The amyloid plaque proteome in early onset Alzheimer's disease and Down syndrome

Abstract

Amyloid plaques contain many proteins in addition to beta amyloid (Aβ). Previous studies examining plaque-associated proteins have shown these additional proteins are important; they provide insight into the factors that drive amyloid plaque development and are potential biomarkers or therapeutic targets for Alzheimer's disease (AD). The aim of this study was to comprehensively identify proteins that are enriched in amyloid plaques using unbiased proteomics in two subtypes of early onset AD: sporadic early onset AD (EOAD) and Down Syndrome (DS) with AD. We focused our study on early onset AD as the drivers of the more aggressive pathology development in these cases is unknown and it is unclear whether amyloid-plaque enriched proteins differ between subtypes of early onset AD. Amyloid plaques and neighbouring non-plaque tissue were microdissected from human brain sections using laser capture microdissection and label-free LC-MS was used to quantify the proteins present. 48 proteins were consistently enriched in amyloid plaques in EOAD and DS. Many of these proteins were more significantly enriched in amyloid plaques than Aβ. The most enriched proteins in amyloid plaques in both EOAD and DS were: COL25A1, SMOC1, MDK, NTN1, OLFML3 and HTRA1. Endosomal/lysosomal proteins were particularly highly enriched in amyloid plaques. Fluorescent immunohistochemistry was used to validate the enrichment of four proteins in amyloid plaques (moesin, ezrin, ARL8B and SMOC1) and to compare the amount of total Aβ, Aβ40, Aβ42, phosphorylated Aβ, pyroglutamate Aβ species and oligomeric species in EOAD and DS. These studies showed that phosphorylated Aβ, pyroglutamate Aβ species and SMOC1 were significantly higher in DS plaques, while oligomers were significantly higher in EOAD. Overall, we observed that amyloid plaques in EOAD and DS largely contained the same proteins, however the amount of enrichment of some proteins was different in EOAD and DS. Our study highlights the significant enrichment of many proteins in amyloid plaques, many of which may be potential therapeutic targets and/or biomarkers for AD.

Keywords: Alzheimer’s disease; Amyloid beta; Amyloid plaques; Down syndrome; Early onset; Mass spectrometry; Proteomics.

Fig. 1

Schematic of methods used in this study. Formalin fixed paraffin embedded human tissue samples containing the hippocampus were used in this study (n = 5/group; all with advanced AD neuropathology [A3, B3, C3]). Laser capture microdissection was used to microdissect plaques or neighboring non-plaque tissue, LC–MS was used to quantify proteins present in each sample and various bioinformatics approaches were used to identify plaque enriched proteins and pathway or cell-type enrichment. Immunohistochemistry and comparison with previous studies through systematic literature searches was used to validate the enrichment of selected proteins in plaques

Fig. 2

Comparison of levels of different Aβ species and oligomers in DS, EOAD and cognitively normal controls. Representative fluorescent immunohistochemistry images show the distribution of total Aβ, Aβ42, Aβ40, phosphorylated Aβ, pyroglutamate Aβ and oligomers in the cortex. A Similar amounts of plaques containing Aβ, Aβ42 and Aβ40 were observed in DS in comparison to EOAD. B Phosphorylated Aβ was observed in plaques and intraneuronally in both DS and EOAD. Intraneuronal phosphorylated Aβ was observed in both DS and EOAD (higher magnification image inserts in B). Pyroglutamate Aβ was only observed in plaques in DS and EOAD. Immunostaining for the conformational oligomer antibody TWF9 was observed intraneuronally, but not in plaques. Significantly higher amounts of plaque-associated phosphorylated Aβ and pyroglutamate Aβ were observed in DS in comparison to EOAD and controls. In contrast, significantly higher amounts of oligomers were observed in EOAD in comparison to DS and controls. Significant differences were determined by one-way ANOVA followed by Tukey’s multiple comparisons test. *p < 0.05; **p < 0.01

Fig. 3

48 proteins were significantly enriched in plaques in both DS and EOAD. A 107 proteins were enriched in DS plaques and 68 proteins were enriched in EOAD plaques. Of these, 48 proteins were enriched in both DS and EOAD. B Aβ was significantly enriched in plaques in comparison to neighboring non-plaque tissue in both DS (11.92 fold enriched) and EOAD (6.96 fold enriched; paired t-test). C There was a highly significant correlation in the abundance (determined by intensity values from LC–MS) of common plaque enriched proteins in DS and EOAD. Apolipoprotein E (APOE), APP and vimentin (VIM) were the three most abundant proteins in plaques in both DS and EOAD. Proteins are coloured to show if each is a previously validated plaque protein (red: 56.2% proteins previously validated as a plaque protein in a targeted immunohistochemistry [IHC] study; blue: 12.5% proteins previously validated as a plaque protein in a proteomics study only) or a novel identified plaque protein (green; 31.3% proteins). D Pathway analysis of the 48 proteins commonly enriched in plaques in both DS and EOAD showed a highly significant degree of protein–protein interactions (p < 1.0 × 10⁻¹⁶). Pathway analysis showed that these proteins were highly enriched extracellular proteins (blue), endosome proteins (green) or lysosome proteins (red). *p < 0.05; **p < 0.01

Fig. 4

Significantly altered proteins in plaques in comparison to neighboring non-plaque tissue. A, B Volcano plots highlight proteins in red that were significantly altered in plaques in comparison to non-plaque tissue. Significance was determined by a combination of p < 0.05 and a fold change difference of greater than 1.5 fold. Proteins that have a fold change difference of greater than 1.5 fold only are shown in green and proteins that had a difference of p < 0.05 only are shown in blue. The total number of proteins included in the analysis was 2059 proteins for DS and 2115 proteins for EOAD. Proteins are identified by gene IDs. C, D Unsupervised clustering heatmaps for proteins that were significantly altered in DS or EOAD. Plaque and non-plaque samples independently clustered, highlighting the significantly different protein expression between plaque and non-plaque samples for DS and EOAD. All gene IDs are indicated for EOAD in each row whilst only genes from cluster 1 and 4 are marked for DS, constituting a divergent cluster and highly enriched cluster of genes respectively for DS plaques

Fig. 5

Comparison of common plaque enriched proteins in DS and EOAD with previous proteomic studies. Plot shows the 30 plaque proteins that were either identified in plaques or correlated with Aβ in at least 3 previous proteomic studies. Proteomic data was obtained from [44] for enrichment in preclinical AD or LOAD plaques and [60] for correlation with Aβ. Blue boxes indicate protein significantly enriched in plaques in comparison to surrounding non-plaque tissue or significantly correlated with Aβ. Red boxes indicate detection in the study but no enrichment in plaques or correlation with Aβ. White boxes indicate instances when a protein was not detected. Proteins are listed in order of fold change enrichment in plaques in EOAD followed by fold change enrichment in DS plaques if not enriched in EOAD

Fig. 6

Validation of moesin as a plaque enriched protein in human brain tissue by immunohistochemistry. Enrichment of moesin (MSN) in amyloid plaques was observed in DS, EOAD and LOAD cases. Moesin was also observed outside of plaques in all tissue examined (including cognitively normal control tissue) in cells consistent with a microglial morphology

Fig. 7

Validation of ezrin as a plaque enriched protein in human brain tissue by immunohistochemistry. Enrichment of ezrin (EZR) was observed in amyloid plaques in DS, EOAD and LOAD cases

Fig. 8

Validation of SMOC1 as a plaque enriched protein in human brain tissue by immunohistochemistry. A Enrichment of SMOC1 was observed in a sub-population of amyloid plaques in DS, EOAD and LOAD cases. B Plot shows percentage of SMOC1 immunoreactive plaques in the hippocampus of DS, EOAD and LOAD cases (n = 3/group). Results generated by an analysis of 321 ± 47 hippocampal plaques (average ± SEM) in each case. The ratio of SMOC1 positive plaques (immunoreactive for both Aβ and SMOC1) over the total number of amyloid plaques was calculated for each case in DS, EOAD and LOAD. C Representative images of diffuse and neuritic plaques immunolabeled with SMOC1. D Representative images of SMOC1, pyroglutamate Aβ and phosphorylated Aβ immunolabelled plaques in the hippocampus of a representative Down syndrome case. Fluorescent immunohistochemistry was used to identify SMOC1, pyroglutamate Aβ or phosphorylated Aβ immunoreactive plaques on three sequential hippocampal sections from the same case. White arrowheads show SMOC1 immunoreactive amyloid plaques that were also immunoreactive for pyroglutamate Aβ and phosphorylated Aβ species. Red arrowheads show pyroglutamate Aβ and/or phosphorylated Aβ immunoreactive plaques negative for SMOC1. *p < 0.05

Fig. 9

Validation of ARL8B as a plaque enriched protein in human brain tissue by immunohistochemistry. A Enrichment of ARL8B in amyloid plaques was observed in DS, EOAD and LOAD cases. B Plot shows percentage of ARL8B immunoreactive plaques in the hippocampus of DS, EOAD and LOAD cases (n = 3/group). Results generated by an analysis of 309 ± 41 hippocampal plaques (average ± SEM) in each case. The ratio of ARL8B positive plaques (immunoreactive for both Aβ and ARL8B) over the total number of amyloid plaques was calculated for each case in DS, EOAD and LOAD. C Representative images showing ARL8B distribution in amyloid plaques. Bright puncta of ARL8B were diffusely present throughout both diffuse and neuritic plaques. Basal ARL8B staining was observed in controls in neuron soma. D Intense ARL8B immunoreactivity was observed in plaque-associated cells (i; arrows). Double-fluorescent immunohistochemistry showed that these plaque-associated cells with intense ARL8B immunoreactivity were a subset of plaque-associated reactive astrocytes (ii; GFAP, red arrows), and not plaque associated reactive microglia (iii; IBA1, white arrows) or neurons (iv; MAP2, white arrows). Insert in ii shows a higher magnification image of the colocalization of ARL8B and GFAP in plaque associated astrocytes

Fig. 10

Co-localization of plaque enriched proteins with vascular amyloid deposition. Representative images of vascular amyloid pathology immunolabeled with moesin (MSN), ezrin (EZR), SMOC1, ARL8B (green) and Aβ (4G8/6E10, red). Moesin, ezrin and SMOC1 co-localized with vascular amyloid pathology while ARL8B did not