Proteomic and Unbiased Post-Translational Modification Profiling of Amyloid Plaques and Surrounding Tissue in a Transgenic Mouse Model of Alzheimer’s Disease

Abstract

Amyloid plaques are one of the hallmarks of Alzheimer’s disease (AD). The main constituent of amyloid plaques is amyloid-β peptides, but a complex interplay of other infiltrating proteins also co-localizes. We hypothesized that proteomic analysis could reveal differences between amyloid plaques and adjacent control tissue in the transgenic mouse model of AD (APPPS1-21) and in similar regions from non-transgenic littermates. Our microproteomic strategy included isolation of regions of interest by laser capture microdissection and analysis by liquid chromatography mass spectrometry-based label-free relative quantification. We consistently identified 183, 224, and 307 proteins from amyloid plaques, adjacent control and non-tg samples, respectively. Pathway analysis revealed 27 proteins that were significantly regulated when comparing amyloid plaques and corresponding adjacent control regions. We further elucidated that co-localized proteins were subjected to post-translational modifications and are the first to report 193 and 117 unique modifications associated to amyloid plaques and adjacent control extracts, respectively. The three most common modifications detected in proteins from the amyloid plaques were oxidation, deamidation, and pyroglutamylation. Together, our data provide novel information about the biological processes occurring within and around amyloid plaques in the APPPS1-21 mouse model of AD.

Keywords

Alzheimer’s disease amyloid plaque mass spectrometry microdissection pyroglutamate

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disorder and the most common cause of dementia [1, 2]. The underlying mechanisms of the disease is not fully understood although two pathological hallmarks are deposition of extracellular plaques, mainly composed of amyloid-β (Aβ) peptides [3, 4], and intracellular neurofibrillary tangles, mainly composed of hyperphosphorylated tau [5, 6]. Aβ peptides originate from sequential cleavage of the membrane-spanning amyloid-β protein precursor (AβPP) by different secretases. The widely accepted ‘amyloid cascade hypothesis’ postulates that deposition of Aβ into senile plaques (Aβ plaques) initiates a pathological cascade leading to AD [7 –11]. The Aβ deposition is thought to begin prior to tau-associated pathology and clinical symptoms of AD underlining the pathological relevance of Aβ proteoforms in AD [12]. The molecular composition of Aβ plaques encompasses a variety of different amyloid-binding and infiltrating proteins [13]. Identification of these co-localizing proteins are reasoned to contain valuable information about processes occurring within and around Aβ plaques. It has been shown that aggregated Aβ peptides contain several post-translational modifications (PTMs) such as oxidation- and pyroglu-modifications [14 –16]. However, it remains to be elucidated whether plaque-infiltrating and associating proteins also are subjected to the same degree of PTMs as aggregated Aβ peptides.

Mass spectrometry (MS)-based technologies have in recent years proven to be a valuable tool to achieve in-depth characterization of the proteomic profile of biological specimens and characterization of PTMs [17, 18]. A commonly used method for proteomic discovery is a bottom-up approach in combination with data-dependent acquisition (DDA) [19]. The method is label-free and enables large coverage of the proteome that presents several advantages for elucidating complex biological systems, including identification of proteins containing PTMs. Although detection and validation of PTMs in MS presents several challenges such as low stoichiometry, site specific assignment and poor stability, recent improvements of algorithms enables a strong PTM identification and quantitative analysis by integrating the database search with results from de novo based analysis [20]. Furthermore, a recent study has shown in-depth PTM identification from a DDA approach which supports the applicability to our study [21].

By using laser capture microdissection (LCM), targets of interest within tissue sections can be isolated and subjected to MS analysis providing localized information. A recent study has illustrated the capability of using LCM to capture different Aβ plaque profiles when staining with tetra- and heptameric formyl-thiophene acetic acid (q/h-FTAA), thioflavin-S, or Congo red [22]. Another example of this approach was applied to Aβ plaques from rapidly progressive AD and sporadic AD patients [23].

Transgenic (tg) mice, like the APPPS1-21 strain [24], are useful mechanistic models that are based on familial AD (FAD) mutations which result in an increased Aβ load and early-onset inherited AD. The APPPS1-21 model carries mutations in the APP gene (KM670/671NL, Swedish) and in the PSEN1 gene (L166P), leading to excessive production of Aβ [24]. Despite differences in the senile plaque composition when comparing mouse and human brain as described by others [25] and us [26] previously, the main response mechanism to Aβ accumulation which is of interest for our study here are considered very similar. For example, Aβ plaque deposition and subsequent activation of microglia and astrogliosis begins at approximately six weeks of age in the mouse cerebral cortex [24], demonstrating glial engagement that similarly is observed in AD. Furthermore, dendritic spine loss is observed approximately four weeks after plaque deposition [27], indicating that the tg mouse model also resembles part of the neurodegenerative aspect of AD. Based on these comparable features, it is reasoned that APPPS1-21 mice are relevant to study Aβ-related pathologies and model how abnormal Aβ production affects the brain.

We used a discovery-based proteomic strategy in combination with LCM to profile the proteome of Aβ plaques and closely surrounding regions extracted from the cerebral cortex of APPPS1-21 mice. We discovered several differentially regulated proteins in Aβ plaques compared to adjacent control tissue by using two independent search algorithms. Additionally, we are the first to elucidate the unique PTM profile of the Aβ plaque associated proteins compared to the surrounding tissue.

MATERIALS AND METHODS

Materials

Acetonitrile (ACN; ≥99%), tris(2-carboxyethyl)phosphine (TCEP), 2-chloroacetamide (CAA; ≥98%), and triethylamonium bicarbonate (TEAB) were purchased from Sigma Aldrich, Denmark. Formic acid (FA; ≥98%) was purchased from Fluka, Denmark, thioflavin-t from Merck Eurolab, Denmark and trifluoroacetic acid (TFA; ≥99%) from Fisher Scientific, Denmark. ProteaseMax™ was purchased from Promega, Denmark.

Animals

Six 12-month-old heterozygous female tg APPPS1-21 and corresponding non-transgenic (non-tg) littermates were used for the analysis (Charles River, Germany). The mice were group-housed (five/cage). Water and food were supplied ad libitum, clean cages were provided twice a week with enrichment and mice were kept on a 12 h/12 h light/dark schedule. Room humidity was 55% ±5%, and room temperature (RT) was 21±2°C. The animal experiments were performed in accordance with the European Communities Council Directive no. 86/609, the directives of the Danish National Committee on Animal Research Ethics, and Danish legislation on experimental animals (license no. 2014-15-0201-00339). Mice were anesthetized with avertin (i.p. administration of 2% solution; 1 ml/100 g) and perfused with KPBS/heparin. Brains were fresh frozen on dry ice and stored at -80°C until further use.

Tissue sectioning

The mouse brains were embedded with Tissue-Tek OCT (Sakura, Denmark) and left for temperature adaption at -16°C for 30 min. The samples were cut into 25-μm thick sagittal sections using a cryostat microtome (Leica CM3050 S, Denmark) and mounted on polyethylene naphthalate (PEN) membrane glass slides (Thermo Fisher, Denmark). Three sections from each hemisphere were applied to a PEN membrane glass slide and fixed with 70% EtOH and 99% EtOH for one minute in each solution. Next, they were dried at RT and stored at -80°C until further use. Adjacent sections from both non-tg littermates and APPPS1-21 mice were similarly attached to SuperFrost Plus (VWR, Denmark) glass slides for thioflavin-T and immunohistochemistry analysis. Here, sections used for thioflavin-T staining were similarly fixed in EtOH while sections used for immunohistochemistry were not.

Thioflavin-T staining and immunohistochemistry

Brain sections from both APPPS1-21 and non-tg littermates were stained in 1% thioflavin-T solution for 10 min. The thioflavin-t solution was removed by rinsing the slides in ddH₂O three times for 3 min each. The tissue was air dried at RT and mounted in pertex (Histolab, Denmark). Adjacent sections from the APPPS1-21 mice were immunolabeled using the Vectastain Elite ABC HPR kit (Vector Laboratories Inc., Denmark). In brief, the fresh frozen tissue sections were fixed in 4% PFA for 10 min and pretreated with 1% hydrogen peroxide (H₂O₂) solution to block endogenous peroxidase activity and 90% formic acid solution for antigen retrieval. Washes in KPBS was performed in between. Sections were permeabilized in blocking buffer (1% bovine serum albumin (BSA), 0.3% triton X-100, 5% normal swine serum in KPBS) followed by over-night (O/N) staining with the anti-Aβ antibody (mouse, 6E10; 1 : 500; BioLegend) at 4°C. Tissue was counterstained with biotinylated anti-mouse biotinylated antibodies (DAKO E0354; 1 : 500), followed by incubation in the avidin-biotin complex (ABC) solution (Vector Laboratories Inc., Denmark). Sections were washed in washing buffer (0.25% BSA, 0.1% triton X-100 in KPBS) in between. The immunoreactivity was visualized with the chromogen, diaminobenzidine (DAB, Sigma, Denmark). Imaging was performed on a Zeiss Axio Scope.A1 equipped with a Zeiss Axiocam 506 color camera.

LCM of mouse brain tissue sections

The PEN membrane glass slides containing tissue sections were defrosted by RT adjustment and a dip wash in ddH₂O prior to staining with a 1% thioflavin-T solution, as stated in the previous section (except mounting in pertex). Three types of regions (Aβ plaques and adjacent control regions from the APPPS1-21 mice, and similar control areas from the non-tg littermates) were microdissected and collected onto CapSure™ Macro LCM caps (Thermo Fisher, Denmark) using the Arcturus Veritas LCM system. A 20X magnification was utilized and 50 regions were extracted from each section with an average area of 5791.2 μm² per extracted plaque, 6449.4 μm² per extracted adjacent control, and 5560.7 μm² per extracted non-tg littermate regions. The microdissection focused on core plaques, and not diffuse plaques, because the latter are mainly thioflavin-T negative. The caps containing the microdissected material were incubated in 50 μl digestion buffer (0.01% ProteaseMax™ in 50 mM TEAB buffer, pH 8.5) O/N at 37°C. 10 μl of fresh digestion buffer was manually resuspended five times on the cap surface to ensure release of potential remaining material (total volume = 100 μl) and collected in 1.5 ml Eppendorf LoBind microcentrifuge tubes (Eppendorf). An in-solution digest protocol was utilized for each tube. In brief, the tubes were sonicated for 2 min, heat treated at 95°C for 5 min and allowed to cool to RT. The samples were reduced and alkylated in 1 : 50 TCEP (v:v; 0.5M) and 1 : 10 CAA (v:v; 0.5M) and incubated in darkness for 30 min at 37°C. One μl trypsin (0.1 μg/μl) was added and incubated O/N at 37°C. TFA (final concentration = 1%) was added and incubated for 30 min at 37°C to stop the reaction of ProteaseMax™ and trypsin. The tubes were inserted into a SpeedVac (miVac, GeneVac), dried for 6 h and stored at -80°C until analysis. The dried material was reconstituted in 12 μl loading buffer (2% ACN, 0.1% TFA, 0.1% FA) prior to MS analysis.

Data acquisition by LC-MS/MS

A pooled sample was created for each subregional group by adding two μl from each reconstituted sample into one 0.5 ml tube (LoBind, Eppendorf). These samples were utilized to create two pooled higher intensity samples per subregion in the data analysis. Nine μl sample material was added to 96-well plate with a random order. Samples were analyzed on quadrupole Orbitrap MS (Q-Exactive HF-X, Thermo Scientific) connected to an RSLC UHPLC nanopump system (Dionex). Runs of each sample were loaded onto the C18 reversed phase column (Dionex; Acclaim PepMap100 C18, 5 μm precolumn and 75 cm Acclaim Pepmap RSLC, 75 μm ID main column, Thermo Scientific) and eluted with a linear gradient of 96% solvent A (1% formic acid) and 4% solvent B (ACN) [28], increasing solvent B to 35% on a 30-min ramp gradient. The MS was operated in DDA, selecting the 10 precursor-ions with the highest intensity for higher energy collisional dissociation (HCD) fragmentation. Full MS settings were as follows: resolutions = 60 000; AGC target = 3e6 (max 60 ms); scan range 375 to 1500 m/z. The following dd-MS² settings were used: resolution = 30 000; AGC target = 1e5 (max 120 ms); loop count = 10; isolation window = 1.5 m/z; scan range 200 – 2000 m/z. The maximum IT was set to 120 ms for all samples except the two pooled higher intensity samples which were acquired at 250 ms and 500 ms, respectively (Supplementary Figure 1). The following dd settings were used: underfill ratio = 30.0%; charge exclusion unassigned, 1, 7, 8,>8; peptide match = preferred; exclude isotopes = on. Intensity threshold was set to 2.5e5 for all samples except the two pooled higher intensity samples which were acquired at 1.2e5 or 6.0e4. Dynamic exclusion was set to 10.0 s for all samples except the two pooled higher intensity samples which were either excluded or set 5.0 s, respectively. Data are available via ProteomeXchange with identifier PXD014274.

Data processing

Datafiles were processed in MaxQuant (v1.6.5.0) [29]. The FASTA file, containing the UniProt mus musculus database (UP_000000589), was downloaded 4 September 2017 and contained 16.911 entries. The human amyloid beta A4 protein/amyloid precursor protein (APP) and presenilin 1 (PSEN1) sequences were added due to the co-expression of these proteins in the APPPS1-21 mouse model while the mouse amyloid beta A4 protein/amyloid precursor protein was removed. The Swedish mutation (KM670/671NL) was not included due to its location prior to the Aβ1-42 peptide sequence. Bioinformatic and statistical analysis was performed in Perseus (v1.6.5.0) [30]. Potential contaminants and reverse sequences were removed followed by filtering the dataset by minimum 50% valid values and presence of > 1 unique peptide. Hierarchical clustering was applied together with Pearson correlation analysis to identify potential outliers. A sample was defined as an outlier when two criteria were met; 1) the average Pearson correlation coefficients (<0.8, compared to the corresponding samples in the group (Aβ plaque versus Aβ plaque, adjacent control versus adjacent control, and non-tg versus non-tg), and 2) a median absolute deviation (MAD) analysis (if the value was five times away from the median it was defined as an outlier). The normal distributed data (histograms not shown) was analyzed with two-sampled test (two-sided student’s t-test; minimum valid value filter = 70%; Permutation-based FDR = 0.05). The same MaxQuant settings, except the match-between-run function, were used to elucidate the consequence of excluding this feature. STRING (v11.0) enrichment analysis was used for pathway analysis [31]. The five Biological Process (GO)-associations with lowest FDR were selected for both up- and downregulated proteins. Variation of the second component in the PCA plot was analyzed with student t-test and Fisher exact test using the Gene Ontology Biological Process (GOBP) class.

A parallel data analysis was performed in BSI PEAKS Studio X (v10; Bioinformatics Solution; Canada). All data samples were imported, analyzed with PEAKS Denovo, PEAKS DB search, PEAKS PTM and PEAKS SPIDER. The following settings were used: parent mass error tolerance 10.00 ppm; fragment mass error tolerance = 0.02 Da; fragment mass search type monoisotopic; trypsin with [D, P]; max missed cleavages 3; semi specific digestion mode; fixed modification: Carbamidomethylation (57.02); Variable modification: citrullination (0.98), oxidation (M; 15.99), phosphorylation (STY; 79.97); max variable PTM per peptide = 3. The database that was used for the MaxQuant previously was similarly used for the PEAKS analysis. The following settings were applied for visualization: FDR = 0.1%; PTM ion intensity≥1%; Proteins = 10logP≥20; unique peptides≥1; De novo ALC≥50%; -10logP≤18.4. Similar settings were applied to PEAKS PTM analysis although the variable modifications now included 312 different modifications.

The label-free quantification (LFQ) analysis of Aβ plaques and adjacent control was performed by using the PEAKS DB and PEAKS PTM searches as references (using 0.1% FDR; significance≥20; fold change≥1; significance method = PEAKSQ; At least 1 unique peptide). LFQ modification outputs were processed using R as reported in [21] to identify unique modifications. In brief, LFQ modification outputs (with minimum a-score of 100) were exported and log2 transformed followed by calculation of the LFQ ratio between the modified and unmodified peptides. A non-zero quantitation value was required in more than half of the samples for inclusion.

RESULTS

Proteomic fingerprint of Aβ plaques in APPPS1-21 model mice

We investigated the protein composition of Aβ plaques from 12-month-old APPPS1-21 mice by a combination of tissue microdissection and label-free LC-MS/MS quantification. APPPS1-21 mice overexpress human APP and PS1 with FAD mutations leading to overexpression of human Aβ species and rapid progression of Aβ plaques in the mouse brain [24]. Tissue was dissected from the cerebral cortex region containing a high Aβ plaques load in the 12-month old APPPS1-21 mice (Fig. 1A). Aβ plaques were identified by thioflavin-T staining exclusively in brains from APPPS1-21 and not in brains from non-tg littermates (Supplementary Figure 2). A 20X objective lens was used for microdissection to ensure a true-plaque morphology and avoid inclusion of potential contaminant caused by the fluorescent background. Additionally, we validated the thioflavin-T staining by orthogonal detection of Aβ plaques by immunohistochemistry with the anti-Aβ 6E10 antibody. To this end, we stained adjacent brain sections with either thioflavin-T or 6E10 and observed that the same plaques in the adjacent sections were detected by both thioflavin-T and 6E10 (Supplementary Figure 2).

Fig.1

A) Experimental workflow. Brain tissue sections from right and left hemisphere were collected from six transgenic (tg) APPPS1-21 and six non-tg mice (total = 12 samples per genotype). Laser microdissection was used to extract regions of interest (50 per brain slice; three slices per hemisphere); amyloid plaques (Aβ plaque, green), adjacent control (blue), or non-tg (red), respectively. Extracted proteins were digested, analyzed by liquid chromatography-tandem mass spectrometry, and processed by bioinformatic analysis. B) Principal component analysis plot of LFQ intensities associated with samples (after Pearson correlation correction and median absolute deviation analysis). Samples from both right and left hemispheres are presented (n = 6 mice; samples from both hemispheres). Components 1 and 2 are presented. C) Venn-diagram of the number of proteins observed in different subgroups; Aβ plaque extracted material (green), adjacent control (blue), corresponding regions from non-tg littermate mice (red). D) Volcano plot of differentiating log2 transformed proteins when comparing Aβ plaques and adjacent control regions. The location of amyloid beta A4 protein (App), apolipoprotein E (Apoe), and clusterin (Clu) are visualized.

We microdissected two different regions from the APPPS1-21 mice either containing Aβ plaques or without Aβ plaques of an adjacent control region (Fig. 1A). From non-tg littermates, we microdissected corresponding control regions. Female mice were used in this study because they have been reported to exhibit a more robust Aβ deposition than male APPPS1-21 mice [32].

For this study, we implemented a detergent enhanced solubilization strategy using low volume acid-labile surfactant, ProteaseMax™, to increase solubility without use of detergents that interferes with the MS analysis. Due to the limited sample material in LCM plaque extracts, we applied a match-between-run strategy matching ID to a pooled higher intensity sample to increase the protein identification rate [29]. This approach included the generation of mixed samples from either of the three subgroups. By adjusting the MS settings for these samples, we could affect the number of full- and MS2 scans and thereby establish spectral references that almost doubled the protein identification (Supplementary Figure 1A-C). The strategy was supported by sequential sample analysis with the same chromatographic method. Further, sample concentrations were in the nanogram range, meaning that high loading effect was not an issue as it was significantly lower than the column capacity. The applied software uses normalization in the LFQ algorithm for label-free quantification ensuring a reliable comparison between protein levels [33].

Investigation of the three different microdissected regions by principal component analysis (PCA) implied a difference between the three respective subgroups: non-tg, adjacent control, and Aβ plaques from APPPS1-21. We included samples from both right and left hemisphere from each individual animal and used 6 animals per genotype (Fig. 1B; n = 6 mice; samples from both hemispheres; total = 12 data points). One Aβ plaque sample associated to the left hemisphere of one mouse was lost during the MS analysis. Additionally, the MS analysis of another APPPS1-21 mouse (adjacent control and plaque extract from same hemisphere) did not meet the test criteria (Person correlation minimum value and median absolute deviation) whereby it was defined as an outlier and excluded from the further analysis. Therefore, we have 10 datapoints from the Aβ plaque group (green circles in the PCA plot) and 11 datapoints from the adjacent control group (blue squares in the PCA plot). A partial overlap was observed on the first component between non-tg control and adjacent control regions from APPPS1-21, while no overlap was observed between Aβ plaques from APPPS1-21 and non-tg samples. This separation between non-tg and APPPS1-21 plaque samples on the first component was anticipated. Additionally, we observed a separation on the second component in top and bottom location, which was not dependent on the genotype by affecting samples from both non-tg and APPPS1-21 mice equally. Interestingly, both samples from the left and right hemispheres from each individual animal (except for one non-tg sample) were either in the top or in the bottom location for the second component, indicating that this separation was associated to the individual animals (Supplementary Figure 3). We identified significant protein differences in pathways associated to “response to stimulus” and “metabolic processes” by Gene Ontology Biological Process (GOBP) class between the top and bottom samples in all subgroups (data not shown). Based on these findings, we suggest that the observed separation on the second component was dependent on the individual state (e.g., hormonal and stress related factors) of the mice prior to the MS experiment and consider this separation as normal variation detected within mouse cohorts. Our further pathway analysis focused on tissue-associated differences between the Aβ plaque and adjacent control regions from APPPS1-21.

A total of 555 proteins were detected in extracted Aβ plaque regions from APPPS1-21 mice, following a database search against the mouse reference proteome. Of these, 183 proteins were consistently observed across Aβ plaque samples. In the adjacent control and non-tg control regions, 224 and 307 proteins were identified consistently, respectively.

A total of 427 proteins were shared within all groups (Fig. 1 C, middle of circles). Fourteen proteins, Pafah1b2, Arfgef3, Camk2d, Spock2, Ncam2, Slc1a4, Pgrmc1, Sept2, Apoa1bp, Uqcr10, Lamtor1, Clstn1, Hapln1, and Twf2, were unique to Aβ plaques. Four proteins, Atp2b4, App, Apoe, and Vtn, were unique to the Aβ plaque and adjacent control regions (Fig. 1C; overlapping area between the green and blue circle). Eleven proteins were unique to the adjacent control regions: Cmas, Rpn2, Dlg3, Baiap2, Mtnd5, P4hb, Prdx3, Inpp1, Psmc5, Rplp2, and Prkca.

The inclusion of the adjacent control regions allowed a subsequent identification of relative proteomic differences. By comparing adjacent control and Aβ plaques, we identified App, Apoe, and Clu as the top three regulated proteins illustrated by the volcano plot (Fig. 1D). Two-sided t-test analysis identified 27 proteins that were significantly different when comparing the Aβ plaque proteome data to the adjacent control samples from APPPS1-21 (Tables 1 and 2). The differentiating proteins are listed in (Tables 1 and 2). The regulated proteins (both up and down) contained between 2 and 83 unique peptides (Tables 1 and 2).

Table 1

Label-free quantification (LFQ) analysis of Aβ plaque extracts compared to adjacent control regions (Upregulated proteins identified by MaxQuant and Perseus analysis; two-sided student’s t-test; minimum valid value filter = 70%; Permutation-based FDR = 0.05; p-value cut-off≤0.05). (*)=Proteins identified as significantly regulated using PEAKS Quantification analysis

Protein	Gene	Fold change	p	Unique peptides	Sequence coverage [%]
Alpha-centractin	Actr1a	1.48	3.66E-04	4	27.7
Alpha-enolase*	Eno1	1.68	2.04E-03	9	50.5
Amyloid beta A4 protein*	App	7.17	3.87E-08	10	21
Apolipoprotein E*	Apoe	14.26	4.20E-07	8	39.9
ATP-dependent 6-phosphofructokinase	Pfkl	1.53	1.35E-03	4	8.1
Clusterin*	Clu	6.51	9.21E-07	10	25.2
Creatine kinase B-type	Ckb	1.71	1.15E-04	18	62.7
Ezrin;Moesin;Radixin	Ezr;Rdx;Msn	2.18	5.65E-06	6	14
Glyceraldehyde-3-phosphate dehydrogenase	Gapdh	1.56	1.12E-03	16	63.8
Glypican-1	Gpc1	4.20	4.64E-03	4	12.6
Neurexin-1	Nrxn1	5.14	8.17E-05	6	8.2
Phosphoglycerate kinase 1	Pgk1	1.52	2.18E-03	12	41.2
Phosphoglycerate mutase 1	Pgam1	1.70	1.22E-03	10	59.8
Phospholipid-transporting ATPase	Atp9a	3.40	1.05E-05	3	5.3
Protein RUFY3	Rufy3	2.13	3.24E-03	3	5.2
Ras-related protein Rab-1A	Rab1	1.36	1.57E-03	5	40.6
Ras-related protein Rab-2A*	Rab2a	1.65	1.36E-03	2	33.5
Ras-related protein Rab-7a*	Rab7a	2.09	5.04E-06	9	46.4
TOM1-like protein 2	Tom1l2	1.66	7.62E-04	5	14.4
Triosephosphate isomerase	Tpi1	1.54	5.01E-04	7	28.8
Polyubiquitin-B	Ubb	2.41	4.67E-03	4	58.4
Vitronectin	Vtn	9.43	3.62E-05	5	20.7

Table 2

Label-free quantification (LFQ) analysis of Aβ plaque extracts compared to adjacent control regions (Downregulated proteins identified by MaxQuant and Perseus analysis; two-sided student’s t-test; minimum valid value filter = 70%; Permutation-based FDR = 0.05; p-value cut-off≤0.05). (*)=Proteins identified as significantly regulated using PEAKS Quantification analysis

Protein	Gene	Fold change	p	Unique peptides	Sequence coverage [%]
Matrin-3*	Matr3	0.46	1.31E-03	7	12.2
Microtubule-associated protein 1B*	Map1b	0.65	3.17E-03	13	11.2
Microtubule-associated protein 2*	Map2	0.62	1.56E-03	34	28.8
Solute carrier family 12 member 5	Slc12a5	0.52	6.66E-04	10	12.9
Spectrin alpha chain*	Sptan1	0.71	9.92E-04	83	41.7

To support our protein identification analysis, we performed a parallel analysis with an de novo and tandem MS based identification algorithm, including all sample extracts. The approach identified 2,032 proteins when analyzing the three subgroups together. Furthermore, the additional unbiased PTM module identified 1,397 proteins that contained PTMs (Ascore≥20; FDR = 0.1%; unique peptides≥1; Supplementary Table 1). Moreover, the analysis enabled a ratio determination of truncated Aβ peptides between Aβ plaques and adjacent control regions (Supplementary Figure 4). We found that truncated Aβ species, such as Aβ_29 - 40 oxidized (Ox), Aβ_29 - 42, Aβ_29 - 42 Ox, Aβ_29 - 43, and Aβ_29 - 43 Ox, were abundant within Aβ plaques and low in adjacent control extracts, indicating that true-plaque content was microdissected in the Aβ plaque group.

The second LFQ analysis identified 95 significantly regulated proteins when comparing the Aβ plaque and adjacent control regions (Supplementary Table 2). Ten proteins were among the previously 27 significantly regulated proteins identified by the MaxQuant and Perseus analysis and thus supported their presence (Indicated by (*) in Tables 1 and 2). The proteins identified by both search algorithms were Eno1, App, Apoe, Clu, Rab2a, Rab7a, Matr3, Map1b, Map2, and Sptan1. The App, Clu, and Apoe proteins were among the most regulated proteins in both LFQ analysis. App, Apoe, and Clu have previously been identified as susceptible genes for AD in different genome-wide association study (GWAS) [34] and identification of these three top upregulated proteins illustrated the validity of our approach in APPPS1-21 mice.

Regulated plaque-proteins are associated to different biological processes

We investigated a possible relationship between the up- and downregulated proteins by using literature based biological process (GO) enrichment analysis of known protein interactions [31]. In general, we observed that up- and downregulated proteins were associated to different biological processes (Table 3).

Table 3

Enrichment analysis association of significant differentiating proteins identified by label-free quantification (LFQ) in MaxQuant and Perseus. FDR = false discovery rate

Upregulated	Observed count/background count	FDR	Downregulated	Observed count/background count	FDR
Glycolytic process: Eno1, Gapdh, Pfkl, Pgam1, Tpi1	5/37	8.05E-07	Negative regulation of protein depolymerization: Map1b, Map2, Sptan1	3/63	7.19E-05
Purine ribonucleoside monophosphate biosynthetic process: Eno1, Gapdh, Pfkl, Pgam1, Tpi1	5/91	2.66E-06	Dendrite development: Map1b, Map2, Slc12a5	3/113	7.94E-05
Regulation of amyloid fibril formation: Apoe, App, Clu	3/7	5.49E-06	Negative regulation of supramolecular fiber organization: Map1b, Map2, Sptan1	3/122	8.29E-05
Endocytosis: Apoe, App, Atp9a, Ezr, Rab1, Rab7, Vtn	7/424	6.85E-06	Negative regulation of cytoskeleton organization: Map1b, Map2, Sptan1	3/130	8.56E-05
Vesicle-mediated transport: Apoe, App, Atp9a, Ezr, Nrxn1, Rab1, Rab2a, Rab7, Vtn	9/1019	1.31E-05	Negative regulation of microtubule depolymerization: Map1b, Map2	2/28	6.3E-4

The upregulated proteins were associated to the following biological processes: glycolytic process (n = 5), purine ribonucleoside monophosphate biosynthetic process (n = 5), regulation of amyloid fibril formation (n = 3), endocytosis (n = 7), and vesicle-mediated transport (n = 9). Several of these proteins (e.g., Rab1a, Rab2a, Rab7a, App, Apoe, and Vtn) are hypothesized to be involved in endocytosis and/or phagocytosis and thus indicated degradational processes [35 –37].

The downregulated proteins were associated to: negative regulation of protein depolymerization (n = 3), dendrite development (n = 3), negative regulation of supramolecular fiber organization (n = 3), negative regulation of cytoskeleton organization (n = 3), and negative regulation of microtubule depolymerization (n = 2). Both proteins, Map1b and Map2, were represented in each of the associations. Map1b and Map2 are constituents of microtubules, suggesting that part of the downregulated proteins was associated with the cytoskeleton of neurons.

Altogether our analysis therefore indicated a functional difference between the up- and downregulated Aβ plaque proteins where the upregulated proteins were associated to endocytosis and glycolysis and downregulated proteins associated to neuronal cytoskeleton.

Proteins co-localizing with Aβ plaques are subjected to unique modifications

PTMs can alter protein structures and thus have an impact on the biological function. This has previously been observed to occur on aggregated Aβ peptides in AD; however, it remains unclear whether co-localizing proteins are subjected to similar degree of modifications. We are the first to report detection of unique PTMs on proteins co-localizing with Aβ plaque extracts isolated from APPPS1-21 mice. Investigation of the modification profile of detected proteins in the Aβ plaques and adjacent control extracts showed 426 shared PTMs (Fig. 2A). Further, there was no clear discriminative difference when depicting the shared PTM data in a PCA plot (Fig. 2B). Interestingly, a total of 193 and 117 modifications were uniquely identified on proteins co-localizing with Aβ plaques and the adjacent control regions, respectively (Ascore > 100 and minimum representation in half of the subgroup; Fig. 2A, C). The identified unique modifications of co-localizing proteins in Aβ plaque extracts were: 2-OH-ethyl thio-Ser, 3-sulfobenzoic succinimidyl ester, acetylation (N-term; Protein N-term), amidation, amidination of lysines or N-terminal amines with methyl acetimidate, aminoethylcysteine, ammonia-loss (N), biotinylation, carbamylation, carboxylation (DKW), carboxymethyl (KW X@N-term), citrullination, crotonaldehyde, deamidation (NQ), deoxy, demethylation (KR), ethanolation (KR), formylation, ISD (z + 2)-series, methionine replacement by azido homoalanine, methylation (KR), methylpyrroline, oxidation (HW), oxidation (M), proline oxidation to pyroglutamic acid, proline oxidation to pyrrolidinone, pyro-glu from Q, replacement of proton with ammonium ion, sulphone, tryptophan oxidation to kynurenin, tryptophan oxidation to oxolactone (Fig. 2 C; Table 4). The three most common unique modifications detected in proteins from the Aβ plaques extracts were oxidation (n = 56), deamidation (n = 26), and pyro-glu (n = 25).

Fig.2

A) Venn-diagram of detected modifications in amyloid-β (Aβ) plaque (green) and adjacent control (blue) extracted regions. B) Principal component analysis plot of the shared modification profile (LFQ intensities) on proteins associated with Aβ plaques (green) and adjacent control (blue) regions. Components 1 and 2 are presented. C) Heatmap representation of unique modifications from Aβ plaque (n = 193; column 1) and adjacent control material (n = 117; column 2). The detected modifications are listed to the left of the heatmap. The number of unique modifications is listed in each cell and illustrated with a rainbow color scale (purple = 0; red = 60). Non-biological modifications, such as ISD (z + 2)-series, were also included to depict all modification identified by the analysis.

Table 4

Identified modifications on proteins co-localizing with Aβ plaques extracts

Protein accession	Gene	Protein name	Modification	Site
Q61781	Krt14	Keratin, type I cytoskeletal 14	2-OH-ethyl thio-Ser	S247
Q9DBG3	Ap2b1	AP-2 complex subunit beta	3-Sulfobenzoic succinimidyl ester	P286
Q04447	Ckb	Creatine kinase B-type	3-Sulfobenzoic succinimidyl ester	H29
P48962	Slc25a4	ADP/ATP translocase 1	Acetylation (N-term)	A285
P0DP27	Calm2	Calmodulin-2	Acetylation (N-term)	H108
P23116	Eif3a	Eukaryotic translation initiation factor 3 subunit A	Acetylation (N-term)	S64
Q9Z331	Krt6b	Keratin, type II cytoskeletal 6B	Acetylation (N-term)	I298
Q3UV11	Krt6b	Keratin, type II cytoskeletal 6B	Acetylation (N-term)	I290
A0A1B0GSX0	Ldha	L-lactate dehydrogenase	Acetylation (N-term)	A31
Q564E2	Ldha	L-lactate dehydrogenase	Acetylation (N-term)	A2
P06151	Ldha	L-lactate dehydrogenase A chain	Acetylation (N-term)	A2
P68373	Tuba1c	Tubulin alpha-1C chain	Acetylation (N-term)	A65
P56135	Atp5mf	ATP synthase subunit f, mitochondrial	Acetylation (Protein N-term)	A2
Q68FD5	Cltc	Clathrin heavy chain 1	Acetylation (Protein N-term)	A2
P17742	Ppia	Peptidyl-prolyl cis-trans isomerase A	Acetylation (Protein N-term)	V2
P51863	Atp6v0d1	V-type proton ATPase subunit d 1	Acetylation (Protein N-term)	S2
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Amidation	G93
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Amidation	S95
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Amidation	G93
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Amidation	S95
P99024	Tubb5	Tubulin beta-5 chain	Amidation	G93
P99024	Tubb5	Tubulin beta-5 chain	Amidation	S95
Q922F4	Tubb6	Tubulin beta-6 chain	Amidation	F90
P56480	Atp5f1b	ATP synthase subunit beta, mitochondrial	Amidination of lysines or N-terminal amines with methyl acetimidate	E172
Q9QYG0	Ndrg2	Protein NDRG2	Aminoethylcysteine	T10
P05067	APP	Amyloid-beta precursor protein	Ammonia-loss (N)	N698
Q03265	Atp5f1a	ATP synthase subunit alpha, mitochondrial	Biotinylation	Y337
P20029	Hspa5	Endoplasmic reticulum chaperone BiP	Biotinylation	N201
P63017	Hspa8	Heat shock cognate 71 kDa protein	Biotinylation	N174
P17156	Hspa2	Heat shock-related 70 kDa protein 2	Biotinylation	N175
P50446	Krt6a	Keratin, type II cytoskeletal 6A	Carbamylation	G414
Q9Z331	Krt6b	Keratin, type II cytoskeletal 6B	Carbamylation	G422
Q3UV11	Krt6b	Keratin, type II cytoskeletal 6B	Carbamylation	G414
P06745	Gpi	Glucose-6-phosphate isomerase	Carboxylation (DKW)	K438
E9Q0F0	Krt78	Keratin 78	Carboxylation (DKW)	D121
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Carboxymethyl (KW X@N-term)	S78
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Carboxymethyl (KW X@N-term)	S78
P99024	Tubb5	Tubulin beta-5 chain	Carboxymethyl (KW X@N-term)	S78
P07146	Prss2	Anionic trypsin-2	Citrullination	R122
P56480	Atp5f1b	ATP synthase subunit beta, mitochondrial	Citrullination	R387
Q68FD5	Cltc	Clathrin heavy chain 1	Citrullination	R437
Q9D0M3	Cyc1	Cytochrome c1, heme protein, mitochondrial	Citrullination	R228
E9PZ00	Psap	Prosaposin	Citrullination	R350
K3W4L3	Psap	Prosaposin	Citrullination	R355
Q8BFQ1	Psap	Prosaposin	Citrullination	R353
Q61207	Psap	Prosaposin	Citrullination	R356
A0A0A0MQA5	Tuba4a	Tubulin alpha chain	Citrullination	R293
P68373	Tuba1c	Tubulin alpha-1C chain	Citrullination	R264
P05214	Tuba3a	Tubulin alpha-3 chain	Citrullination	R264
Q9JJZ2	Tuba8	Tubulin alpha-8 chain	Citrullination	R422
P02535	Krt10	Keratin, type I cytoskeletal 10	Crotonaldehyde	K205
P63101	Ywhaz	14-3-3 protein zeta/delta	Deamidation (NQ)	Q219
P56480	Atp5f1b	ATP synthase subunit beta, mitochondrial	Deamidation (NQ)	Q101
Q8R5C5	Actr1b	Beta-centractin	Deamidation (NQ)	Q264
Q9EPL2	Clstn1	Calsyntenin-1	Deamidation (NQ)	Q300
O08553	Dpysl2	Dihydropyrimidinase-related protein 2	Deamidation (NQ)	Q122
P20029	Hspa5	Endoplasmic reticulum chaperone BiP	Deamidation (NQ)	N507
P56564	Slc1a3	Excitatory amino acid transporter 1	Deamidation (NQ)	N483
A2AE89	Gstm1	Glutathione S-transferase Mu 1	Deamidation (NQ)	Q192
E9Q197	Glod4	Glyoxalase domain-containing protein 4	Deamidation (NQ)	Q241
P21279	Gnaq	Guanine nucleotide-binding protein G(q) subunit alpha	Deamidation (NQ)	Q197
P63017	Hspa8	Heat shock cognate 71 kDa protein	Deamidation (NQ)	N194
P08730	Krt13	Keratin, type I cytoskeletal 13	Deamidation (NQ)	N201
Q6IFX2	Krt42	Keratin, type I cytoskeletal 42	Deamidation (NQ)	Q217
Q61735	Cd47	Leukocyte surface antigen CD47	Deamidation (NQ)	N93
A0A0U1RPX4	Mapk3	Mitogen-activated protein kinase	Deamidation (NQ)	Q288
P20917	Mag	Myelin-associated glycoprotein	Deamidation (NQ)	Q198
E9PUL5	Prrt2	Proline-rich transmembrane protein 2	Deamidation (NQ)	Q36
E9PUL5	Prrt2	Proline-rich transmembrane protein 2	Deamidation (NQ)	Q44
Q80TL0	Ppm1e	Protein phosphatase 1E	Deamidation (NQ)	N439
G5E8R3	Pcx	Pyruvate carboxylase	Deamidation (NQ)	N185
P50396	Gdi1	Rab GDP dissociation inhibitor alpha	Deamidation (NQ)	Q236
Q6PIC6	Atp1a3	Sodium/potassium-transporting ATPase subunit alpha-3	Deamidation (NQ)	N628
Q6PIC6	Atp1a3	Sodium/potassium-transporting ATPase subunit alpha-3	Deamidation (NQ)	N710
P14094	Atp1b1	Sodium/potassium-transporting ATPase subunit beta-1	Deamidation (NQ)	N104
O08599	Stxbp1	Syntaxin-binding protein 1	Deamidation (NQ)	Q250
Q9ERD7	Tubb3	Tubulin beta-3 chain	Deamidation (NQ)	N226
P17751	Tpi1	Triosephosphate isomerase	Deoxy	S165
P62631	Eef1a2	Elongation factor 1-alpha 2	Dimethylation (KR)	K55
P50446	Krt6a	Keratin, type II cytoskeletal 6A	Ethanolation (KR)	R84
E9Q0F0	Krt78	Keratin 78	Formylation	K122
P02535	Krt10	Keratin, type I cytoskeletal 10	Formylation	T369
Q9QWL7	Krt17	Keratin, type I cytoskeletal 17	Formylation	I145
Q61171	Prdx2	Peroxiredoxin-2	Formylation	K109
Q9CQV8	Ywhab	14-3-3 protein beta/alpha	ISD (z + 2)-series	K214
P62259	Ywhae	14-3-3 protein epsilon	ISD (z + 2)-series	K215
P68510	Ywhah	14-3-3 protein eta	ISD (z + 2)-series	K217
P61982	Ywhag	14-3-3 protein gamma	ISD (z + 2)-series	K217
O70456	Sfn	14-3-3 protein sigma	ISD (z + 2)-series	K214
P63101	Ywhaz	14-3-3 protein zeta/delta	ISD (z + 2)-series	K212
P05067	APP	Amyloid-beta precursor protein	ISD (z + 2)-series	L688
Q8BFZ3	Actbl2	Beta-actin-like protein 2	ISD (z + 2)-series	L172
P43006	Slc1a2	Excitatory amino acid transporter 2	ISD (z + 2)-series	F487
A0A0A0MQA5	Tuba4a	Tubulin alpha chain	ISD (z + 2)-series	F296
P68373	Tuba1c	Tubulin alpha-1C chain	ISD (z + 2)-series	F267
P05214	Tuba3a	Tubulin alpha-3 chain	ISD (z + 2)-series	F267
Q9CPN9	2210010C04Rik	RIKEN cDNA 2210010C04 gene	Methionine replacement by azido homoalanine	M110
P63017	Hspa8	Heat shock cognate 71 kDa protein	Methylation (KR)	R469
Q792Z1	Try10	MCG140784	Methylation (KR)	K112
Q922F4	Tubb6	Tubulin beta-6 chain	Methylpyrroline	K122
O08599	Stxbp1	Syntaxin-binding protein 1	Oxidation (HW)	W288
Q01853	Vcp	Transitional endoplasmic reticulum ATPase	Oxidation (HW)	W551
P63038	Hspd1	60 kDa heat shock protein, mitochondrial	Oxidation (M)	M356
P68134	Acta1	Actin, alpha skeletal muscle	Oxidation (M)	M134
H7BWZ3	Arpc3	Actin-related protein 2/3 complex subunit 3	Oxidation (M)	M11
P05067	APP	Amyloid-beta precursor protein	Oxidation (M)	M170
P05202	Got2	Aspartate aminotransferase, mitochondrial	Oxidation (M)	M38
P05202	Got2	Aspartate aminotransferase, mitochondrial	Oxidation (M)	M244
Q9DCX2	Atp5pd	ATP synthase subunit d, mitochondrial	Oxidation (M)	M20
G3X9L6	G3X9L6	ATP synthase subunit d, mitochondrial	Oxidation (M)	M20
Q8BFZ3	Actbl2	Beta-actin-like protein 2	Oxidation (M)	M154
Q8BFZ3	Actbl2	Beta-actin-like protein 2	Oxidation (M)	M191
A8DUK4	Hbb-bs	Beta-globin	Oxidation (M)	M56
Q68FD5	Cltc	Clathrin heavy chain 1	Oxidation (M)	M1538
Q04447	Ckb	Creatine kinase B-type	Oxidation (M)	M207
O08553	Dpysl2	Dihydropyrimidinase-related protein 2	Oxidation (M)	M37
P39053	Dnm1	Dynamin-1	Oxidation (M)	M332
P97807	Fh	Fumarate hydratase, mitochondrial	Oxidation (M)	M85
P48774	Gstm5	Glutathione S-transferase Mu 5	Oxidation (M)	M67
E9QKR0	Gnb2	Guanine nucleotide-binding protein G(I)/G(S)/G(T) subunit beta-2	Oxidation (M)	M367
P07901	Hsp90aa1	Heat shock protein HSP 90-alpha	Oxidation (M)	M98
G3UVV4	Hk1	Hexokinase 1, isoform CRA_f	Oxidation (M)	M254
P17710	Hk1	Hexokinase-1	Oxidation (M)	M311
A0A1W2P768	Hist2h3c1	Histone H3.2	Oxidation (M)	M136
Q564E2	Ldha	L-lactate dehydrogenase	Oxidation (M)	M264
P06151	Ldha	L-lactate dehydrogenase A chain	Oxidation (M)	M264
P08249	Mdh2	Malate dehydrogenase, mitochondrial	Oxidation (M)	M315
Q9QUK9	Try5	MCG15083	Oxidation (M)	M109
Q9QUK9	Try5	MCG15083	Oxidation (M)	M183
Q4VA93	Prkca	Protein kinase C	Oxidation (M)	M343
Q9D051	Pdhb	Pyruvate dehydrogenase E1 component subunit beta, mitochondrial	Oxidation (M)	M133
Q9Z2Q6	sept5	Septin-5	Oxidation (M)	M325
Q8VDN2	Atp1a1	Sodium/potassium-transporting ATPase subunit alpha-1	Oxidation (M)	M164
Q6PIC6	Atp1a3	Sodium/potassium-transporting ATPase subunit alpha-3	Oxidation (M)	M497
A3KGU9	Sptan1	Spectrin alpha chain, non-erythrocytic 1	Oxidation (M)	M278
P56677	St14	Suppressor of tumorigenicity 14 protein homolog	Oxidation (M)	M205
P60879	Snap25	Synaptosomal-associated protein 25	Oxidation (M)	M49
O08599	Stxbp1	Syntaxin-binding protein 1	Oxidation (M)	M38
A0A0A0MQA5	Tuba4a	Tubulin alpha chain	Oxidation (M)	M406
A0A0A0MQA5	Tuba4a	Tubulin alpha chain	Oxidation (M)	M442
P68373	Tuba1c	Tubulin alpha-1C chain	Oxidation (M)	M377
P68373	Tuba1c	Tubulin alpha-1C chain	Oxidation (M)	M413
P05214	Tuba3a	Tubulin alpha-3 chain	Oxidation (M)	M377
Q9JJZ2	Tuba8	Tubulin alpha-8 chain	Oxidation (M)	M377
A2AQ07	Tubb1	Tubulin beta-1 chain	Oxidation (M)	M257
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Oxidation (M)	M147
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Oxidation (M)	M164
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Oxidation (M)	M257
Q9ERD7	Tubb3	Tubulin beta-3 chain	Oxidation (M)	M164
Q9ERD7	Tubb3	Tubulin beta-3 chain	Oxidation (M)	M257
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Oxidation (M)	M164
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Oxidation (M)	M257
P99024	Tubb5	Tubulin beta-5 chain	Oxidation (M)	M164
P99024	Tubb5	Tubulin beta-5 chain	Oxidation (M)	M257
Q922F4	Tubb6	Tubulin beta-6 chain	Oxidation (M)	M164
Q922F4	Tubb6	Tubulin beta-6 chain	Oxidation (M)	M257
P50516	Atp6v1a	V-type proton ATPase catalytic subunit A	Oxidation (M)	M95
P62814	Atp6v1b2	V-type proton ATPase subunit B, brain isoform	Oxidation (M)	M25
P38647	Hspa9	Stress-70 protein, mitochondrial	Proline oxidation to pyroglutamic acid	P223
F8WGL3	Cfl1	Cofilin-1	Proline oxidation to pyrrolidinone	P67
P05214	Tuba3a	Tubulin alpha-3 chain	Proline oxidation to pyrrolidinone	P72
Q91V92	Acly	ATP-citrate synthase	Pyro-glu from Q	Q986
A2ALL9	Atp2b3	Calcium-transporting ATPase	Pyro-glu from Q	Q787
Q0VF55	Atp2b3	Calcium-transporting ATPase	Pyro-glu from Q	Q787
Q8C048	Atp2b3	Calcium-transporting ATPase	Pyro-glu from Q	Q787
E9Q828	Atp2b4	Calcium-transporting ATPase	Pyro-glu from Q	Q779
Q06890	Clu	Clusterin	Pyro-glu from Q	Q167
Q9JHU4	Dync1h1	Cytoplasmic dynein 1 heavy chain 1	Pyro-glu from Q	Q4229
Q8BMF4	Dlat	Dihydrolipoyllysine-residue acetyltransferase component of pyruvate dehydrogenase complex, mitochondrial	Pyro-glu from Q	Q500
O08553	Dpysl2	Dihydropyrimidinase-related protein 2	Pyro-glu from Q	Q44
Q99LC5	Etfa	Electron transfer flavoprotein subunit alpha, mitochondrial	Pyro-glu from Q	Q20
P11881	Itpr1	Inositol 1,4,5-trisphosphate receptor type 1	Pyro-glu from Q	Q584
P50446	Krt6a	Keratin, type II cytoskeletal 6A	Pyro-glu from Q	Q390
Q9Z331	Krt6b	Keratin, type II cytoskeletal 6B	Pyro-glu from Q	Q398
Q3UV11	Krt6b	Keratin, type II cytoskeletal 6B	Pyro-glu from Q	Q390
G5E829	Atp2b1	Plasma membrane calcium-transporting ATPase 1	Pyro-glu from Q	Q790
P62821	Rab1A	Ras-related protein Rab-1A	Pyro-glu from Q	Q104
Q8C266	Rab5c	Ras-related protein Rab-5C	Pyro-glu from Q	Q122
A0A0J9YUL3	Sept11	Septin 11, isoform CRA_b	Pyro-glu from Q	Q254
Q8VDN2	Atp1a1	Sodium/potassium-transporting ATPase subunit alpha-1	Pyro-glu from Q	Q708
Q6PIE5	Atp1a2	Sodium/potassium-transporting ATPase subunit alpha-2	Pyro-glu from Q	Q705
Q6PIC6	Atp1a3	Sodium/potassium-transporting ATPase subunit alpha-3	Pyro-glu from Q	Q698
A3KGU9	Sptan1	Spectrin alpha chain, non-erythrocytic 1	Pyro-glu from Q	Q486
P01831	Thy1	Thy-1 membrane glycoprotein	Pyro-glu from Q	Q20
P46460	Nsf	Vesicle-fusing ATPase	Pyro-glu from Q	Q170
Q9Z1G3	Atp6v1c1	V-type proton ATPase subunit C 1	Pyro-glu from Q	Q200
Q8R5H6	Wasf1	Wiskott-Aldrich syndrome protein family member 1	Replacement of proton with ammonium ion	R341
A0A0A0MQA5	Tuba4a	Tubulin alpha chain	Sulphone	M331
Q7TMM9	Tubb2a	Tubulin beta-2A chain	Tryptophan oxidation to kynurenin	W101
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Tryptophan oxidation to kynurenin	W101
Q9D6F9	Tubb4a	Tubulin beta-4A chain	Tryptophan oxidation to kynurenin	W344
P99024	Tubb5	Tubulin beta-5 chain	Tryptophan oxidation to kynurenin	W101
Q9CPN9	2210010C04Rik	RIKEN cDNA 2210010C04 gene	Tryptophan oxidation to oxolactone	W239

The identified modifications in adjacent control regions were: acetylation (N-term), acetylation (Protein N-term), alpha-amino adipic acid, amidation, carboxylation (DKW), citrullination, deamidation (NQ), glycidamide adduct, hexose (NSY), iminobiotinylation, ISD (z + 2)-series, maleimide, membrane protein extraction, methionine replacement by homopropargylglycine, methylation (KR), oxidation (M), phosphorylation (STY), pyro-glu from Q, sodium adduct, and ubiquitin (Fig. 2 C). The three most common modifications in the adjacent control group were oxidation (n = 49), deamidation (n = 22), and citrullination (n = 9).

Our findings indicated that oxidation and deamidation were the most common unique modification of proteins co-localizing with both Aβ plaque and adjacent control regions. The presence of the pyro-glu modification was identified to be much higher in proteins co-localizing with Aβ plaque compared to adjacent control regions.

DISCUSSION

Aβ plaques have previously been observed to encompass hundreds of different infiltrating proteins [23]. We identified 27 up- and downregulated proteins that associated with Aβ plaques in APPPS1-21 mice. To account for the current pathological state of the APPPS1-21 mice, we compared Aβ plaques to the adjacent control areas instead of comparing to brain areas in non-tg littermate mice reflecting normal conditions. Using this approach, we elucidated local proteomic differences and enabled identification of unique modifications associated to Aβ plaques and the surrounding closely related brain tissue of APPPS1-21 mice. On the technical side, we combined a few innovative methods to enhance the data quality. We used a detergent enhanced solubilization strategy with limited sample transfer, minimizing the potential risk of losing sample material during the workflow while enhancing protein solubility and trypsin digestion efficiency [17]. We further implemented a match-between-run strategy to increase the identification rate.

Our MS analysis identified 183 proteins that were consistently observed across Aβ plaque samples in the cerebral cortex of tg APPPS1-21 mice. These included proteins such as Clu, Apoe, Gapdh, and Pgam1. The respective human genes have been identified as susceptible genes for AD in different GWAS [34 , 39]. The APOE gene has been described as the most susceptibility gene for SAD. The gene exist in three types of alleles (ɛ2, ɛ3, and ɛ4) were the ɛ4 allele has been reported to be a strong risk factor for both FAD and SAD [40]. Despite reports of reduced GAPDH activity in AD brain, studies have observed interaction with fibrillar Aβ_1 - 42 [41, 42]. CLU and PGAM1 have been reported to be elevated in specific brain regions in AD patients [43, 44].

We detected 10 unique peptides for Clu, 8 for Apoe, 16 for Gapdh, and 10 for Pgam1 in comparison with App itself which was detected with 10 unique peptides, indicating a relevance and presence of these proteins within the Aβ plaques. Based on our findings, a relationship may be indicated in which an overexpression of Aβ in the APPPS1-21 mice can cause co-localization of these proteins to Aβ plaques. However, the proteins that we have detected in our study are translated from genes that are reasoned not to include the single nucleotide polymorphisms (SNPs) that have been identified in GWAS studies. The fact that these proteins co-localize with the plaques in the APPPS1-21 model may therefore indicate a physiological mechanism that occurs in response to the Aβ plaque deposition. For example, Apoe has been found to co-localize with Aβ plaques in tg APP mice and hypothesized to influence the degree of Aβ fibrillization [45] whereas Clu has been reported to interact directly with Aβ peptides [46]. The presence of SNPs may change the amino acid sequence of these proteins, impair the biological function and thereby impair the engagement of Aβ plaques. Notably, several differences between human and mouse Apoe has been reported [47]. However, despite the differences, we observed high degree of co-localization of Apoe within the Aβ plaque extracts compared to adjacent control regions in our study. Previous studies have generated mice expressing the human APOE, and crossed them with APP tg mice [49]. Such a double tg mouse strain would be a valuable tool to test the impact of human Apoe on the plaque pathology in the context of our study.

The majority of identified proteins (96.7%; 177/183; list included 2,422 proteins) has previously been detected in plaque extracts from AD patients when comparing to a complete list of identified proteins [23, 50]. In addition, the consistently observed proteins (69.9%; 128/183; list included 277 proteins) were in line with an earlier study [23] supporting the validity of our findings (data not shown). The previous comparison was based on findings in SAD patients whereas the mouse model, that was used in our study, was based on aggressive FAD mutations [24, 51]. The underlying cause for Aβ plaque generation in SAD is still insufficiently understood, and knowledge gained from mouse models with FAD mutations is a necessary approach. Comparing the identified proteins in the APPPS1-21 (FAD) mouse model here to plaque-associated proteins in SAD described previously, reveals high degree of alignment, despite the potential pathological- and species-related differences. The proteomic similarity on a Aβ plaque level in the APPPS1-21 mouse model compared to SAD therefore underlines the relevance of at least this tg mouse model in a human Aβ plaque context.

Our approach identified significantly upregulated proteins associated to two general processes; metabolism and endocytosis. In contrast, the downregulated proteins were associated to neuronal cytoskeleton. Glial cells have frequently been observed to engage and surround plaques in post mortem tissue analysis of AD patients and tg mouse models of AD [52 –54]. Astrocytes are hypothesized to release mediators that affect the inflammatory response, for example by production of proteases that induce the Aβ degradation and Apoe that lipidate Aβ which increase the clearance rate [55]. Similarly, activated microglia can migrate to Aβ plaques and initiate phagocytosis and/or release proteases [55]. We reason that the relationship between the up- and downregulated proteins identified in our study could be reflected in the plaque engagement by glial cells and degeneration of neuronal processes (neuritic plaques). The consistent identification of astrocytic markers such as GFAP, Glul, and Slc1a2 in the plaque extracts further supports this reasoning. Engaging glial cells are considered to exhibit increased metabolism that is associated with the degradation process. For example, Pgk1 and Eno1 have previously been reported as enzymes involved in glycolysis and gluconeogenesis but also linked to AD which supports their identification [56, 57].

A recent study used a similar approach to compare the proteomic similarity of Aβ plaques in AD and APPPS1-21 mice [58]. The study identified 4,092 ortholog pairs when comparing AD and APPPS1-21 brain samples indicating high degree of translatability between mouse and human brain proteome. The study also compared differential pathways (proteomic patterns) between plaque and adjacent control regions in both APPPS1-21 mice and AD [58]. They identified pathways that were either similar, such as proteins associated with cholesterol efflux and humoral immune responses, or different; such as proteins associated with AMPA receptor activity and glutathione metabolic processes. Among the most regulated pathways in APPPS1-21 mice were proteins associated with the TYRO protein tyrosine kinase-binding protein (TYROBP) network and microglia pathogen phagocytosis [58]. TYROBP (also known as DAP12) is a sensing and phagocytosis-related, microglia-enriched, phosphotyrosine phosphoprotein that forms functional complexes with TREM2 and the complement receptor CR3, among others [59, 60]. We observed a high number of proteins associated with glial sensing, indicating similarity to the findings of Xiong and colleagues, who also reported that proteins associated with complement activation and inflammatory response were highly regulated in AD. We observed cellular mediators of inflammation (e.g., peroxiredoxin-2 and 14-3-3 protein) among the consistent plaque-proteins in the APPPS1-21 mice, also supporting comparability with the findings from Xiong and colleagues. However, we did not observe proteins linked to the complement activation cascade associated with plaque extracts in our study. In contrast to Xiong and colleagues, we focused the microdissection to the cerebral cortex wherefore regional differences could explain the discrepancy on identifying proteins associated with the complement system.

Studies have indicated a notable involvement of different PTMs in the pathology of AD. These include for example, pyroglutamylation (pyro-glu), phosphorylation, tyrosine nitration, and oxidation of Aβ peptides [14]. However, it remains to be elucidated whether co-localizing proteins are subjected to similar degree of PTMs as Aβ plaques. We are the first to describe unique modification associated to proteins co-localizing with Aβ plaques. We identified 426 common modifications on proteins co-localizing with Aβ plaques and adjacent control extracts. Here, no difference could discriminate the two subgroups in the PCA plot of the LFQ intensities associated to the modifications.

A clear difference could be observed in the number of unique modifications in the two subgroups. The number of unique modifications on proteins co-localizing with Aβ plaque extracts were higher compared to adjacent control regions (193 versus 117), indicating that proteins co-localized with aggregated Aβ are more susceptible for modifications.

On a technical note, to limit false-positive identification of unique PTMs by random stochastic selection of ions during MS/MS, we applied the following validity criteria: 1) The modification was present in more than 50% of samples within a group, and 2) had a minimum a-score of 100 determined by PEAKS X Studio. The software uses a feature-based peptide identification method based on MS1 detection with a femtomole to sub femtomole detection range on the Q-Exactive HF-X instrument that support a true-positive identification.

We observed that oxidation and deamidation were the most abundant modifications on proteins within the Aβ plaque and adjacent control extracts in the APPPS1-21 mice. Oxidative stress and generation of free radicals have been observed in AD brains. These have been hypothesized to originate from Aβ peptides [61]. Our findings suggest that oxidative stress is also occurring in the brain of APPPS1-21 mice, reflected by the increase levels of oxidized proteins. Increased levels of deamidation have similarly been linked to dementia [62]. Our detection of deamidated proteins in APPPS1-21 mice is therefore also in line with a previous study using human brain tissue. On the other hand, deamidation and oxidation of peptides have been reported to be generated during sample preparation and should therefore be interpreted with caution [63, 64]. We therefore handled samples from all subgroups in a parallel approach, suggesting a similar degree of experimentally-generated oxidation and deamidation across the subgroups.

Citrullination was observed as a unique modification in proteins co-localizing with both Aβ plaques and adjacent control regions. This protein modification was identified as the third most abundant modification after oxidation and deamidation in the adjacent control group. Accumulation of citrullinated proteins have been observed in AD patients [65]. Citrullination has also been associated with autoimmune diseases such as rheumatoid arthritis and multiple sclerosis [66, 67], and detection of citrullination in the APPPS1-21 brains could indicate widespread inflammatory response in this mouse model.

The number of unique pyro-glu modification of proteins co-localizing with Aβ plaques was found to be much higher compared to the number in adjacent control regions (25 versus 6) in APPPS1-21 mice. We observed the unique pyro-glu modification on the following plaque co-localizing proteins: Acly, Atp2b3, Atp2b4, Clu, Dync1h1, Dlat, Dpysl2, Etfa, Itpr1, Krt6a, Krt6b, Atp2b1, Rab1A, Rab5c, Sept11, Atp1a1, Atp1a2, Atp1a3, Sptan1, Thy1, Nsf, and Atp6v1c1. Of these, Rab1A, Clu, and Sptan1 were among the significantly regulated proteins identified in our LFQ analysis in MaxQuant and Perseus. The pyro-glu modification of Aβ peptides has been observed to localize within the core of Aβ plaques and hypothesized to serve as seeds for new plaque deposition [68]. The pyro-glu modification is generated by cleavage of the N-terminus followed by a conversion catalyzed by the enzyme glutaminyl cyclase. The levels of glutaminyl cyclase have been found to correlate with pyro-glu modified Aβ peptides in AD patients [69]. Based on our findings from the APPPS1-21 mice, we suggest that glutaminyl cyclase activity is also increased in Aβ plaque regions in the tg mice, leading to increased pyro-glu modifications on co-localizing proteins.

Several immunotherapies targeting Aβ proteoforms have failed in the clinic, for many reasons, one of which could be a poor understanding of the molecular nature of aberrant neurotoxic Aβ species, processes, and PTMs driving pathogenesis. We identified unique modifications on App, Mapk3, Itpr1, Gnaq, and Cyc1 from Aβ plaque regions in the APPPS1-21 mice. These five proteins are reported in Kyoto Encyclopedia of Genes and Genomes (KEGG) AD pathway, highlighting their relevance in the human disease. Furthermore, proteins associated with endocytosis, such as CD47, Dnm1, and Rab1 were also identified to express unique modifications in Aβ plaque regions in the APPPS1-21 mice. As PTMs can affect the biological function of proteins, these modifications could have a potential impact and importance of the AD pathology. Therefore, findings from this study could increase our knowledge of the complex plaque-brain interaction and how PTMs might be involved in the process. Our study may support efforts to design and develop molecules that could target PTMs in AD.

In conclusion, this study has optimized a protocol for extracting and investigating the Aβ plaque proteome composition by LC-MS/MS. We extracted Aβ plaques from the APPPS1-21 mouse model and identified several proteins that support the current biological view of glial cell engagement in response to Aβ plaque deposition. The data further implicated that the level of proteins associated to degradation were increased within Aβ plaque extracts compared to the surrounding tissue in APPPS1-21 mice. Furthermore, we identified unique PTM differences between Aβ plaque and adjacent control extracts in APPPS1-21 mice which also could have relevant impact in AD pathogenesis.

ACKNOWLEDGMENTS

This work was partly funded by the Innovation Fund Denmark (IFD) under File No. 5189-00044B and the Lundbeck Foundation grant (R247-2017-239). The Danish National Mass Spectrometry Platform for Functional Proteomics (PRO-MS; grant no. 5072-00007B); the Obelske Family Foundation and the Svend Andersen Foundation are acknowledged for grants to the analytical platform enabling parts of this study.

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/19-0652r1).

Footnotes

The supplementary material is available in the electronic version of this article: .

References

Barker

, Luis

, Kashuba

, Luis

, Harwood

, Loewenstein

, Waters

, Jimison

, Shepherd

, Sevush

, Graff-Radford

, Newland

, Todd

, Miller

, Gold

, Heilman

, Doty

, Goodman

, Robinson

, Pearl

, Dickson

, Duara

(2002) Relative frequencies of Alzheimer disease, Lewy body, vascular and frontotemporal dementia, and hippocampal sclerosis in the State of Florida Brain Bank. Alzheimer Dis Assoc Disord 16, 203–12.

Bowler

, Munoz

, Merskey

, Hachinski

(1998) Fallacies in the pathological confirmation of the diagnosis of Alzheimer’s disease. J Neurol Neurosurg Psychiatry 64, 18–24.

Hardy

, Selkoe

(2002) The amyloid hypothesis of Alzheimer’s disease: Progress and problems on the road to therapeutics. Science 297, 353–6.

Masters

, Simms

, Weinman

, Multhaupt

, Mcdonald

(1985) Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc Natl Acad Sci U S A 82, 4245–4249.

Grundke-iqbal

, Iqbal

, Tung

, Quinlan

, Wisniewski

, Bindert

(1986) Abnormal phosphorylation of the microtubule-associated protein tau (tau) in Alzheimer cytoskeletal pathology. Proc Natl Acad Sci U S A 83, 4913–4917.

DeTure

, Dickson

(2019) The neuropathological diagnosis of Alzheimer’s disease. Mol Neurodegener 14, 32.

Selkoe

(1991) The molecular of Alzheimer’s pathology disease review. Neuron 6, 487–498.

Hardy

, Higgins

(1992) Alzheimer’s disease: the amyloid cascade hypothesis. Science 256, 184–185.

Haass

, Selkoe

(2007) Soluble protein oligomers in neurodegeneration: lessons from the Alzheimer’s amyloid β-peptide. Nat Rev Mol Cell Biol 8, 101–112.

10.

Walsh

, Selkoe

(2004) Deciphering the molecular basis of memory failure in Alzheimer’s disease. Neuron 44, 181–193.

11.

Souza

, Araújo

, Costa

, Oliveira

(2016) Combined genome-wide CSF Aβ-42’s associations and simple network properties highlight new risk factors for Alzheimer’s disease. J Mol Neurosci 58, 120–128.

12.

Jack

, Knopman

, Jagust

, Shaw

, Aisen

, Weiner

, Petersen

, Trojanowski

(2010) Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 9, 119–128.

13.

Atwood

, Martins

, Smith

, Perry

(2002) Senile plaque composition and posttranslational modification of amyloid-β peptide and associated proteins. Peptides 23, 1343–1350.

14.

Kummer

, Heneka

(2014) Truncated and modified amyloid-beta species. Alzheimers Res Ther 6, 1–9.

15.

Kakuda

, Miyasaka

, Iwasaki

, Nirasawa

, Wada-kakuda

(2017) Distinct deposition of amyloid-β species in brains with Alzheimer’s disease pathology visualized with MALDI imaging mass spectrometry. Acta Neuropathol Commun 5, 1–8.

16.

Moro

, Phillips

, Gaimster

, Paul

, Mudher

, Nicoll

JAR

, Boche

(2018) Pyroglutamate and Isoaspartate modified amyloid-beta in ageing and Alzheimer’s disease. Acta Neuropathol Commun 6, 3.

17.

Pirmoradian

, Budamgunta

, Chingin

, Zhang

, Astorga-wells

, Zubarev

(2013) Rapid and deep human proteome analysis by single-dimension shotgun proteomics. Mol Cell Proteomics 12, 3330–3338.

18.

Claridge

, Kastaniegaard

, Stensballe

, Greening

(2019) Post-translational and transcriptional dynamics - regulating extracellular vesicle biology. Expert Rev Proteomics 16, 17–31.

19.

Aebersold

, Mann

(2016) Mass-spectrometric exploration of proteome structure and function. Nature 537, 347–355.

20.

Han

, He

, Xin

, Shan

, Ma

(2011) PeaksPTM: Mass spectrometry-based identification of peptides with unspecified modifications. J Proteome Res 10, 2930–2936.

21.

Nasiri Kenari

, Kastaniegaard

, Greening

, Shambrook

, Stensballe

, Cheng

, Hill

(2019) Proteomic and post-translational modification profiling of exosome-mimetic nanovesicles compared to exosomes. Proteomics 19, 1800161.

22.

Michno

, Nyström

, Wehrli

, Lashley

, Brinkmalm

, Guerard

, Syvänen

, Sehlin

, Kaya

, Brinet

, Nilsson

KPR

, Hammarström

, Blennow

, Zetterberg

, Hanrieder

(2019) Pyroglutamation of amyloid-βx-42 (Aβx-42) followed by Aβ1-40 deposition underlies plaque polymorphism in progressing Alzheimer’s disease pathology. J Biol Chem 294, 6719–6732.

23.

Drummond

, Nayak

, Faustin

, Pires

, Hickman

, Askenazi

, Cohen

, Haldiman

, Kim

, Han

, Shao

, Safar

(2017) Proteomic differences in amyloid plaques in rapidly progressive and sporadic Alzheimer’s disease. Acta Neuropathol 133, 933–954.

24.

Radde

, Bolmont

, Kaeser

, Coomaraswamy

, Lindau

, Stoltze

, Calhoun

, Jäggi

, Wolburg

, Gengler

, Haass

, Ghetti

, Czech

, Hölscher

, Mathews

, Jucker

(2006) Aβ42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology. EMBO Rep 7, 940–946.

25.

Güntert

, Döbeli

, Bohrmann

(2006) High sensitivity analysis of amyloid-beta peptide composition in amyloid deposits from human and PS2APP mouse brain. Neuroscience 143, 461–475.

26.

Bastrup

, Birkelund

, Asuni

, Volbracht

, Stensballe

(2019) Dual strategy for reduced signal suppression effects in MALDI mass spectrometry imaging. Rapid Commun Mass Spectrom 33, 1711–1721.

27.

Bittner

, Burgold

, Dorostkar

, Fuhrmann

, Wegenast-braun

, Schmidt

, Herms

(2012) Amyloid plaque formation precedes dendritic spine loss. Acta Neuropathol 124, 797–807.

28.

Bennike

, Carlsen

, Ellingsen

, Bonderup

, Glerup

, Bøgsted

, Christiansen

, Birkelund

, Stensballe

, Andersen

(2015) Neutrophil extracellular traps in ulcerative colitis: A proteome analysis of intestinal biopsies. Inflamm Bowel Dis 21, 2052–2067.

29.

Cox

, Mann

(2008) MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat Biotechnol 26, 1367–1372.

30.

Tyanova

, Temu

, Sinitcyn

, Carlson

, Hein

, Geiger

, Mann

, Cox

(2016) The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat Methods 13, 731–740.

31.

Szklarczyk

, Franceschini

, Wyder

, Forslund

, Heller

, Huerta-Cepas

, Simonovic

, Roth

, Santos

, Tsafou

, Kuhn

, Bork

, Jensen

, von Mering

(2015) STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res 43, D447–D452.

32.

Ulrich

, Finn

, Wang

, Shen

, Mahan

, Jiang

, Stewart

, Piccio

, Colonna

, Holtzman

(2014) Altered microglial response to Aβ plaques in APPPS1-21 mice heterozygous for TREM2. Mol Neurodegener. 9, 9.

33.

Cox

, Hein

, Luber

, Paron

, Nagaraj

, Mann

(2014) Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ. Mol Cell Proteomics 13, 2513–2526.

34.

Shen

, Jia

(2016) An overview of genome-wide association studies in Alzheimer’s disease. Neurosci Bull 32, 183–190.

35.

, Kanekiyo

, Shinohara

, Zhang

, LaDu

, Xu

, Bu

(2012) Differential regulation of amyloid-β endocytic trafficking and lysosomal degradation by apolipoprotein E isoforms. J Biol Chem 287, 44593–44601.

36.

Akiyama

, Kawamata

, Dedhar

, McGeer

(1991) Immunohistochemical localization of vitronectin, its receptor and beta-3 integrin in Alzheimer brain tissue. J Neuroimmunol 32, 19–28.

37.

Ikezu

, Luo

, Weber

, Zhao

, McCabe

, Buescher

, Ghorpade

, Zheng

, Xiong

(2003) Amyloid precursor protein-processing products affect mononuclear phagocyte activation: pathways for sAPP- and Aβ-mediated neurotoxicity. J Neurochem 85, 925–934.

38.

Lee

, Cheng

, Rogaeva

, Meng

, Stern

, Santana

, Lantigua

, Medrano

, Jimenez-Velazquez

, Farrer

, St. George-Hyslop

, Mayeux

(2008) Further examination of the candidate genes in chromosome 12p13 locus for late-onset Alzheimer disease. Neurogenetics 9, 127–138.

39.

Talwar

, Sinha

, Grover

, Rawat

, Kushwaha

, Agarwal

, Taneja

, Kukreti

(2016) Dissecting complex and multifactorial nature of Alzheimer’s disease pathogenesis: A clinical, genomic, and systems biology perspective. Mol Neurobiol 53, 4833–4864.

40.

Liu

C-C

, Kanekiyo

, Xu

, Bu

(2013) Apolipoprotein E and Alzheimer disease: risk, mechanisms and therapy. Nat Rev Neurol 9, 106–118.

41.

Verdier

, Földi

, Sergeant

, Fülöp

, Penke

, Janáky

, Szücs

, Penke

(2008) Characterization of the interaction between Aβ1-42 and glyceraldehyde phosphodehydrogenase. J Pept Sci 14, 755–762.

42.

Kish

, Lopes-Cendes

, Guttman

, Furukawa

, Pandolfo

, Rouleau

, Ross

, Nance

, Schut

, Ang

, DiStefano

(1998) Brain glyceraldehyde-3-phosphate dehydrogenase activity in human trinucleotide repeat disorders. Arch Neurol 55, 1299.

43.

Lidström

, Bogdanovic

, Hesse

, Volkman

, Davidsson

, Blennow

(1998) Clusterin (Apolipoprotein J) protein levels are increased in hippocampus and in frontal cortex in Alzheimer’s disease. Exp Neurol 154, 511–521.

44.

Kim

, Franck

, Kang

, Heinsen

, Ravid

, Ferrer

, Hee Cheon

, Lee

, Shin Yoo

, Steinbusch

, Salzet

, Fournier

, Mok Park

(2015) Proteome-wide characterization of signalling interactions in the hippocampal CA4/DG subfield of patients with Alzheimer’s disease. Sci Rep 5, 11138.

45.

Bales

, Verina

, Dodel

, Du

, Altstiel

, Bender

, Hyslop

, Johnstone

, Little

, Cummins

, Piccardo

, Ghetti

, Paul

(1997) Lack of apolipoprotein E dramatically reduces amyloid β-peptide deposition. Nat Genet 17, 263–264.

46.

, Ma

, Wei

, Li

, Wu

, Zhuang

, Zhao

(2014) Clusterin in Alzheimer’s disease: a player in the biological behavior of amyloid-beta. Neurosci Bull 30, 162–168.

47.

Maloney

, Ge

Y-W

, Alley

, Lahiri

(2007) Important differences between human and mouse APOE gene promoters: limitation of mouse APOE model in studying Alzheimer’s disease. J Neurochem 103, 1237–1257.

48.

Sullivan

, Mace

, Maeda

, Schmechel

(2004) Marked regional differences of brain human apolipoprotein e expression in targeted replacement mice. Neuroscience 124, 725–733.

49.

Kuszczyk

, Sanchez

, Pankiewicz

, Kim

, Duszczyk

, Guridi

, Asuni

, Sullivan

, Holtzman

, Sadowski

(2013) Blocking the interaction between Apolipoprotein E and Aβ reduces intraneuronal accumulation of Aβ and inhibits synaptic degeneration. Am J Pathol 182, 1750–1768.

50.

Liao

, Cheng

, Wang

, Duong

, Losik

, Gearing

, Rees

, Lah

, Levey

, Peng

(2004) Proteomic characterization of postmortem amyloid plaques isolated by laser capture microdissection. J Biol Chem 279, 37061–37068.

51.

Moehlmann

, Winkler

, Xia

, Edbauer

, Murrell

, Capell

, Kaether

, Zheng

, Ghetti

, Haass

, Steiner

(2002) Presenilin-1 mutations of leucine 166 equally affect the generation of the Notch and APP intracellular domains independent of their effect on Aβ42 production. Proc Natl Acad Sci U S A 99, 8025–8030.

52.

Hickman

, Allison

, El Khoury

(2008) Microglial dysfunction and defective β-amyloid clearance pathways in aging Alzheimer’s disease mice. J Neurosci 28, 8354–8360.

53.

Mattiace

, Davies

, Yen

, Dickson

(1990) Microglia in cerebellar plaques in Alzheimer’s disease. Acta Neuropathol 80, 493–498.

54.

Itagaki

, Mcgeer

, Akiyama

, Zhu

, Selkoe

(1989) Relationship of microglia and astrocytes to amyloid deposits of Alzheimer disease. J Neuroimmunol 24, 173–182.

55.

Heneka

, Golenbock

, Latz

(2015) Innate immunity in Alzheimer’s disease. Nat Immunol 16, 229–236.

56.

Cutler

, Kay

, Marangos

, Burg

(1986) Cerebrospinal fluid neuron-specific enolase is reduced in Alzheimer’s disease. Arch Neurol 43, 153–154.

57.

Butterfield

, Lange

(2009) Multifunctional roles of enolase in Alzheimer disease brain: beyond altered glucose metabolism. J Neurochem 111, 915–933.

58.

Xiong

, Ge

, Ma

(2019) Quantitative proteomics reveals distinct composition of amyloid plaques in Alzheimer’s disease. Alzheimers Dement 15, 429–440.

59.

Hickman

, Kingery

, Ohsumi

, Borowsky

, Wang

, Means

, El Khoury

(2013) The microglial sensome revealed by direct RNA sequencing. Nat Neurosci 16, 1896–1905.

60.

Konishi

, Kiyama

(2018) Microglial TREM2/DAP12 signaling: A double-edged sword in neural diseases. Front Cell Neurosci 12, 1–14.

61.

Hensley

, Hall

, Subramaniam

, Cole

, Harris

, Aksenov

, Aksenova

, Gabbita

, Wu

, Carney

, Lovell

, Markesbery

, Butterfield

(1995) Brain regional correspondence between Alzheimer’s disease histopathology and biomarkers of protein oxidation. J Neurochem 65, 2146–2156.

62.

Adav

, Gallart-Palau

, Tan

, Lim

, Tam

, Sze

(2016) Dementia-linked amyloidosis is associated with brain protein deamidation as revealed by proteomic profiling of human brain tissues. Mol Brain 9, 20.

63.

Hao

, Ren

, Datta

, Tam

, Sze

(2015) Evaluation of the effect of trypsin digestion buffers on artificial deamidation. J Proteome Res 14, 1308–1314.

64.

Boys

, Kuprowski

, Noël

, Konermann

(2009) Protein oxidative modifications during electrospray ionization: solution phase electrochemistry or corona discharge-induced radical attack? Anal Chem 81, 4027–4034.

65.

Ishigami

, Ohsawa

, Hiratsuka

, Taguchi

, Kobayashi

, Saito

, Murayama

, Asaga

, Toda

, Kimura

, Maruyama

(2005) Abnormal accumulation of citrullinated proteins catalyzed by peptidylarginine deiminase in hippocampal extracts from patients with Alzheimer’s disease. J Neurosci Res 80, 120–128.

66.

Valesini

, Gerardi

, Iannuccelli

, Pacucci

, Pendolino

, Shoenfeld

(2015) Citrullination and autoimmunity. Autoimmun Rev 14, 490–497.

67.

Yang

, Tan

, Piao

(2016) Myelin basic protein citrullination in multiple sclerosis: a potential therapeutic target for the pathology. Neurochem Res 41, 1845–1856.

68.

Sullivan

, Berg

, Elliott-Bryant

, Fishman

, McKee

, Morin

, Shia

, Fine

(2011) Pyroglutamate-Aβ 3 and 11 colocalize in amyloid plaques in Alzheimer’s disease cerebral cortex with pyroglutamate-Aβ 11 forming the central core. Neurosci Lett 505, 109–112.

69.

Schilling

, Zeitschel

, Hoffmann

, Heiser

, Francke

, Kehlen

, Holzer

, Hutter-Paier

, Prokesch

, Windisch

, Jagla

, Schlenzig

, Lindner

, Rudolph

, Reuter

, Cynis

, Montag

, Demuth

H-U

, Rossner

(2008) Glutaminyl cyclase inhibition attenuates pyroglutamate Aβ and Alzheimer’s disease-like pathology. Nat Med 14, 1106–1111.