated three times for 10 s to isolate the Insoluble protein fraction. Protein fractions were stored at -80ºC for further experiments.
Extraction of total protein from tissue. Temporal cortex from Control, SAD, E-AOFAD and LOFAD cases ( Table 3) were cleared of meninges and only grey matter was used for the procedure. Approximately 250 mg of tissue were cut in small pieces, poured into a glass Dounce tissue grinder type B and homogenized with ten even strokes in 1 mL of lysis buffer containing 150 mM NaCl, 20 mM Tris pH 7.4, 1 mM EDTA, 10% Glycerol, 1% NP40 and a cocktail of phosphatase and protease inhibitors (Roche, Mannheim, Germany). The homogenate was centrifuged at 13,000 g for 10 min at 4ºC and the proteins present in the supernatant were quantified using the bicinchoninic acid method (BCA Protein Assay Kit, Thermo, Dreieich, Germany). The protein extracts were stored at -80ºC for further experiments.
Western blotting. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Once proteins were quantified, SDS-PAGE was carried out using a 20 well electrophoresis system (VWR, Radnor, PA, USA) or a Miniprotean system (BioRad, München, Germany). Samples were mixed with loading buffer (0.375 M Tris pH 6.8, 50% glycerol, 10% SDS, 0.5 M DTT and 0.002% bromophenol blue) and heated to 95ºC for 5 minutes. About 25 -30 μg of protein were loaded into each well. After electrophoresis, proteins were transferred to nitrocellulose membranes (BioRad, München, Germany) using a Trans-blot Turbo Transfer system (BioRad, München, Germany) at 300 mA for 2 h. The membranes were incubated for 1 h in 5% non-fat milk dissolved in TTBS (100 mMTris pH 7.5, 500 mMNaCl, 0.02% Tween-20) and incubated overnight at 4°C with primary antibody (Table 16). Subsequently, membranes were washed with TTBS and incubated with secondary antibody (Table 17) coupled to peroxidase for 1 h at room temperature. Immunoreactive signal was developed with the ECL Western Blotting chemiluminescence system (SuperSignal West Pico Chemiluminiscent Substrate, Thermo, Dreieich, Germany) and detected with a ChemiDoc system (BioRad, München, Germany). For some of total vs phosphorylated protein kinase blotting where the species of each antibody allowed it, the same membrane was first incubated with the phosphorylated antibody, it was developed and re-incubated with the total protein antibody of a different species. The images were analyzed using the quantification software QuantityOne (BioRad, München, Germany). The results of each sample were normalized to GAPDH and compared between groups. To minimize interassay variation, the samples from all experimental groups were processed in parallel. Regarding A oligomers, there is an intrinsic difficulty in distinguishing between Aoligomers and APP fragments still containing the A sequence (4) and although it has been suggested that A oligomers rather than plaques are the main factor in A-related pathogenicity, existing evidence regarding their possible role in AD is unclear (5). Therefore, only small oligomers (< 40 kDa) were considered for quantification.
In vitro gamma secretase activity assay. Detergent resistant membrane preparation from human SAD (n=5) and FAD brains (n=23) CHAPSO resistant membranes were prepared for human brains (frontal cortex) which were frozen within 12 h postmortem as previously described in (7) with minor modifications. After carefully removing leptomeninges and blood vessels, < 250 mg blocks of tissue were homogenized in ~ 10 volumes of 10% sucrose in MES buffer (25 mM MES, pH 6.5, 150 mMNaCl) containing 1% CHAPSO (Sigma) and protease inhibitors. The homogenate was mixed with equal volume of 70% sucrose in MES buffer. 4 ml was placed at the bottom of an ultracentrifuge tube (Beckman, 344059) and overlaid with 4 ml of 35% sucrose and 4 ml of 5% sucrose prepared accordingly. The obtained gradients were centrifuged at 39,000 rpm for 20 h at 4°C on a SW 41Ti rotor (Beckman). After centrifugation the raft fraction (interface of 5%/35% sucrose) was carefully collected and re-centrifuged (50,000 rpm, 60 min, 4°C) in 20 mM PIPES, pH 7, 250 mM sucrose and 1M EGTA. The resultant pellet was re-suspended with above buffer using a 26G syringe and stored at -80°C until use. We adjusted CHAPSO resistant membrane fractions to 1 μg/μl in protein concentration with 20 mM PIPES, pH 7.0, 250 mM sucrose and 1 mM EGTA. Kinome profile characterization. 50 mg of temporal cortex from selected cases ( Table 3) was lysed at 0 °C using M-PER (Mammalian Protein Extraction Reagent, Thermo Scientific, MA, USA) lysis buffer (0.1 g/ml) containing Protease Inhibitor Cocktail (Roche, Manheim, Germany) and Phosphatase Inhibitor Cocktail (Roche, Manheim, Germany), and centrifuged at 10.000× g 10 min, 4 °C. Supernatants we snap frozen in 100μl aliquots and stored at -80 °C. The protein concentration was determined using the Bradford Lowry Assay (Pierce Coomasie assay, Thermo, Dreieich, Germany). Frozen aliquots were never re-frozen but used directly for kinase activity determination. Kinase activity profiles were determined using the PamChip ® 96 serine/threonine (STK) and protein tyrosine (PTK) peptide microarray system from PamGene International B.V. ('s-Hertogenbosch, The Netherlands) according to the instructions of the manufacturer, and as described previously (8). All PamChip ® 96 array plates used in this study came from the same production batch and all plates were run on the same PamStation instrument. For each assay, 0.5μg of protein was used. Arrays were incubated 30 cycles in blocking buffer and 60 cycles in reaction buffer containing ATP (final concentration 100 μM; Sigma-Aldrich, St. Louis, MO, USA). Arrays were washed and incubated for 60 min with a secondary antibody (polyclonal swine anti-rabbit Immunoglobulin/FITC). Images at 50 ms exposure time were captured every 10 min with an integrated CCD-based optical system in combination with Evolve software (version 1.5, PamGene International BV). After removal of the secondary antibody and a wash step, post-wash images were taken at different exposure times (20, 50, 100, and 200 ms). The PTK assay mixture contained the same kinase assay buffer, 100μM ATP and 0.01% BSA, supplemented with 4μl protein kinase (PK)-additive (PamGene International BV), 10 mMDithiothreitol (DTT, Fluka, Sigma-Aldrich, St. Louis, MO, USA) and fluorescein isothiocyanate (FITC) labeled anti-phosphotyrosine antibody (PamGene International BV, 's-Hertogenbosch, The Netherlands). For each PTK assay, 7.5μg of protein was used. Since a labeled antibody is present in the PTK assay mixture, peptide phosphorylation was monitored during the incubation with assay mixture, by taking images every 5 min at 50 ms exposure time, allowing real time recording of the reaction kinetics (onestep reaction). After washing of the array, fluorescence was detected at different exposure times (20, 50, 100, and 200 ms). The fluorescent signal intensity for each peptide was analyzed using BioNavigator 6.1 software (PamGene International BV,'s-Hertogenbosch, The Netherlands) a statistical analysis and visualization software tool with an App-based infrastructure (https://www.pamgene.com/en/bionavigator.htm). For signal quantification, the slope of the fluorescent signal versus exposure time was calculated in order to increase the dynamic range and to filter out time differences between plates. Saturated signals were excluded. Visual quality control was performed to exclude defective arrays from the analysis. A linear mixed-effects model that analyzed the signals of all peptides jointly while taking the correlation structure into account was used. Change of log (signals) over log (time) was calculated. The obtained STK and PTK median kinase signal intensities were analyzed for common effects (for all peptides) and peptide-specific plate, strip and array random effects. The measurement error was modeled using a peptide-specific variance component covariance matrix that allowed for heterogeneous variances among exposure time points. The upstream protein kinases able to phosphorylate residues in peptides on the PTK and STK arrays were identified in the Human Protein Reference Database (http://www.hprd.org) (9), in Phosphosite (http://www.phosphosite.org) and Reactome (http://www.reactome.org) (10). These kinases were projected on the kinase phylogenetic tree using the Kinome Render tool (http://bcb.med.usherbrooke.ca/kinomerender.php). When databases used different names to indicate a kinase, the kinase names were converted to those used in Kinome Render via their UniProtID. For kinases linked to multiple UniProtIDs, only the ID used in the Kinome Render tool was retained (8).
Chymotrypsin 20S proteasome activity assay. Chymotrypsin 20S proteasome activity was tested in temporal cortex from controls and PSEN1 E280A cases using the 20S proteasome activity assay kit APT280 ( pTau Seeding assay. Finally, 300 mg of frozen temporal cortices were homogenized in 1500 μL of PBS + protease inhibitor (Roche) in a 2 mL glass dounce homogenizer (30 up/down strokes on ice by hand). The homogenate was transferred to a 1.5 mL Eppendorf tube and centrifuged at 10,000 x g for 10 min at 4°C. The supernatant was collected and aliquoted to avoid excessive freeze/thaw cycles. A bicinchoninic acid assay (BCA, Thermo Scientific Pierce) was performed to determine total protein concentration following the manufacturer's protocol. The in vitro seeding assay has been previously described and widely characterized (18,19). Briefly, The Tau RD P301S FRET Biosensor (ATCC® CRL-3275™) cells stably expressing the repeat domain of Tau with the P301S mutation conjugated to either the cyan fluorescent protein (CFP) or the yellow fluorescent protein (YFP) (TauRD-P301S-CFP/YFP) were cultured at 37°C, 5% CO2 in DMEM, 10% v/v fetal bovine serum, 0.5% v/v penicillin/streptomycin. Cells were plated on Costar Black, clear bottom 96well plates (previously coated with 1:20 poly-D-lysine) at a density of 40,000 cells per well. Brain extracts (1 µg of total protein quantified by BCA per well) were then incubated with Lipofectamine 2000 (Invitrogen, final concentration 1% v/v) in opti-MEM (final volume 50 μL per well) for 10 min at room temperature before being added to the cells. Each condition was applied in triplicate or quadruplicate. After 24 h, Tau seeding was subsequently analysed using flow cytometry : Medium was removed and 50 μL trypsin 1x was added for 7 min at 37°C. Chilled DMEM + 10% fetal bovine serum (150 μL) was added to the trypsin and cells were transferred to 96-well U-bottom plates (Corning). Cells were pelleted at 500 x g, resuspended in freshly-made 2% v/v paraformaldehyde in PBS (Electron Microscopy Services) for 10 min at room temperature in the dark, and pelleted at 500 × g. Cells were resuspended in chilled PBS and run on the MACSQuant VYB (Miltenyi) flow cytometer. CFP and Forster resonance energy transfer (FRET) were both measured by exciting the cells using the 405 nm laser and reading fluorescence emission at the 405/50 nm and 525/50 nm filters, respectively. To quantify the FRET signal, a bivariate plot of FRET vs. the CFP donor was generated and cells that received control brain extract alone were used to identify the FRET-negative population. Using this gate, the integrated FRET density (IFD) value for each well was calculated by multiplying the percent of FRET-positive cells by the median fluorescence intensity of that FRET-positive population. 40,000 events per well were analysed. Data was analysed using the MACSQuantify software (Miltenyi).

Genetic and protein network analysis methods
Fourteen patients with PSEN1 E280A FAD placed at the extremes of the AoO distribution (Table 3) were included for whole-exome capture (WEC). DNA was extracted from brain tissue and genomic DNA was (v) variant calling; and (vi) assigning quality scores to variants as described elsewhere (11,12). Genotype files were processed in SVS 8.3.0. Samples with calls below Illumina®'s expected 99% SNVs call rates were excluded. Single nucleotide variants (SNVs) were excluded when (i) deviated from Hardy-Weinberg equilibrium with P< 2x10-7, (ii) the minimum genotype call rate was <90%, (iii) the number of alleles was one or more than two, and (iv) the MAF<1%. Genotype and allelic frequencies were estimated by maximum likelihood. Subsequently, a filtering phase including the identification of de novo SNVs; filtering of potentially pathogenic variants using SIFT PolyPhen-2 MutationTaster, Gerp++ and PhyloP; and filtering of damaging variants based on genes known to be associated with AD, was performed using information from dbSNP and the 1K Exome Project. De novo SNVs were defined according to the DNA-seq Analysis module in SVS 8.3.0. Potential relationships between AoO and SNVs were individually examined using one-way analysis of variance (ANOVA). P-values were obtained based on the F-statistic and corrected for multiple testing using the false discovery rate (FDR) and a method based on extremes-value theory, as explained elsewhere (13). Network analysis and pathway analysis was performed using NetworkAnalyst (14) webpage tools and Cytoscape software (15). Protein -protein interaction was assessed InnateDB (16) webpage tools.                 There were not significant differences between groups. B.Quantification of Ab 1-42 plaque load in temporal cortex in SAD patients (n=10) vs PSEN1 E280A FAD patients (n=23). There is not significant differences between groups. C. Densitometric analysis of sAPP, small Ab oligomers and Ab monomers. Semi denaturing electrophoresis of TBS soluble fractions from temporal cortex in SAD patients (n=10) vs PSEN1 E280A FAD patients (n=23) were blotted with 6E10 antibody and analyzed according to their band distribution in kDa. D. Densitometric analysis scatterplots for sAPP, small AβOs and Aβ monomers. There are not significant differences between groups for sAPP, both oligomers and monomers show differences between SAD and FAD patients (**=p≤0.01).      . Synaptophysin particle distribution according to size in PS1 E280A FAD cases grouped by age of onset. A. Particle count according to size in PS1 E280A FAD cases grouped by age of onset. B. Average Synaptophysin-positive particle size in PS1 E280A FAD cases grouped by age of onset. C. Bar graphs for the density of small Synaptophysin-positive particles in temporal cortices of EOFAD (n=5), AOFAD (n=5) and LOFAD (n=5) cases. AOFAD cases showed significantly higher particle density when compared with EOFAD cases. (* = p ≤ 0.05). D. Correlation analysis between Synaptophysin particles density and soluble pTau-S400 / Tau. E. Correlation analysis between Synaptophysin particles density and disease duration.       Gene  ETNK1  MDM1  SCAPER  11138217  TPST2  ZNF234  HKDC1  UBQLN1  DNAH11  MED16  TAF2  USH2A  CSN1S1  USP36  NME4  KCNJ12  NRG1  TEKT4P2  OVCH2  OVCH2  OVCH2  OR2T11  EPB41L4A  NAT2  GPR124  MATN2  MAST2  HLA-B  FMN1  TRAK2  ZAN  GOLGA6L2  SLC7A7  RNF213  USP40  DHRS4L1  ZSCAN3  OVCH2  SLC7A7  TMEM180  50643184  LINC00094  CYP2D7P  TRAPPC3  Uncommon Variants Common Variants Figure 16. Proof of principle for co-immunoprecipitation with anti-polyubiquitinated and anti-Tau antibodies in brain tissue. First, we tested the specificity of the pTau S400 antibody using temporal cortex total homogenate and dephosphorylating the membrane with Akaline phosphatase (AP) 1u/ug of protein during 1 hour at 37 ⁰C before incubation with primary and secondary antibodies. Afterwards, using temporal cortex TBS fractions from one average onset and one late onset FAD cases we tested immunoprecipitation and western blots using antibodies against total Tau and polyubiquitin as baits and blotting with total tau, pTau s400 and polyubiquitin. At the left side for each test we placed how standard western blot looks for each case. Figure 17. Co-immunoprecipitation using monoclonal polyubiquitin antibody as bait and immunoblots for total Tau, pTau-S400 and pTau-S422. A. Co-immunoprecipitation using monoclonal polyubiquitin antibody as bait and immunoblots for total Tau and pTau-S400 in TBS soluble fractions from temporal cortex of early and average AoO FAD (E-AOFAD, n = 5) and LOFAD (n = 7). B. Co-immunoprecipitation using monoclonal polyubiquitin antibody as bait and immunoblots for total Tau and pTau-S422 in TBS soluble fractions from temporal cortex of early and average AoO FAD (E-AOFAD, n = 5) and LOFAD (n = 7).