Proteomics analysis identifies new markers associated with capillary cerebral amyloid angiopathy in Alzheimer’s disease

Proteomics analysis identifies new markers associated with capillary cerebral amyloid angiopathy... Alzheimer’s disease (AD) is characterized by amyloid beta (Aβ) deposits as plaques in the parenchyma and in the walls of cortical and leptomeningeal blood vessels of the brain called cerebral amyloid angiopathy (CAA). It is suggested that CAA type-1, which refers to amyloid deposition in both capillaries and larger vessels, adds to the symptomatic manifestation of AD and correlates with disease severity. Currently, CAA cannot be diagnosed pre-mortem and disease mechanisms involved in CAA are elusive. To obtain insight in the disease mechanism of CAA and to identify marker proteins specifically associated with CAA we performed a laser dissection microscopy assisted mass spectrometry analysis of post-mortem human brain tissue of (I) AD cases with only amyloid deposits in the brain parenchyma and no vascular related amyloid, (II) AD cases with severe CAA type-1 and no or low numbers of parenchymal amyloid deposits and (III) cognitively healthy controls without amyloid deposits. By contrasting the quantitative proteomics data between the three groups, 29 potential CAA-selective proteins were identified. A selection of these proteins was analysed by immunoblotting and immunohistochemistry to confirm regulation and to determine protein localization and their relation to brain pathology. In addition, specificity of these markers in relation to other small vessel diseases including prion CAA, CADASIL, CARASAL and hypertension related small vessel disease was assessed using immunohistochemistry. Increased levels of clusterin (CLU), apolipoprotein E (APOE) and serum amyloid P-component (APCS) were observed in AD cases with CAA. In addition, we identified norrin (NDP) and collagen alpha-2(VI) (COL6A2) as highly selective markers that are clearly present in CAA yet virtually absent in relation to parenchymal amyloid plaque pathology. NDP showed the highest specificity to CAA when compared to other small vessel diseases. The specific changes in the proteome of CAA provide new insight in the pathogenesis and yields valuable selective biomarkers for the diagnosis of CAA. Keywords: Cerebral amyloid angiopathy, Amyloid beta, Alzheimer’s disease, Proteomics, Biomarker, Laser microdissection, Human brain, Post-mortem tissue Introduction pathology in varying degrees. When restricted to the Alzheimer’s disease (AD) pathology is characterized by larger blood vessels, including leptomeningeal vessels, the deposition of amyloid beta (Aβ) in the brain paren- cortical arteries and arterioles, this is referred to as chyma as amyloid plaques and at the brain vasculature. CAA type-2. In approximately 50% of the AD cases also The latter is referred to as cerebral amyloid angiopathy brain capillaries are affected, which is designated as CAA (CAA). Approximately 80% of AD cases have CAA type-1 [1, 2]. Especially around the capillaries the Aβ de- posits can extend into the parenchyma as perivascular Aβ also referred to as dyshoric changes [3]. In AD the observed * Correspondence: D.C.Hondius@vu.nl; D.Hondius@vumc.nl August B. Smit and Annemieke J. M. Rozemuller contributed equally to this plaque pathology and CAA type-1 capillary deposits have work. an inverse correlation [4]. Department of Pathology, Amsterdam Neuroscience, VU University Medical Aβ deposition at the vessel wall in CAA correlates Center, PO Box 7057, 1007, MB, Amsterdam, The Netherlands Department of Molecular and Cellular Neurobiology, Center for with an increase in the occurrence of cerebral infarction, Neurogenomics and Cognitive Research, Amsterdam Neuroscience, VU cerebral haemorrhage and micro-bleeds. In addition, it University Amsterdam, Amsterdam, The Netherlands © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 2 of 19 causes a structural disruption of the vascular wall and neurofibrillary tangles and neuritic plaques is staged might indirectly deteriorate the integrity of the micro- [12–14] and indicated conform the ABC criteria [15]. vascular network [5, 6]. Aβ peptide transport through the blood brain barrier (BBB) or via perivascular drain- Fast immunohistochemistry for LCM age is an important mechanism to clear the brain from Sections (10 μm) of fresh frozen occipital tissue were Aβ. Disruption of Aβ clearance is thought to lead to in- mounted on PEN-membrane slides (Leica), air-dried and crease in Aβ deposition in the walls of capillaries and fixed in 100% ethanol for 1 min. After air drying the tis- blood vessels, which in turn further decreases drainage sue was wetted with sterile PBS. Anti-Aβ (clone IC16, capacity resulting in further enhancement of Aβ depos- detecting N-terminal part of Aβ [16]) was applied at a 1: ition [7, 8]. 100 dilution in sterile PBS (pH 7.5) and incubated for CAA type-1 is clinically highly relevant, as it contributes 20 min at RT. After washing 3 times for 30 s in sterile to the symptomatic appearance of AD and, in severe form, PBS, HRP labelled rabbit anti-mouse (DAKO) was ap- CAA type-1 can present itself as the primary cause of plied at a 1:100 dilution in sterile PBS and incubated for rapidly progressive dementia [9, 10]. 15 min at RT. Sections were briefly washed (3 × 30 s) Currently, definitive diagnosis of AD and the occurrence and freshly prepared 3,3′ diaminobenzidine (DAB) solu- of CAA can only be determined post-mortem. However, tion was applied and left to incubate for 5 min to visualize the presence or absence of CAA in AD patients might antibody binding. Sections were thoroughly washed in alter therapeutic options. In particular, a biomarker to de- ultra-pure H O and incubated with 1% (w/v) toluidine tect CAA in patients might aid in stratification of patient blue in ultrapure H O for 1 min as a counterstain. Sec- groups, which is highly important when initiating, inter- tions were then washed in ultra-pure H O twice for 1 min preting and improving outcome in clinical trials. More- and twice in 100% ethanol for 1 min and air dried. over, proteins selectively involved in CAA may function as therapeutic targets. Brain tissue preparation and laser capture Proteomics analysis using mass spectrometry is a pre- microdissection (LCM) ferredmethodtoobtainanunbiasedinsightinto proteins Laser capture microdissection (LCM) was performed as involved in disease. For this, 20 cases were selected encom- described previously [17]. LCM was performed using a passing a group of AD patients with severe CAA type-1, a Leica AS LMD system (Leica). Cortical layers II to VI group with AD bearing severe plaque pathology but devoid which were randomly selected from control tissue and of CAA, and a cognitively healthy control group without selected based on the presence of severe Aβ pathology any pathology in the occipital lobe. Subsequently, we per- in the case of AD and CAA were collected in Eppendorf formed a proteomics analysis of small laser dissected oc- tubes containing 30 μl M-PER lysis buffer (Thermo cipital tissue sections containing either high plaque load, or Scientific) supplemented with reducing SDS sample severe CAA or no Aβ deposits. By contrasting the protein buffer (Thermo Scientific). Between 10 and 20 tissue sec- expression profiles of these subject groups we discovered tions with a thickness of 10 μm were captured using LCM, 9 3 proteins that are highly selective for CAA. These proteins yielding an equal volume each of 1.0 × 10 μm .Micro- also provide insight in specific pathogenic components of dissected tissue was stored at − 80 °C until further use. CAA, which might offer new targets for therapy. Protein separation by electrophoresis and in-gel digestion Material and methods Micro-dissected tissue lysates were incubated at 95 °C Case selection for 5 min to denature the proteins, followed by incu- Post mortem brain tissue was obtained from the bation with 50 mM iodoacetamide for 30 min at RT Netherlands Brain Bank (NBB), Netherlands Institute in the dark to alkylate the cysteine residues. To re- for Neuroscience (NIN), Amsterdam. All brain tissue duce protein complexity, samples were size separated was collected from donors with written informed con- on a NuPAGE® 4–12% Bis-Tris acrylamide gel using sent forbrainautopsyand theuse of thematerialand MOPS SDS running buffer (Invitrogen) according to clinical information for research purposes has been the manufacturers’ protocol. obtained by the NBB. Brain tissue was selected based Gels were fixed in a solution containing 50% (v/v) on clinical and neuropathological reports. Three groups ethanol and 3% (v/v) phosphoric acid in H O for 3 h at were composed. Cognitively healthy control cases lacking RT and stained with Colloidal Coomassie Blue (34% (v/v) any pathology, AD cases with severe plaque pathology but methanol, 3% (v/v) phosphoric acid, 15% (w/v) ammonium devoid of CAA and CAA type-1 cases with severe and Sulphate, and 0.1% (w/v) Coomassie brilliant blue G-250 (nearly) pure capillary CAA pathology (Thal stages 2 (Thermo Scientific), overnight while shaking. Destaining and 3 for CAA) [11]. All cases are listed in Table 1. was performed in ultra-pure water under gentle agitation Alzheimer’s disease pathology present as Aβ deposits, for several hours to reduce background staining (Additional Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 3 of 19 Table 1 Patient data MS/Validation Case Diagnosis M/F Age (years) Abeta Tau CERAD PMD APOE MS 1 CAA type-1 F 75 A3 B3 C0 6:00 44 MS 2 CAA type-1 F 96 A3 B3 C0 4:20 43 a c MS 3 CAA type-1 M 68 A3 B1 C0 6:05 44 MS 4 CAA type-1 F 78 A3 NA C0 4:20 44 MS 5 CAA type-1 M 81 A3 B3 C2 6:30 44 MS 6 CAA type-1 F 95 A3 B3 C2 4:35 44 MS 7 CAA type-1 M 80 A3 B3 C0 5:05 44 MS 8 AD F 82 A3 B3 C3 6:00 42 MS 9 AD F 72 A3 B3 C3 6:30 44 MS 10 AD F 81 A3 B3 C3 6:00 33 MS 11 AD F 73 A3 B3 C3 5:55 44 MS 12 AD M 84 A3 B3 C3 8:05 NA MS 13 AD F 87 A3 B3 C3 5:45 43 MS 14 AD F 72 A3 B3 C3 5:55 23 MS 15 Control M 74 A0 B0 C0 8:05 33 MS 16 Control F 80 A1 B1 C0 6:58 43 MS 17 Control M 82 A0 B1 C0 5:10 23 MS 18 Control M 78 A0 B1 C0 17:40 33 MS 19 Control F 79 A0 B1 C0 18:13 33 MS 20 Control F 81 A0 B1 C0 4:25 33 V 21 CAA type-1 F 94 A3 B3 C3 04:30 43 V 22 CAA type-1 M 74 A3 B3 C3 03:25 NA V 23 CAA type-1 F 87 A3 B3 C3 08:00 44 V 24 CAA type-1 F 84 A3 B3 C2 04:45 NA V 25 CAA type-1 M 88 A3 B3 C3 03:55 NA V 26 CAA type-1 M 75 A3 B3 C0 03:15 NA V 27 AD M 64 A3 B3 C3 07:30 33 V 28 AD F 81 A3 B3 C3 05:15 43 V 29 AD F 90 A3 B3 C3 04:45 33 V 30 AD M 65 A3 B3 C3 06:00 43 V 31 AD F 73 A3 B3 C3 NA NA V 32 AD F 90 A3 B3 C3 03:55 32 V 33 AD M 88 A3 B3 C3 04:40 43 V 34 AD M 74 A3 B3 C3 05:10 NA V 35 Control M 73 A0 B0 C0 24:45 33 V 36 Control M 71 A0 B1 C0 07:40 33 V 37 Control F 82 A0 B1 C0 07:00 33 V 38 Control M 56 A0 B0 C0 09:15 43 V 39 Control M 62 A0 B1 C0 07:20 33 V 40 Control M 76 A0 B0 C0 06:45 33 V 41 Control M 93 A0 B1 C0 05:05 33 V 42 Control F 60 A0 B0 C0 08:10 32 V 43 Cotton wool M 72 A3 B3 C0 05:15 43 V 44 Prp-CAA F 57 A0 B0 C0 24:00 NA Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 4 of 19 Table 1 Patient data (Continued) MS/Validation Case Diagnosis M/F Age (years) Abeta Tau CERAD PMD APOE V 45 CADASIL M 73 A0 B0 C0 31:45 NA V 46 CARASAL F 55 A1 B1 C0 04:00 NA V 47 Hyper tension related SVD F 92 A1 B2 C0 07:25 NA Alzheimer’s disease: AD, cerebral amyloid angiopathy: CAA, M: male, F: female, post mortem delay: PMD, not available/not applicable: NA, used for mass a b spectrometry analysis: MS, used for validation: V. ( Aβ only present as dysphoric CAA) ( Focal tau accumulation around blood vessels with prp-amyloid deposits) ( only dyshoric angiopathy in gallyas staining) file 1: Figure S1 ). Each gel lane was sliced into 12 equal modifications. Mass deviation tolerance was set to sized parts to reduce sample complexity during later mass 20 ppm for monoisotopic precursor ions and 0.5 Da for spectrometry analysis and each part was cut into blocks of MS/MS peaks. False-discovery rate cut-offs for peptide approximately 1 mm and collected in an Eppendorf tube. and protein identifications were set to 1% for both. The Gel fragments were destained in ultrapure water with minimum peptide length was seven amino acids. Identi- 50 mM NH HCO and 50% (v/v) acetonitrile overnight. fied proteins that had the same set of peptides or a sub- 4 3 Gel fragments were dehydrated using acetonitrile for set of peptides compared to another protein, were 20 min and dried for 30 min using a SpeedVac. The merged into one protein group. Peptides that were shared gel parts were rehydrated in 70 μl of ultra-pure water between different proteins were assigned to the protein containing 50 mM NH HCO and 10 μg/ ml trypsin with most peptide evidence (so-called ‘Razor’ peptides). 4 3 (sequence grade; Promega) and incubated overnight at Only protein groups with at least a single unique and a 37 °C to facilitate digestion of the proteins. Peptides single Razor peptide were included. For relative protein were extracted twice with a solution containing 0.1% quantification MaxQuant LFQ intensities based on at least (v/v) trifluoric acid and 50% (v/v) acetonitrile for a single shared peptide ratio were used [19]. 20 min. The samples were dried using a SpeedVac and stored at − 20 °C until further analysis. Statistical analysis of differential protein expression To identify proteins that differ in abundance between Mass spectrometry analysis the different experimental groups an ANOVA (Kruskal– The peptides of the individual sample fractions were dis- Wallis test) was performed using the Perseus software solved in 15 μL of 0.1% (v/v) acetic acid of which 10 μL platform [20], adhering to a significance cut-off of p ≤ 0.05. was loaded onto a nano-liquid chromatography (nano- The p values were not corrected for multiple testing to in- LC) system (Eksigent). The peptides were separated cludemoreproteinsand providea broadimpressionofthe using a capillary reversed phase C18 column that had differences in the proteome. been equilibrated with 0.1% (v/v) acetic acid at a flow Conditions that were set for inclusion of CAA selective rate of 400 nL/min. The peptides were eluted by increas- proteins comprise of three approaches (A, B and C) that ing the acetonitrile concentration linearly from 5 to 40% are visualized in Fig. 2. Approach A: T-tests (two-sided, in 80 min and to 90% in 10 min, using the same flow rate. assuming unequal variances, performed using Excel Eluted peptides were transferred into the LTQ/Orbitrap (Microsoft)) were performed contrasting the three ex- MS (Thermo Scientific) by Electro Spray Ionisation (ESI). perimental groups. When there was a significant differ- The Orbitrap was operated in the range of m/z 350–2000 ence (p < 0.05) between both the control group versus at a full width at half maximum resolution of 30,000 after CAA, andthe AD groupversusCAA,aproteinwas la- accumulation to 500,000 in the LTQ with one microscan. belled as CAA specific. Approach B: If the number of The five most abundant precursor ions were selected for quantitative values in the control group was zero or fragmentation by collision-induced dissociation (CID) with one while the AD and CAA groups both had two or an isolation width of 2 Da. more quantitative values, than a t-test was performed between theAD and theCAA group. When theAD Protein inference and relative protein quantification group had zero or one quantitative values while the MaxQuant software was used for spectrum annotation, control and CAA groups both had two or more quanti- protein inference, and relative protein quantification tative values a t-test was performed between the CAA [18]. Spectra were annotated against the Uniprot human and control group. Approach C: In the case of zero or reference proteome database (version 2016_04). Enzyme single quantitative values in both the control and AD specificity was set to Trypsin/P, allowing at most two groups, proteins were included based exclusively on a missed cleavages. Carbamido-methylation of cysteine minimum of four quantitative values in the CAA group. was set as a fixed modification, and N-acetylation and Also, we included proteins with zero or single quantitative methionine oxidation were set as variable Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 5 of 19 values in the CAA group and four or more values in both staining, the sections were fixed in 100% acetone for the AD and control groups. 10 min. For paraffin sections the paraffin was removed by ANOVA (Kruskal–Wallis test) and posthoc Dunn’s washing in xylene. Next, the sections were washed in multiple comparison tests on immunoblot data and im- decreasing concentrations of ethanol (100%, 96% and 70% munohistochemical data was performed using Graphpad (v/v)). Endogenous peroxidase activity was quenched by Prism (GraphPad Software). incubating in methanol with 0.3% H O for 30 min at RT. 2 2 Next, antigen retrieval was performed by submerging the Immunoblot analysis slides in citrate buffer (pH 6) and heating in an autoclave. Protein extracts were prepared by lysis of whole occipital Primary antibodies were diluted in antibody diluent lobe tissue in reducing SDS sample buffer using a 1:20 (VWR) and incubation was performed overnight at 4 °C. tissue weight to lysis buffer ratio. Proteins were dena- All primary antibodies and corresponding dilutions used tured at 95 °C for 5 min and separated by SDS-PAGE are listed in Table 2. After incubation the sections were using precast Stain Free gradient gels (Bio-Rad) and thoroughly washed in PBS (pH 7.4) for 30 min followed transferred (40 V overnight at 4 °C) onto a 0.45 μm by incubation of an HRP-labelled secondary antibody, PVDF membrane (Merck Millipore), which was pre- Envision (DAKO) for 30 min. Again, the sections were incubated in 100% methanol. The PVDF membrane was thoroughly washed in PBS (pH 7.4) for 30 min and then incubated in Odyssey blocking buffer for 1 h and subse- incubated with DAB to visualize antibody binding. quently incubated with the primary antibody overnight. Counterstaining of the nuclei was performed by incuba- After washing in Tris-buffered saline (pH 7.5) with 0.1% tion in hematoxylin for 3 min followed by extensive (v/v) Tween-20 (TBST) for 3 × 10 min, the membrane washing in running tab water for 5 min. Next, the slides was incubated for 3 h with the secondary antibody. were dehydrated by incubation in increasing concentra- Visualization was achieved using an Odyssey imaging tions of ethanol consisting of 70% (v/v), 96% (v/v) and system using excitation wavelengths of 700 nm and 100% (v/v) ethanol. The slides were then incubated in 800 nm. Total protein load was visualized using a chemidoc xylene and mounted using Quick-D mounting medium. EZ (Bio-Rad) after electro blotting (Additional file 2:Figure A negative control was made by omission of the pri- S2) and the protein densitometric values were then used to mary antibody. Quantification of the staining was done normalize for total protein input. Primary antibodies and using ImageJ using the threshold colour plugin. dilutions are shown in Table 2. Secondary antibodies used were IRDye 800 CW Goat anti-Rabbit (LI-COR) and IRDye Results 680 conjugated Goat anti-Mouse (LI-COR) both were Selection of cases, controls and analysis of brain tissue used at a 1:7.000 dilution. All anti-bodies were diluted Three groups with a total of 20 cases were assembled in Odyssey blocking buffer (LI-COR). Quantification based on careful neuro-pathological inspection: 1) cogni- was performed using ImageJ software. tively healthy control cases (n =6) without any Aβ path- ology or tau pathology, 2) AD cases with severe Aβ plaque Immunohistochemical analysis pathology but no vascular deposits (no CAA) (n =7) and Fresh frozen or paraffin embedded human occipital tissue 3) AD cases with severe nearly pure CAA type-1 path- was cut (5 μm). For frozen tissue the sections were placed ology and a negligible amount of plaque pathology (n =7). on a SuperFrost Microscope Slide (VWR, PA, USA) and From here, these groups will be mentioned as “control”, air-dried overnight at room temperature (RT). Prior to “AD” and “CAA”, respectively. Inclusion of these cases Table 2 Antibodies used in this study Antibody Source Species Ordernr. Clone Dilution (IHC) Amyloid-beta Kind gift of Prof. Dr. Korth, Heinrich Heine Mouse IC16 1:200 University, Düsseldorf, Germany APOE Abcam Mouse ab1907 E6D7 1:3200 APOE Santa Cruz Biotechnology Mouse sc-13521 A1.4 Used for immunoblot NDP Novus Biologicals Rabbit NBP1–84769 polyclonal 1:400 NDP R&D systems Mouse MAB3014 #343711 1:800 HTRA1 R&D systems Mouse MAB2916 #275615 1:6400 APCS Statens Serum Institut, SSI Antibodies Mouse #56585 HYB281–05 1:1600 COL6A2 Abnova Mouse H00001292-M01 2C5-F2 1:3200 COL6A2 Santa Cruz Biotechnology Rabbit SC-83607 polyclonal 1:1600 Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 6 of 19 was done based on histochemical analysis using Congo- quantification of 2427 proteins in total and approximately red and additional IHC for Aβ on the occipital frozen tis- 1500 proteins identified per individual case (Additional sue intended for LC-MS-MS analysis. file 3:FigureS3),withaminimumofone tryptic We focussed our analysis on the occipital lobe as this peptide detected. All quantified proteins are listed in region is the most frequently and severely affected by Additional file 4: Table S1. CAA pathology. Tissue sections of human occipital lobe To gain insight into the global similarities and differ- from all selected cases were mounted on PEN-foil slides ences between the three groups and the individual cases and Aβ pathology was visualized using fast immunohis- an ANOVA (Kruskall Wallis) was performed. This yielded tochemistry. Grey matter tissue was isolated using LCM. 309 proteins that have a significant difference (p <0.05) in Tissue isolation from the AD cases and CAA cases was abundance between any of the experimental groups. Using focused on occipital lobe grey matter areas with severe these proteins in an unsupervised clustering analysis, three Aβ pathology, i.e. high plaque load or high CAA type-1 different expression signatures were obtained. The protein burden, respectively. This was done to selectively enrich expression signatures of the AD and CAA groups ap- the input material for the proteomics analysis for these peared largely similar, whereas both were different from types of Aβ pathology. For control cases occipital lobe the control group (Additional file 5: Figure S4A). Un- grey matter areas from the same anatomical region were supervised clustering analysis of the individual cases using selected for isolation. LCM-collected tissue samples were the 309 ANOVA-identified proteins, separated the con- lysed and proteins were separated using SDS-PAGE. trols from the disease cases (Additional file 5:FigureS4B). Each PAGE sample lane was divided into 12 fractions The CAA and AD cases were not separated on the basis and subjected to in-gel trypsin digestion (Fig. 1). of the full set of differentially expressed proteins indicating that overall their protein expression profile is largely simi- Protein quantification and global protein expression profiles lar. One CAA case (case #5) clustered with the control To identify and quantify proteins, liquid chromatography cases indicating that the protein expression profile of this followed by mass spectrometry (LC-MS-MS) was performed sample is more similar to the control cases than to other on the 20 laser-dissected tissue samples. This allowed CAA or AD cases. Visualizing the expression profile of Fig. 1 Workflow used in this study. Amyloid Beta pathology was visualized in human postmortem occipital lobe tissue. Unaffected grey matter was isolated from healthy control cases. Grey matter with high burden of Aβ pathology was isolated from the AD and CAA cases thereby isolating tissue with high plaque load or high CAA type-1 burden, respectivelyTissue was lysed and the proteins were separated using SDS-PAGE and subjected to in-gel trypsin digestion. Peptides were analysed using LC-MS-MS. A database search for protein identification and protein quantification was performed using MaxQuant software. ANOVA (Kruskall Wallis) and t-tests were performed to identify significantly regulated proteins Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 7 of 19 case #5 next to the average expression profiles of the three Approach C (Fig. 2) resulted in the identification of groups confirmed the resemblance of case #5 to the con- HLA-DRA, HLA-DQA2, HTRA1, APCS, COL6A2, trol group, but also showed several proteins that are simi- MOB2, POTEI, KIAA1468, TMF1 and SGIP1 (Fig. 3c lar in expression to the AD and or CAA groups and Table 3c) as CAA specific proteins. (Additional file 6: Figure S5A). Earlier, Case #5 was identified as having an expression profile resembling a control case. Case #5 was found positive for Alzheimer type 2 astrocytes, possibly related Identification of proteins selectively altered in CAA type-1 to high alcohol intake, and exhibited relatively low tau To identify proteins that have a significantly different pathology. Otherwise, this case showed no pathological abundance in CAA type-1 compared to both the control abnormalities when compared to the rest of the CAA and the AD group, and therefore represent unique fea- type-1 group. However, the expression of several CAA se- tures of CAA type-1, we performed student t-tests (two- lective markers that we identified was inspected for case #5. sided, assuming unequal variances) for those proteins The levels of these markers correspond well with the other where at least two quantitative values per groups were cases of the CAA group (Additional file 6:FigureS5B), in- available. When there was a significant difference (p <0. dicating that these proteins are inseparably linked to the 05) between both the control group versus CAA, and pathology of CAA type-1. In addition, although the number the AD group versus CAA, a protein was designated as of cases is too small to do valid statistics, we observed CAA-specific (Fig. 3a and Table 3a). CLU, APOE, SUCLG2, no clear relation between gender and expression of the PPP2R4, KTN1, ACTG1, TNR, COL6A3 and NFASC met markers (Additional file 7: Figure S6). these criteria. In addition, levels of CLU, APOE, SUCLG2, To determine whether the above-described approaches PPP2R4 and ACTG1 were also significantly different were indeed appropriate in selecting CAA specific pro- (p < 0.05) when comparing the AD group with the teins, we performed additional immunoblotting and im- control group. munohistochemical (IHC) analysis. After calculating the multiple testing corrected false discovery rate (FDR) only CLU was considered signifi- Confirmation of MS data using immunoblotting and cant. This is likely due to the relatively low sample size immunohistochemical analysis of this exploratory study and the high inter-individual Of the proteins described in Table 3 we selected APOE variance that is inevitably associated with the use of hu- (approach A), NDP (approach B), HTRA1, APSC and man tissue. Given the explorative nature of this study we COL6A2 (approach C), based on the fold change or spe- relaxed criteria and adhered to the uncorrected p-values cific expression in the CAA type-1 group compared to for protein inclusion. the AD and control groups, to confirm our mass spectrom- Importantly, using label-free mass spectrometry to etry results. Immunoblotting was performed on whole identify and quantify proteins, the absence of data for a tissue lysates of the same cases as used for the mass spec- number of proteins is observed. Despite great improve- trometry analysis. When comparing the CAA group with ments in the speed and sensitivity of MS analysers miss- the control group we found significant differences in NDP ing data is almost unavoidable. When quantitative data expression (Fig. 4). For APOE, APCS and COL6A2, the are absent in one group while being present in the other data showed the same trend of increased abundance in the group(s), this likely indicates differences in abundance, CAA group as the proteomics data, but the differences did which might represent interesting candidate marker pro- not reach significance. A likely explanation for this is the teins. Therefore, absence of data in one or more patient higher variation of expression of these proteins in the tissue groups required 2 additional approaches to also consider used for immunoblotting, which in contrast to the mass these proteins in this study. An overview of the 3 comple- spectrometry exploratory analysis, was not selectively menting strategies for protein inclusion is shown in Fig. 2 enriched for pathological burden using LCM, and in- and a complete description is present in the methods sec- stead included white matter, leptomeningeal vessels and tion. Note that any given protein is only considered using a grey matter with a lower pathological burden. To un- single approach as these approaches are mutually exclusive. equivocally demonstrate CAA related expression, we Using approach B (Fig. 2), proteins with a significant turned to IHC analysis of these same proteins, which, difference were included, and APP, UBLCP1, SRI, NDP, in contrast to immunoblotting, allows region specific PNP, C1orf123, DHX15, SYNPO, TPM1, CADPS2 and analysis similar to the LCM-LC-MS-MS analysis. For SERPINA3 (Fig. 3b and Table 3b) were identified as pro- this a separate cohort was used consisting of cognitively teins selectively present in CAA type-1. Peptide data on healthy control cases (n = 8) without any Aβ or tau APP indicates that quantification was based on two pep- pathology, 2) AD cases with severe Aβ plaque path- tides in which the most abundantly detected peptide ology but no vascular deposits (no CAA) (n =8) and 3) (LVFFAEDVGSNK) is part of Aβ. AD cases with severe CAA type-1 pathology (n =6). Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 8 of 19 Table 3 A, B and C Proteins identified as selectively altered in CAA type-1 Gene Protein P-val C vs P-val AD vs FDR C vs FDR AD vs FC C vs FC AD vs # detections # detections # detections CAA CAA CAA CAA CAA CAA Control AD CAA A. Significant CAA versus control and CAA versus Alzheimer's disease CLU Clusterin;Clusterin beta chain; 0.000 0.000 0.001 0.007 4.47 2.33 6 7 7 Clusterin alpha chain APOE Apolipoprotein E 0.001 0.001 ns ns 4.97 2.11 6 7 7 SUCLG2 Succinyl-CoA ligase [GDP- 0.002 0.002 ns ns 2.17 0.61 5 7 7 forming] subunit beta, mitochondrial PPP2R4 Serine/threonine-protein 0.026 0.010 ns ns 2.18 0.80 6 7 7 phosphatase 2A activator KTN1 Kinectin 0.015 0.021 ns ns 0.34 0.40 2 2 3 ACTG1 Actin, cytoplasmic 2;Actin, 0.000 0.035 ns ns 0.80 0.90 6 7 7 cytoplasmic 2, N-terminally processed TNR Tenascin-R 0.007 0.017 ns ns 0.82 0.82 6 7 7 COL6A3 Collagen alpha-3(VI) chain 0.016 0.024 ns ns 7.79 4.95 3 5 7 NFASC Neurofascin 0.028 0.040 ns ns 0.86 0.87 6 7 7 B. Significant CAA versus control or Alzheimer’s disease and ≤1 detection in other group APP Amyloid beta A4 protein;N- NA 0.000 NA ns NA 6.65 1 7 7 APP;Soluble APP-alpha;Soluble APP-beta; UBLCP1 Ubiquitin-like domain- NA 0.007 NA ns NA 1.96 1 2 4 containing CTD phosphatase 1 SRI Sorcin NA 0.011 NA ns NA 0.25 1 3 5 NDP Norrin NA 0.020 NA ns NA 5.16 0 4 7 PNP Purine nucleoside NA 0.030 NA ns NA 1.72 0 3 3 phosphorylase C1orf123 UPF0587 protein C1orf123 NA 0.047 NA ns NA 0.71 0 6 6 DHX15 Putative pre-mRNA-splicing 0.005 NA ns NA 0.44 NA 3 0 5 factor ATP- dependent RNA helicase DHX15 SYNPO Synaptopodin 0.015 NA ns NA 0.44 NA 4 1 4 TPM1 Tropomyosin alpha-1 chain 0.021 NA ns NA 0.37 NA 3 1 3 CADPS2 Calcium-dependent secretion 0.042 NA ns NA 1.22 NA 2 1 3 activator 2 Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 9 of 19 Table 3 A, B and C Proteins identified as selectively altered in CAA type-1 (Continued) Gene Protein P-val C vs P-val AD vs FDR C vs FDR AD vs FC C vs FC AD vs # detections # detections # detections CAA CAA CAA CAA CAA CAA Control AD CAA SERPINA3 Alpha-1- 0.042 NA ns NA −0.86 NA 2 1 3 antichymotrypsin;Alpha-1- antichymotrypsin His-Pro-less C. ≤1 detection in control and Alzheimer's disease and ≥4 in CAA OR ≤1 in CAA and ≥4 in Alzheimer’s disease and control HLA-DRA;HLA- HLA class II histocompatibility NA NA NA NA NA NA 0 1 7 DQA2 antigen, DR alpha chain;HLA class II histocompatibility antigen, DQ alpha 2 chain HTRA1 Serine protease HTRA1 NA NA NA NA NA NA 0 1 7 APCS Serum amyloid P- NA NA NA NA NA NA 0 1 6 component;Serum amyloid P-component(1–203) COL6A2 Collagen alpha-2(VI) chain NA NA NA NA NA NA 0 1 5 MOB2 MOB kinase activator 2 NA NA NA NA NA NA 1 1 5 POTEI POTE ankyrin domain family NA NA NA NA NA NA 1 0 4 member I KIAA1468 LisH domain and HEAT NA NA NA NA NA NA 1 0 4 repeat-containing protein KIAA1468 TMF1 TATA element modulatory NA NA NA NA NA NA 0 1 4 factor SGIP1 SH3-containing GRB2-like NA NA NA NA NA NA 4 4 1 protein 3-interacting protein 1 Proteins were found using the three different selection methods as described in Fig. 2 AD Alzheimer’s disease, CAA cerebral amyloid angiopathy, FDR false discovery rate, NA not applicable, NS not significant, FC fold change Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 10 of 19 Fig. 2 Three strategies used to select proteins that are differentially expressed in CAA type-1 compared to control and AD brains. Criteria of each of the selection strategies are specified, numbers of resulted proteins indicated, and selected proteins are listed in tables and figures as indicated First Aβ pathology was visualized and its presence was vessels in the CAA cases and showed plaque pathology in confirmed in AD and CAA type-1 cases showing plaques the AD cases without CAA. Control cases were all nega- and vascular Aβ pathology, respectively (Fig. 5b and c). tive for HTRA1 (Fig. 5j-l). IHC for APOE resulted in pro- Then, IHC analysis was performed to gain information on nounced staining of the vasculature in CAA cases and the localization of the selected proteins. IHC for NDP appeared related to compact deposits as well as more dif- showed pronounced immunoreactivity in CAA type-1 fuse dyshoric deposits. Also immunoreactivity of APOE cases that appeared associated to the vasculature. NDP was observed in the AD cases related to the Aβ plaques, immunostaining in CAA, appeared to be associated with although the staining was less intense than that related to both compact Aβ depositions as well as more diffuse the vascular amyloid in the CAA cases (Fig. 5p-r). APCS staining in the parenchyma in cases that exhibit dyshoric IHC illustrated the presence of this protein in relation Aβ deposits. Staining was more pronounced related to ca- with both diffuse and compact Aβ pathology in both the pillaries compared to larger vessels. The AD cases with CAA and AD group. However, staining related to the plaques were nearly devoid of immunoreactivity, controls plaque pathology was less intense than that related to the did not show any immunoreactivity for NDP (Fig. 5d-f). vascular Aβ pathology (Fig. 5m-o). Different antibodies against NDP showed similar results For quantification of the IHC, images were obtained at (data not shown). sites that, for the AD and CAA cases, had high Aβ COL6A2 IHC showed some immunoreactivity in con- pathological burden in nearby sections of the same tis- trol and AD cases which was restricted to leptomeningeal sue block. The percentage of positively stained pixels vessels (Additional file 8: Figure S7) and a few large vessels over a total of 5 images from each case was determined. in the brain tissue. In CAA type-1, immunoreactivity for Although this method is semi-quantitative, it allowed a COL6A2 was highly increased and includes brain capillar- region-specific analysis in line with the tissue obtained ies and larger vessels. Immunoreactivity was mostly by laser capture dissection that was at the basis of the associated with the endothelium and/or the adventitia original mass spectrometry analysis. Using IHC quantifi- (Fig. 5g-i). Similar results were obtained using two differ- cation we observed strongly increased immunoreactivity ent antibodies for COL6A2 (data not shown). HTRA1 in CAA type-1 cases for NDP and COL6A2, and moder- IHC showed clear overlap with Aβ in both AD and CAA. ate increases for APCS, HTRA1 and APOE, confirming Compact and diffuse staining was observed related to the the MS results (Fig. 6). Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 11 of 19 Fig. 3 Relative abundance of proteins with altered expression in CAA type-1 compared to Alzheimer’s disease cases and controls as determined by MS. Three groups of selected proteins (panels a-c), with altered levels (MS-derived, log2 LFQ intensity values) in CAA type-1 compared to the control groups and the AD groups. Selection criteria are specified in Fig. 2. Gene names are indicated Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 12 of 19 Fig. 4 Immunoblotting analysis of proteins with altered expression in CAA type-1. a Immunoblotting was performed for, NDP, COL6A2, APCS and APOE on occipital lobe tissue lysates of non-demented controls (N) AD cases (A) and CAA cases (C). Immunoreactivity was observed at the correct molecular weight for each protein. b A significant difference (p < 0.05) was found in the expression of NDP in the CAA group when compared to control but not when compared to the AD group. For COL6A2, APCS and APOE significance was not reached between any of the groups using this technique. Data are expressed as mean ± SEM Specificity in relation to other small vessel diseases vessels (Fig. 7g). IHC for APOE and APCS presented a To assess the specificity of Aβ, NDP, COL6A2, APCS highly positive, staining that appeared similar to the Aβ and APOE in relation to other small vessel diseases we staining. performed additional IHC on cases that present various The PrP-CAA tissue was confirmed negative for Aβ types of vascular defects, i.e., cotton wool plaque path- pathology (Fig. 7k) and showed positive for Prp (data ology, prion CAA, cerebral autosomal dominant arterio- not shown). NDP immunoreactivity was highly increased pathy with subcortical infarcts and leukoencephalopathy and localised to the affected vessels. NDP staining was (CADASIL), hypertension related small vessel disease more pronounced around affected capillaries than at af- and Cathepsin A-related arteriopathy with strokes and fected larger vessels. COL6A2 immunoreactivity was leukoencephalopathy (CARASAL). IHC was performed clearly present around capillaries and larger affected ves- on sections that exhibited the relevant pathological char- sels, of which the vascular pathology was clearly ob- acteristics of each disease including an additional CAA served using haematoxylin. Although co-occurring in type 1 case. Immunoreactivity for Aβ, NDP, COL6A2, largely the same vessels COL6A2 did not generally co- APOE and APCS was confirmed in the CAA type-1 case localize with the deposits, but instead localized more in- (Fig. 7a-e). Immunoreactivity for these proteins was also ternally as a component of the basal membrane. APOE assessed and confirmed in case exhibiting hereditary and APCS were also highly present and co-localized with cerebral haemorrhage with amyloidosis Dutch type the Prp deposits. (HCHWA-D), which is a heredity form of CAA-type-1 In the white matter no immunoreactivity was seen for (Additional file 9: Figure S8). Aβ or any of the marker proteins (Fig. 7p-t) in the con- In tissue with cotton wool plaques the pathology was trol case. The CADASIL case showed characteristic confirmed using IHC for Aβ displaying pathology pathology in the white matter, thickened vessel walls, in around capillaries and larger vessels with dyshoric the white matter as is common with this condition. No changes extending deep into the parenchyma (Fig. 7f). Aβ pathology was present (Fig. 7u) Mild immunoreactiv- Intense NDP immunoreactivity was observed related to ity for NDP was present and of the assessed proteins capillaries and the dyshoric changes and to a lesser ex- COL62A staining was most pronounced. APOE and tend directly lining the larger vessels. COL6A2 showed APCS also displayed mild immunoreactivity related to intense immunoreactivity lining the capillaries and larger the affected vessels (Fig. 7v-y). Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 13 of 19 Fig. 5 Immunohistochemical analysis of selected proteins. Representative images were taken. Aβ pathology was visualized. a The control case does not have any Aβ pathology. b Plaque pathology is confirmed in the AD case and (c) CAA type-1 pathology is confirmed in the CAA type-1 case. d-f Extensive NDP immunoreactivity is observed in the CAA type-1 cases whereas absent in both control and AD cases without CAA. g-i COL6A2 immunoreactivity is hardly observed in the control and AD cases, however, extensive immunoreactivity is observed in the CAA type cases and includes both capillaries and large vessels. j-l Immunoreactivity for HTRA1 is absent in control tissue, however, is observed both related to plaque pathology and CAA at comparable intensity. m-o Immunoreactivity for APCS is absent in control cases but is observed both related to plaque and CAA type-1 pathology. However, the intensity of the staining observed in the AD cases is considerably less. p-r Also, APOE immunoreactivity is observed related to both plaque and CAA type-1 pathology, yet its intensity in CAA type-1 is far greater. Scale bar, 100 μm in images A to R. Scale bar in image (C`) 10 μm and in all zoomed images, which are marked with a grave accent (`) In hypertension related small vessel disease the presence in CAA (type-1 and cotton wool) and Prp-CAA. Involve- of COLA6A2 was most prominent while immunoreactivity ment of NDP, APOE and APCS in other small vessel dis- for NDP, APOE and APCS was low but present (Fig. 7z-ad). eases is varying from non (CARASAL) to mild Interestingly IHC analysis of the CARASAL cases showed (CADASIL). Importantly, NDP is explicitly suitable to evi- only the prominent presence of COL6A2 in affected vessels dently separate CAA from Aβ plaque pathology (Table 4). while NDP, APOE and APCS were absent. Taken our data together, from the tested panel of pro- Discussion teins, we recognize COL6A2 as a general small vessel dis- One of the most prevalent cerebro-vascular diseases in ease marker. NDP, APOE and APCS are most prominent the elderly is sporadic CAA, characterized by vascular Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 14 of 19 Fig. 6 Semi-quantitative analysis of immunohistochemical data of proteins with altered expression in CAA type-1. Immunohistochemical stainings were quantified by measuring the percentage of pixels that showed positive immunoreactivity. Significance was calculated using a one-way ANOVA (Kruskal-Wallis test) and posthoc Dunn’s Multiple Comparison Test. A significant increase (p < 0.05) in immunoreactivity in the CAA group compared to both control and AD groups was observed for NDP and COL6A2. For APOE, APCS and HTRA1 significant differences were only found when comparing CAA with control, but not with the AD group. Data are expressed as mean ± SEM deposition of amyloid-beta protein. CAA can occur as plaque pathology, in accordance with previous findings an isolated disease or as part of the pathology in AD. [24, 25]. Interestingly, increased levels of CLU have been Several studies have indicated CAA as an important reported in the plasma of CAA patients diagnosed ac- cause of cognitive decline [10, 21, 22]. Currently, there is cording the modified Boston criteria [26]. no treatment for CAA and its presence cannot be diag- Two recent proteomics studies focussed on CAA ana- nosed pre-mortem. Therefore insight in the pathogenic lysing leptomeningeal vessels [27] and leptomeningeal mechanisms and the need for biomarkers are urgent. vessel combined with neocortical arterioles [28]. Some We performed an exploratory laser dissection-assisted similarities were found with these studies, such as the LC-MS-MS analysis of AD brain tissue exhibiting severe increase in CLU and APOE. As expected also several dif- CAA type-1 pathology, AD brain tissue without appar- ferences exist as our analysis focussed on grey matter of ent involvement of CAA, and control brain tissue with- CAA type-1 using micro dissected tissue that is enriched out AD related pathology. We show that the proteome for areas with very high capillary associated Aβ path- of CAA type-1 is different from that of parenchymal ology. These differences might indicate that other mech- plaque pathology in AD, which led to the identification anisms are involved in the pathogenesis in CAA related of proteins selectively associated with CAA. to capillaries compared to larger vessels, e.g., the find- Next to identification of new CAA selective proteins, ings of NDP and COL6A2. In contrast to these previous this study also confirmed the presence proteins already studies we compared CAA cases with both control and known to be involved in CAA pathology, e.g., CLU, AD cases with plaque pathology and without CAA. APOE and APCS [23, 24]. Interestingly, CLU was de- There are many similarities in the response to CAA and tected in all samples included in the study and its abun- AD and these proteins and processes can be cancelled dance was sufficient to completely separate the CAA out against each other, allowing CAA selective proteins group from both the control group and the AD group. to become apparent. Moreover, CLU, APOE and APCS were markedly in- Importantly, we identified potential new key players in creased in AD compared to controls and APOE and the development of CAA. NDP is found highly upregu- APCS showed moderate immunoreactivity related to lated in CAA type-1, cotton wool Aβ pathology and Prp- Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 15 of 19 Fig. 7 Expression of CAA type-1 markers in other small vessel diseases. IHC for Aβ, NDP, COL6A2, APCS and APOE was performed. In a CAA type- 1 case immunoreactivity for all marker proteins is confirmed (a-e). Also immunoreactivity is seen for all markers in the cotton wool case and Aβ pathology was confirmed (f). For NDP, APOE and APCS immunoreactivity is also seen localizing to severe dyshoric angiopathy (g, i and j). COL6A2 immunoreactivity is restricted to the vessel wall (h). In de Prp-CAA case Aβ pathology was absent (k). Extensive immunoreactivity was observed for NDP, COL6A2, APOE and APCS (l-o). No immunoreactivity was observed for Aβ, NDP, COL6A2, APOE and APCS in the white matter of control tissue (p-t). In the CADASIL case no Aβ pathology was present (u). Mild immunoreactivity for NDP (v) COL62A staining was most pronounced (w). APOE and APCS also displayed mild immunoreactivity related to the affected vessels (x, y). In hypertension related small vessel disease no Aβ was detected (z) immunoreactivity of COLA6A2 was moderate (ab) while immunoreactivity for NDP, APOE and APCS were low but present (aa, ac and ad). In the CARASAL case only prominent immunoreactivity of COL6A2 was seen in affected vessels (ag) while NDP, APOE and APCS were absent (af, ah and ai). Scale bar in (a) indicates 100 μm Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 16 of 19 Table 4 Scoring of relative immunoreactivity of 5 CAA-1 positive highly increased in CAA type-1. COL6A2 immunoblot- markers in different small vessel diseases and AD plaque ting did not show a significant difference between the pathology experimental groups. This can be explained by the inclu- Aβ NDP COL6A2 APOE APCS sion of leptomeningeal vessels, that express high levels CAA type-1 3 3 3 3 3 of COL6A2 in all cases in varying amounts in the tissue lysates used for immunoblotting, as shown using IHC. Cotton wool 3 3 3 3 3 COL6A2 was also found increased in the other small Prp-CAA 0 3 3 3 3 vessel diseases that were included for IHC analysis. This Control white matter 0 0 0 0 0 indicates that COL6A2 might be used as valuable indica- CADASIL 0 2 2 1 1 tor of vascular pathology in general, but not specific for Hypertension 0 1 2 1 1 CAA type pathology. COL6A2 is a non-fibril collagen CARASAL 0 0 2 0 0 and COL6 isoforms are present in various tissues includ- ing the vasculature [37]. COL6A2 encodes one of the AD plaques 3 0 0 2 2 three alpha chains of type VI collagen, which is found in 0: no immunoreactivity, 1: mild immunoreactivity, 2: moderate immunoreactivity, 3: extensive immunoreactivity most connective tissues. Type VI collagen anchors endo- thelial basement membranes by interacting with type IV CAA, and localizes around the affected vasculature. collagen [38]. In the brain, collagen VI was shown neu- NDP immunoreactivity was only mildly increased in roprotective and its expression increased in animal CADASIL and hypertension related small vessel disease. models of AD [39]. In addition, these diseases affect different anatomical re- HTRA1 is a trypsin-like serine protease which was de- gions, clearly identifiable with imaging studies, and tected in all CAA samples, with a single value in the AD present a distinct clinical picture, leaving NDP a promis- group and zero quantitative values in the control group. ing biomarker for CAA. Using immunohistochemistry we found a significant dif- NDP is a small, secreted protein with a molecular ference with the control groups but not with the AD weight of approximately 15 kD. It has important func- group as HTRA1 marks normal plaque pathology as tion in the formation of the brain vasculature during de- well. HTRA1 is relevant in neurodegeneration as this velopment and in maintenance of a proper functioning protease is involved in the degradation of APP and Aβ BBB [29]. In the adult brain the NDP gene is primarily [40]. In addition, HTRA1 was found to degrade APOE4 expressed by astrocytes [30]. NDP activates the canon- more efficiently than APOE3 and the presence of ical Wnt/β-catenin signalling pathway via the frizzled APOE4 reduces digestion of MAPT by HTRA1 [41]. (Fzd)4/low-density lipoprotein receptor-related protein Moreover, mutations in HTRA1 are the cause of the her- (Lrp)5/6 receptor complex [31]. editary small vessel disease CARASIL (cerebral auto- In neural progenitor cells (NPCs) derived from FAD somal recessive arteriopathy with subcortical infarcts mutant PSEN1 subjects it was found that NDP mRNA is and leukoencephalopathy) [42, 43]. upregulated, but no increase in mRNA was found in AD As part of our LC-MS-MS exploratory study we iden- human temporal lobe [32]. In the retina NDP was found tified other proteins that are potentially interesting for to promote regrowth of capillaries and formation of additional research in relation to CAA, but were not intra-retinal vessels after oxygen-induced retinal damage specifically followed up in this study. For instance, we [33]. Mice overexpressing NDP, had significantly less observed significant high levels of HLA-DR/HLA-DQ in vascular loss following oxygen exposure. Mutations in CAA type-1. This protein is associated with inflamma- the NDP gene result in Norrie disease, which is primar- tion and high numbers of activated microglia [4]. ily an eye disease that leads to blindness. Interestingly, PNP for which a single nucleotide polymorphism was 30–50% of these patients display developmental delay, found to be associated with faster progression of AD intellectual disability, behavioural abnormalities, or [44]. Its relation to CAA is yet unknown. psychotic-like features [34, 35]. In addition, NDP has SUCLG2 which is involved in clearance of Aβ1–42 been shown to protect neurons against excitotoxicity in- [45] was found increased in CAA compared to control duced by NMDA [36]. NDP seems to have protective and even higher levels were observed in the AD cases. properties for both endothelial cells and neurons, but APP was identified with increased levels in CAA. Pep- whether NDP upregulation is beneficial in the context of tide data indicates that the most abundantly detected CAA pathology is unknown. peptide is a part of Aβ, however this analysis cannot dis- Our proteomics data show COL6A2 expression in criminate between Aβ or APP. CAA cases, which is supported by a strong increase in The proteins in this study, that show selective associ- COL6A2 immunoreactivity in the affected brain paren- ation with CAA type-1 pathology might serve as poten- chyma. In addition to COL6A2 we also found COL6A3 tial CAA type-1 biomarkers in patients. As larger vessels Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 17 of 19 are also positive for the markers that were assessed using of the control group (2nd row), AD group (3rd row) and the CAA group IHC, these markers might also be relevant in CAA type- (4th row). Green, expression below the overall mean; red, above the overall mean. The expression profile of case #5 is largely similar to that of 2, although in the case of NDP the intensity of immuno- the control groups but some proteins show a similar expression as in the reactivity is less in larger vessels compared to affected AD and/or CAA groups. (B) Expression values (LFQ values) of several CAA capillaries in CAA. specific proteins identified in this study with case #5 indicated as empty triangle pointing down. Case #5 does not differ from the CAA group in CAA selective markers might be used for pathological these markers. (TIF 1835 kb) assessment of the severity of CAA. The association of Additional file 7: Figure S6. Protein expression of males versus Aβ with the vasculature, and in particular capillaries, is females. Quantitative data on several CAA selective data was plotted with not always obvious in thin microscopic sections. Also, males represented as triangles and females as dots. No clear relationship between gender and protein abundance was observed. (TIF 24739 kb) the use of these proteins as potential diagnostic markers Additional file 8: Figure S7. Immunoreactivity for COL6A2 is equally should be explored. present in leptomeningeal vessels in control, AD and CAA tissue. (TIF 3794 kb) The need for a biomarker for CAA is urgent, in part Additional file 9: Figure S8. Immunohistochemistry of Amyloid-beta, for (early) diagnosis of CAA, but also for stratification of NDP, COL6A2, APOE and APCS on multiple brain regions of a HCHWA-D CAA patients involved in clinical trials for AD. For instance, type-1 case. Brain tissue of a case exhibiting a hereditary form of CAA type-1 was analyzed by immunohistochemistry of Amyloid-beta, NDP, COL6A2, APOE anti-amyloid immunotherapies in development may war- and APCS. Aβ pathology was confirmed and immunoreactivity associated with rant separation of AD patients with or without CAA be- CAA type-1 pathology was found present for all markers. Scale bar in upper cause of expected side effects associated with CAA, left picture represents 100 μm. (TIF 34988 kb) including vasogenic oedema and cerebral microhemor- rhages [46, 47]. In addition, these markers would help to Acknowledgements The authors thank the Netherlands Brain Bank (Amsterdam, the Netherlands) improve the assessment of the safety of anticoagulation for supplying human brain tissue. The authors want to thank Will Hermsen, therapy in patients with CAA as they increase the risk of University Medical Center Utrecht, for performing the immunohistochemistry intracerebral haemorrhage [48]. on the prion tissue. This work was financially supported by Amsterdam Neuroscience and Alzheimer Nederland, grant number AN-16054. David Hondius was supported by the CAVIA project (nr. 733050202), which has Conclusion been made possible by ZonMW, part of the Dutch national ‘Deltaplan for Dementia’: zonmw.nl/dementiaresearch”. In conclusion, we present a set of marker proteins con- taining known and new markers representing valuable Authors’ contributions tools for both clinical and neuropathological diagnosis DCH, KWL, ABS and AJMR designed the experiments. DCH, KNE, THJM, RCvdS performed the experiments. DCH, KWL, JJMH, ABS and AJMR which can contribute to studies investigating the role of interpreted the results. AJMR and MB provided samples and performed the CAA in AD pathology. In addition to their use as bio- pathological characterization. DCH was responsible for writing of the markers, the newly found proteins might be further in- manuscript. KWL, JJMH, MB, PvN, ABS, AJMR made intellectual contributions and contributed to the writing of the manuscript. All authors read and vestigated to increase our understanding of etiology and approved the final manuscript. disease mechanism related to CAA, and ultimately may be used as therapeutic targets. Competing interests A selection of proteins including, but not limited to, NDP, CLU, APOE, HTRA1, APCS, COL6A2 and COL6A3 are part of the patent application P113281EP00. Additional files Publisher’sNote Additional file 1: Figure S1. Coomassie blue staining of the SDS PAGE Springer Nature remains neutral with regard to jurisdictional claims in gels containing the microdissected tissue lysates. (TIF 478 kb) published maps and institutional affiliations. Additional file 2: Figure S2. Total protein fluorescent signal from blots Received: 22 March 2018 Accepted: 23 April 2018 used for immunoblot analysis. Total protein load was visualized using a chemidoc EZ (Bio-Rad) after electroblotting and used to obtain densitometric values which were then used to normalize for total protein References input. (TIF 553 kb) 1. Attems J, Jellinger K, Thal DR, Van Nostrand W (2011) Review: sporadic Additional file 3: Figure S3. Number of proteins detected per individual cerebral amyloid angiopathy. Neuropathol Appl Neurobiol 37:75–93. https:// case. Proteins were quantified based on a minimum of one peptide and doi.org/10.1111/j.1365-2990.2010.01137.x adhering to an FDR of < 0.01. (TIF 19662 kb) 2. Rudolf Thal D, Sue GriYn WT, I de Vos RA, Ghebremedhin E (2008) Cerebral Additional file 4: Table S1. Complete dataset, containing log2 amyloid angiopathy and its relationship to Alzheimer’s disease. Acta transformed quantitative values (LFQ values) of all quantified proteins Neuropathol 115:599–609. https://doi.org/10.1007/s00401-008-0366-2 per individual case. (XLSX 665 kb) 3. Attems J (2005) Sporadic cerebral amyloid angiopathy: pathology, clinical implications, and possible pathomechanisms. Acta Neuropathol 110:345–359. Additional file 5: Figure S4. Clustering analysis of experimental groups https://doi.org/10.1007/s00401-005-1074-9 and individual cases. Clustering analysis and heat maps of the different 4. Richard E, Carrano A, Hoozemans JJ, Van Horssen J, Van Haastert ES, experimental groups (A) and individual cases (B) based on proteins with Eurelings LS, De Vries HE, Thal DR, Eikelenboom P, Van Gool WA, a significant difference (ANOVA, p < 0.05) in expression between any of Rozemuller AJM (2010) Characteristics of dyshoric capillary cerebral the groups. (TIF 709 kb) amyloid angiopathy. J Neuropathol Exp Neurol 69:1158–1167. https:// Additional file 6: Figure S5. Protein expression of CAA case #5 relative doi.org/10.1097/NEN.0b013e3181fab558 to the experimental groups and individual cases. (A) On the left the 5. van Veluw SJ, Kuijf HJ, Charidimou A, Viswanathan A, Biessels GJ, Rozemuller expression profile of case #5 compared to the average expression profile AJM, Frosch MP, Greenberg SM (2016) Reduced vascular amyloid burden at Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 18 of 19 microhemorrhage sites in cerebral amyloid angiopathy. Acta Neuropathol: 24. Verbeek MM, Otte-Höller I, Veerhuis R, Ruiter DJ, De Waal RMW (1998) 1–7. https://doi.org/10.1007/s00401-016-1635-0 Distribution of Aβ-associated proteins in cerebrovascular amyloid of 6. Weller RO, Nicoll J a R (2003) Cerebral amyloid angiopathy: pathogenesis Alzheimer’s disease. Acta Neuropathol 96:628–636. https://doi.org/10. and effects on the ageing and Alzheimer brain. Neurol Res 25:611–616. 1007/s004010050944 https://doi.org/10.1179/016164103101202057 25. Zhan SS, Veerhuis R, Kamphorst W, Eikelenboom P (1995) Distribution of beta amyloid associated proteins in plaques in Alzheimer’s disease and in 7. Bakker ENTP, Bacskai BJ, Arbel-Ornath M, Aldea R, Bedussi B, Morris AWJ, the non-demented elderly. Neurodegeneration 4:291–297 Weller RO, Carare RO (2016) Lymphatic clearance of the brain: perivascular, Paravascular and significance for neurodegenerative diseases. Cell Mol 26. Montañola A, de Retana SF, López-Rueda A, Merino-Zamorano C, Penalba A, Neurobiol 36:181–194. https://doi.org/10.1007/s10571-015-0273-8 Fernández-Álvarez P, Rodríguez-Luna D, Malagelada A, Pujadas F, Montaner 8. Weller RO, Subash M, Preston SD, Mazanti I, Carare RO (2008) Perivascular J, Hernández-Guillamon M (2016) ApoA1, ApoJ and ApoE plasma levels and drainage of amyloid-?? Peptides from the brain and its failure in cerebral genotype frequencies in cerebral amyloid Angiopathy. NeuroMolecular Med amyloid angiopathy and Alzheimer’s disease. In: Brain Pathol, pp 253–266 18:99–108. https://doi.org/10.1007/s12017-015-8381-7 9. Attems J, Jellinger KA (2004) Only cerebral capillary amyloid angiopathy 27. Manousopoulou A, Gatherer M, Smith C, Nicoll JAR, Woelk CH, Johnson M, correlates with Alzheimer pathology?A pilot study. Acta Neuropathol 107: Kalaria R, Attems J, Garbis SD, Carare RO (2017) Systems proteomic analysis 83–90. https://doi.org/10.1007/s00401-003-0796-9 reveals that clusterin and tissue inhibitor of metalloproteinases 3 increase in 10. Eurelings LSM, Richard E, Carrano A, Eikelenboom P, van Gool WA, leptomeningeal arteries affected by cerebral amyloid angiopathy. Rozemuller AJM (2010) Dyshoric capillary cerebral amyloid angiopathy Neuropathol Appl Neurobiol 43:492–504. https://doi.org/10.1111/nan.12342 mimicking Creutzfeldt–Jakob disease. J Neurol Sci 295:131–134. https://doi. 28. Inoue Y, Ueda M, Tasaki M, Takeshima A, Nagatoshi A, Masuda T, Misumi Y, Kosaka T, Nomura T, Mizukami M, Matsumoto S, Yamashita T, Takahashi H, org/10.1016/j.jns.2010.04.020 Kakita A, Ando Y (2017) Sushi repeat-containing protein 1: a novel disease- 11. Thal DR, Ghebremedhin E, Orantes M, Wiestler OD (2003) Vascular pathology associated molecule in cerebral amyloid angiopathy. Acta Neuropathol 134: in Alzheimer disease: correlation of cerebral amyloid Angiopathy and 605–617. https://doi.org/10.1007/s00401-017-1720-z arteriosclerosis/Lipohyalinosis with cognitive decline. J Neuropathol Exp Neurol 29. Engelhardt B, Liebner S (2014) Novel insights into the development and 62:1287–1301. https://doi.org/10.1093/jnen/62.12.1287 maintenance of the blood–brain barrier. Cell Tissue Res 355:687–699. 12. Braak H, Braak E (1991) Neuropathological stageing of Alzheimer-related https://doi.org/10.1007/s00441-014-1811-2 changes. Acta Neuropathol 82:239–259 13. Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, Vogel FS, 30. Ye X, Smallwood P, Nathans J (2011) Expression of the Norrie disease gene Hughes JP, van Belle G, Berg L (1991) The consortium to establish a registry (Ndp) in developing and adult mouse eye, ear, and brain. Gene Expr for Alzheimer’s disease (CERAD): part II. Standardization of the Patterns. https://doi.org/10.1016/j.gep.2010.10.007 neuropathologic assessment of Alzheimer’s disease. Neurology 41:479–479. 31. Xu Q, Wang Y, Dabdoub A, Smallwood PM, Williams J, Woods C, Kelley MW, https://doi.org/10.1212/WNL.41.4.479 Jiang L, Tasman W, Zhang K, Nathans J (2004) Vascular development in the retina and inner ear: control by Norrin and Frizzled-4, a high-affinity ligand- 14. Thal DR, Rüb U, Orantes M, Braak H (2002) Phases of Aβ-deposition in the receptor pair. Cell 116:883–895. https://doi.org/10.1016/S0092-8674(04)00216-8 human brain and its relevance for the development of AD. Neurology 58: 1791–1800. https://doi.org/10.1212/WNL.58.12.1791 32. Sproul AA, Jacob S, Pre D, Kim SH, Nestor MW, Navarro-Sobrino M, Santa- 15. Montine TJ, Phelps CH, Beach TG, Bigio EH, Cairns NJ, Dickson DW, Maria I, Zimmer M, Aubry S, Steele JW, Kahler DJ, Dranovsky A, Arancio O, Duyckaerts C, Frosch MP, Masliah E, Mirra SS, Nelson PT, Schneider JA, Thal Crary JF, Gandy S, Noggle SA (2014) Characterization and molecular DR, Trojanowski JQ, Vinters HV, Hyman BT (2012) National institute on profiling of PSEN1 familial alzheimer’s disease iPSC-derived neural aging-Alzheimer’s association guidelines for the neuropathologic assessment progenitors. PLoS One. https://doi.org/10.1371/journal.pone.0084547 of Alzheimer’s disease: a practical approach. Acta Neuropathol 123:1–11. 33. Ohlmann A, Seitz R, Braunger B, Seitz D, Bösl MR, Tamm ER (2010) Norrin https://doi.org/10.1007/s00401-011-0910-3 promotes vascular regrowth after oxygen-induced retinal vessel loss and 16. Verwey NA, Hoozemans JJM, Korth C, van Royen MR, Prikulis I, Wouters D, suppresses retinopathy in mice. J Neurosci 30 HAM T, van Haastert ES, Schenk D, Scheltens P, Rozemuller AJM, Blankenstein 34. Braunger BM, Tamm ER (2012) The different functions of Norrin. Adv Exp MA, Veerhuis R (2013) Immunohistochemical characterization of novel med biol. https://doi.org/10.1007/978-1-4614-0631-0_86 monoclonal antibodies against the N-terminus of amyloid β-peptide. Amyloid 35. Sims KB (1993) NDP-related retinopathies. University of Washington, Seattle 20:179–187. https://doi.org/10.3109/13506129.2013.797389 36. Seitz R, Hackl S, Seibuchner T, Tamm ER, Ohlmann A (2010) Norrin mediates 17. Hondius DC, Van Nierop P, Li KW, Hoozemans JJM, Van Der Schors RC, neuroprotective effects on retinal ganglion cells via activation of the Wnt/- Van Haastert ES, Van Der Vies SM, Rozemuller AJM, Smit AB (2016) catenin signaling pathway and the induction of neuroprotective growth Profiling the human hippocampal proteome at all pathologic stages of factors in Muller cells. J Neurosci 30:5998–6010. https://doi.org/10.1523/ Alzheimer’s disease. Alzheimers Dement 12:654–668. https://doi.org/10. JNEUROSCI.0730-10.2010 1016/j.jalz.2015.11.002 37. Ricard-Blum S (2011) The collagen family. Cold Spring Harb Perspect 18. Cox J, Mann M (2008) MaxQuant enables high peptide identification rates, Biol 3:a004978–a004978. https://doi.org/10.1101/cshperspect.a004978 individualized p.P.B.-range mass accuracies and proteome-wide protein 38. Kuo HJ, Maslen CL, Keene DR, Glanville RW (1997) Type VI collagen anchors quantification. Nat Biotechnol 26:1367–1372. https://doi.org/10.1038/nbt.1511 endothelial basement membranes by interacting with type IV collagen. 19. Cox J, Hein MY, Luber CA, Paron I, Nagaraj N, Mann M (2014) Accurate J Biol Chem 272:26522–26529 proteome-wide label-free quantification by delayed normalization and 39. Cheng JS, Dubal DB, Kim DH, Legleiter J, Cheng IH, Yu G-Q, Tesseur I, maximal peptide ratio extraction, termed MaxLFQ. Mol Cell Proteomics 13: Wyss-Coray T, Bonaldo P, Mucke L (2009) Collagen VI protects neurons 2513–2526. https://doi.org/10.1074/mcp.M113.031591 against Abeta toxicity. Nat Neurosci 12:119–121. https://doi.org/10.1038/ 20. Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, Mann M, Cox J nn.2240 (2016) The Perseus computational platform for comprehensive analysis of 40. Grau S, Baldi A, Bussani R, Tian X, Stefanescu R, Przybylski M, Richards P, (prote)omics data. Nat Methods 13:731–740. https://doi.org/10.1038/nmeth.3901 Jones SA, Shridhar V, Clausen T, Ehrmann M (2005) Implications of the 21. Arvanitakis Z, Leurgans SE, Wang Z, Wilson RS, Bennett DA, Schneider JA serine protease HtrA1 in amyloid precursor protein processing. Proc Natl (2011) Cerebral amyloid angiopathy pathology and cognitive domains in Acad Sci 102:6021–6026. https://doi.org/10.1073/pnas.0501823102 older persons. Ann Neurol 69:320–327. https://doi.org/10.1002/ana.22112 41. Chu Q, Diedrich JK, Vaughan JM, Donaldson CJ, Nunn MF, Lee K-F, 22. Boyle PA, Yu L, Nag S, Leurgans S, Wilson RS, Bennett DA, Schneider JA Saghatelian A (2016) HtrA1 proteolysis of ApoE in vitro is allele selective. (2015) Cerebral amyloid angiopathy and cognitive outcomes in community- J Am Chem Soc 138:9473–9478. https://doi.org/10.1021/jacs.6b03463 based older persons. Neurology 85:1930–1936. https://doi.org/10.1212/WNL. 42. Hara K, Shiga A, Fukutake T, Nozaki H, Miyashita A, Yokoseki A, Kawata H, Koyama A, Arima K, Takahashi T, Ikeda M, Shiota H, Tamura M, Shimoe Y, 23. Manousopoulou A, Gatherer M, Smith C, Nicoll JAR, Woelk CH, Johnson M, Hirayama M, Arisato T, Yanagawa S, Tanaka A, Nakano I, Ikeda S, Yoshida Kalaria R, Attems J, Garbis SD, Carare RO (2016) Systems proteomic analysis Y,Yamamoto T,Ikeuchi T,KuwanoR,Nishizawa M,Tsuji S,OnoderaO reveals that clusterin and tissue inhibitor of metalloproteinases 3 increase in (2009) Association of HTRA1 mutations and familial ischemic cerebral leptomeningeal arteries affected by cerebral amyloid angiopathy. small-vessel disease. N Engl J Med 360:1729–1739. https://doi.org/10.1056/ Neuropathol Appl Neurobiol. https://doi.org/10.1111/nan.12342 NEJMoa0801560 Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 19 of 19 43. Tikka S, Baumann M, Siitonen M, Pasanen P, Pöyhönen M, Myllykangas L, Viitanen M, Fukutake T, Cognat E, Joutel A, Kalimo H (2014) CADASIL and CARASIL. Brain Pathol 24:525–544. https://doi.org/10.1111/bpa.12181 44. Tumini E, Porcellini E, Chiappelli M, Conti CM, Beraudi A, Poli A, Caciagli F, Doyle R, Conti P, Licastro F (2007) The G51S purine nucleoside phosphorylase polymorphism is associated with cognitive decline in Alzheimer’s disease patients. Hum Psychopharmacol Clin Exp 22:75–80. https://doi.org/10.1002/hup.823 45. Ramirez A, van der Flier WM, Herold C, Ramonet D, Heilmann S, Lewczuk P, Popp J, Lacour A, Drichel D, Louwersheimer E, Kummer MP, Cruchaga C, Hoffmann P, Teunissen C, Holstege H, Kornhuber J, Peters O, Naj AC, Chouraki V, Bellenguez C, GerrishA, HeunR,Frolich L,Hull M,Buscemi L, Herms S, KolschH, ScheltensP, Breteler MM, Ruther E, Wiltfang J, Goate A, Jessen F, Maier W, Heneka MT, Becker T, Nothen MM (2014) SUCLG2 identified as both a determinator of CSF a 1-42 levels and an attenuator of cognitive decline in Alzheimer’sdisease. Hum Mol Genet 23:6644–6658. https://doi.org/10.1093/hmg/ddu372 46. Boche D, Zotova E, Weller RO, Love S, Neal JW, Pickering RM, Wilkinson D, Holmes C, Nicoll JAR (2008) Consequence of Abeta immunization on the vasculature of human Alzheimer’s disease brain. Brain 131:3299–3310. https://doi.org/10.1093/brain/awn261 47. Sperling R, Salloway S, Brooks DJ, Tampieri D, Barakos J, Fox NC, Raskind M, Sabbagh M, Honig LS, Porsteinsson AP, Lieberburg I, Arrighi HM, Morris KA, Lu Y, Liu E, Gregg KM, Brashear HR, Kinney GG, Black R, Grundman M (2012) Amyloid-related imaging abnormalities in patients with Alzheimer’s disease treated with bapineuzumab: a retrospective analysis. Lancet Neurol 11:241–249. https://doi.org/10.1016/S1474-4422(12)70015-7 48. Banerjee G, Carare R, Cordonnier C, Greenberg SM, Schneider JA, Smith EE, Van Buchem M, Van Der Grond J, Verbeek MM, Werring DJ (2017) The increasing impact of cerebral amyloid angiopathy: essential new insights for clinical practice. J Neurol Neurosurg Psychiatry 88:982–994. https://doi.org/ 10.1136/jnnp-2016-314697 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Acta Neuropathologica Communications Springer Journals

Proteomics analysis identifies new markers associated with capillary cerebral amyloid angiopathy in Alzheimer’s disease

Free
19 pages

Loading next page...
 
/lp/springer_journal/proteomics-analysis-identifies-new-markers-associated-with-capillary-dSkV4imeGm
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Biomedicine; Neurosciences; Pathology; Neurology
eISSN
2051-5960
D.O.I.
10.1186/s40478-018-0540-2
Publisher site
See Article on Publisher Site

Abstract

Alzheimer’s disease (AD) is characterized by amyloid beta (Aβ) deposits as plaques in the parenchyma and in the walls of cortical and leptomeningeal blood vessels of the brain called cerebral amyloid angiopathy (CAA). It is suggested that CAA type-1, which refers to amyloid deposition in both capillaries and larger vessels, adds to the symptomatic manifestation of AD and correlates with disease severity. Currently, CAA cannot be diagnosed pre-mortem and disease mechanisms involved in CAA are elusive. To obtain insight in the disease mechanism of CAA and to identify marker proteins specifically associated with CAA we performed a laser dissection microscopy assisted mass spectrometry analysis of post-mortem human brain tissue of (I) AD cases with only amyloid deposits in the brain parenchyma and no vascular related amyloid, (II) AD cases with severe CAA type-1 and no or low numbers of parenchymal amyloid deposits and (III) cognitively healthy controls without amyloid deposits. By contrasting the quantitative proteomics data between the three groups, 29 potential CAA-selective proteins were identified. A selection of these proteins was analysed by immunoblotting and immunohistochemistry to confirm regulation and to determine protein localization and their relation to brain pathology. In addition, specificity of these markers in relation to other small vessel diseases including prion CAA, CADASIL, CARASAL and hypertension related small vessel disease was assessed using immunohistochemistry. Increased levels of clusterin (CLU), apolipoprotein E (APOE) and serum amyloid P-component (APCS) were observed in AD cases with CAA. In addition, we identified norrin (NDP) and collagen alpha-2(VI) (COL6A2) as highly selective markers that are clearly present in CAA yet virtually absent in relation to parenchymal amyloid plaque pathology. NDP showed the highest specificity to CAA when compared to other small vessel diseases. The specific changes in the proteome of CAA provide new insight in the pathogenesis and yields valuable selective biomarkers for the diagnosis of CAA. Keywords: Cerebral amyloid angiopathy, Amyloid beta, Alzheimer’s disease, Proteomics, Biomarker, Laser microdissection, Human brain, Post-mortem tissue Introduction pathology in varying degrees. When restricted to the Alzheimer’s disease (AD) pathology is characterized by larger blood vessels, including leptomeningeal vessels, the deposition of amyloid beta (Aβ) in the brain paren- cortical arteries and arterioles, this is referred to as chyma as amyloid plaques and at the brain vasculature. CAA type-2. In approximately 50% of the AD cases also The latter is referred to as cerebral amyloid angiopathy brain capillaries are affected, which is designated as CAA (CAA). Approximately 80% of AD cases have CAA type-1 [1, 2]. Especially around the capillaries the Aβ de- posits can extend into the parenchyma as perivascular Aβ also referred to as dyshoric changes [3]. In AD the observed * Correspondence: D.C.Hondius@vu.nl; D.Hondius@vumc.nl August B. Smit and Annemieke J. M. Rozemuller contributed equally to this plaque pathology and CAA type-1 capillary deposits have work. an inverse correlation [4]. Department of Pathology, Amsterdam Neuroscience, VU University Medical Aβ deposition at the vessel wall in CAA correlates Center, PO Box 7057, 1007, MB, Amsterdam, The Netherlands Department of Molecular and Cellular Neurobiology, Center for with an increase in the occurrence of cerebral infarction, Neurogenomics and Cognitive Research, Amsterdam Neuroscience, VU cerebral haemorrhage and micro-bleeds. In addition, it University Amsterdam, Amsterdam, The Netherlands © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 2 of 19 causes a structural disruption of the vascular wall and neurofibrillary tangles and neuritic plaques is staged might indirectly deteriorate the integrity of the micro- [12–14] and indicated conform the ABC criteria [15]. vascular network [5, 6]. Aβ peptide transport through the blood brain barrier (BBB) or via perivascular drain- Fast immunohistochemistry for LCM age is an important mechanism to clear the brain from Sections (10 μm) of fresh frozen occipital tissue were Aβ. Disruption of Aβ clearance is thought to lead to in- mounted on PEN-membrane slides (Leica), air-dried and crease in Aβ deposition in the walls of capillaries and fixed in 100% ethanol for 1 min. After air drying the tis- blood vessels, which in turn further decreases drainage sue was wetted with sterile PBS. Anti-Aβ (clone IC16, capacity resulting in further enhancement of Aβ depos- detecting N-terminal part of Aβ [16]) was applied at a 1: ition [7, 8]. 100 dilution in sterile PBS (pH 7.5) and incubated for CAA type-1 is clinically highly relevant, as it contributes 20 min at RT. After washing 3 times for 30 s in sterile to the symptomatic appearance of AD and, in severe form, PBS, HRP labelled rabbit anti-mouse (DAKO) was ap- CAA type-1 can present itself as the primary cause of plied at a 1:100 dilution in sterile PBS and incubated for rapidly progressive dementia [9, 10]. 15 min at RT. Sections were briefly washed (3 × 30 s) Currently, definitive diagnosis of AD and the occurrence and freshly prepared 3,3′ diaminobenzidine (DAB) solu- of CAA can only be determined post-mortem. However, tion was applied and left to incubate for 5 min to visualize the presence or absence of CAA in AD patients might antibody binding. Sections were thoroughly washed in alter therapeutic options. In particular, a biomarker to de- ultra-pure H O and incubated with 1% (w/v) toluidine tect CAA in patients might aid in stratification of patient blue in ultrapure H O for 1 min as a counterstain. Sec- groups, which is highly important when initiating, inter- tions were then washed in ultra-pure H O twice for 1 min preting and improving outcome in clinical trials. More- and twice in 100% ethanol for 1 min and air dried. over, proteins selectively involved in CAA may function as therapeutic targets. Brain tissue preparation and laser capture Proteomics analysis using mass spectrometry is a pre- microdissection (LCM) ferredmethodtoobtainanunbiasedinsightinto proteins Laser capture microdissection (LCM) was performed as involved in disease. For this, 20 cases were selected encom- described previously [17]. LCM was performed using a passing a group of AD patients with severe CAA type-1, a Leica AS LMD system (Leica). Cortical layers II to VI group with AD bearing severe plaque pathology but devoid which were randomly selected from control tissue and of CAA, and a cognitively healthy control group without selected based on the presence of severe Aβ pathology any pathology in the occipital lobe. Subsequently, we per- in the case of AD and CAA were collected in Eppendorf formed a proteomics analysis of small laser dissected oc- tubes containing 30 μl M-PER lysis buffer (Thermo cipital tissue sections containing either high plaque load, or Scientific) supplemented with reducing SDS sample severe CAA or no Aβ deposits. By contrasting the protein buffer (Thermo Scientific). Between 10 and 20 tissue sec- expression profiles of these subject groups we discovered tions with a thickness of 10 μm were captured using LCM, 9 3 proteins that are highly selective for CAA. These proteins yielding an equal volume each of 1.0 × 10 μm .Micro- also provide insight in specific pathogenic components of dissected tissue was stored at − 80 °C until further use. CAA, which might offer new targets for therapy. Protein separation by electrophoresis and in-gel digestion Material and methods Micro-dissected tissue lysates were incubated at 95 °C Case selection for 5 min to denature the proteins, followed by incu- Post mortem brain tissue was obtained from the bation with 50 mM iodoacetamide for 30 min at RT Netherlands Brain Bank (NBB), Netherlands Institute in the dark to alkylate the cysteine residues. To re- for Neuroscience (NIN), Amsterdam. All brain tissue duce protein complexity, samples were size separated was collected from donors with written informed con- on a NuPAGE® 4–12% Bis-Tris acrylamide gel using sent forbrainautopsyand theuse of thematerialand MOPS SDS running buffer (Invitrogen) according to clinical information for research purposes has been the manufacturers’ protocol. obtained by the NBB. Brain tissue was selected based Gels were fixed in a solution containing 50% (v/v) on clinical and neuropathological reports. Three groups ethanol and 3% (v/v) phosphoric acid in H O for 3 h at were composed. Cognitively healthy control cases lacking RT and stained with Colloidal Coomassie Blue (34% (v/v) any pathology, AD cases with severe plaque pathology but methanol, 3% (v/v) phosphoric acid, 15% (w/v) ammonium devoid of CAA and CAA type-1 cases with severe and Sulphate, and 0.1% (w/v) Coomassie brilliant blue G-250 (nearly) pure capillary CAA pathology (Thal stages 2 (Thermo Scientific), overnight while shaking. Destaining and 3 for CAA) [11]. All cases are listed in Table 1. was performed in ultra-pure water under gentle agitation Alzheimer’s disease pathology present as Aβ deposits, for several hours to reduce background staining (Additional Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 3 of 19 Table 1 Patient data MS/Validation Case Diagnosis M/F Age (years) Abeta Tau CERAD PMD APOE MS 1 CAA type-1 F 75 A3 B3 C0 6:00 44 MS 2 CAA type-1 F 96 A3 B3 C0 4:20 43 a c MS 3 CAA type-1 M 68 A3 B1 C0 6:05 44 MS 4 CAA type-1 F 78 A3 NA C0 4:20 44 MS 5 CAA type-1 M 81 A3 B3 C2 6:30 44 MS 6 CAA type-1 F 95 A3 B3 C2 4:35 44 MS 7 CAA type-1 M 80 A3 B3 C0 5:05 44 MS 8 AD F 82 A3 B3 C3 6:00 42 MS 9 AD F 72 A3 B3 C3 6:30 44 MS 10 AD F 81 A3 B3 C3 6:00 33 MS 11 AD F 73 A3 B3 C3 5:55 44 MS 12 AD M 84 A3 B3 C3 8:05 NA MS 13 AD F 87 A3 B3 C3 5:45 43 MS 14 AD F 72 A3 B3 C3 5:55 23 MS 15 Control M 74 A0 B0 C0 8:05 33 MS 16 Control F 80 A1 B1 C0 6:58 43 MS 17 Control M 82 A0 B1 C0 5:10 23 MS 18 Control M 78 A0 B1 C0 17:40 33 MS 19 Control F 79 A0 B1 C0 18:13 33 MS 20 Control F 81 A0 B1 C0 4:25 33 V 21 CAA type-1 F 94 A3 B3 C3 04:30 43 V 22 CAA type-1 M 74 A3 B3 C3 03:25 NA V 23 CAA type-1 F 87 A3 B3 C3 08:00 44 V 24 CAA type-1 F 84 A3 B3 C2 04:45 NA V 25 CAA type-1 M 88 A3 B3 C3 03:55 NA V 26 CAA type-1 M 75 A3 B3 C0 03:15 NA V 27 AD M 64 A3 B3 C3 07:30 33 V 28 AD F 81 A3 B3 C3 05:15 43 V 29 AD F 90 A3 B3 C3 04:45 33 V 30 AD M 65 A3 B3 C3 06:00 43 V 31 AD F 73 A3 B3 C3 NA NA V 32 AD F 90 A3 B3 C3 03:55 32 V 33 AD M 88 A3 B3 C3 04:40 43 V 34 AD M 74 A3 B3 C3 05:10 NA V 35 Control M 73 A0 B0 C0 24:45 33 V 36 Control M 71 A0 B1 C0 07:40 33 V 37 Control F 82 A0 B1 C0 07:00 33 V 38 Control M 56 A0 B0 C0 09:15 43 V 39 Control M 62 A0 B1 C0 07:20 33 V 40 Control M 76 A0 B0 C0 06:45 33 V 41 Control M 93 A0 B1 C0 05:05 33 V 42 Control F 60 A0 B0 C0 08:10 32 V 43 Cotton wool M 72 A3 B3 C0 05:15 43 V 44 Prp-CAA F 57 A0 B0 C0 24:00 NA Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 4 of 19 Table 1 Patient data (Continued) MS/Validation Case Diagnosis M/F Age (years) Abeta Tau CERAD PMD APOE V 45 CADASIL M 73 A0 B0 C0 31:45 NA V 46 CARASAL F 55 A1 B1 C0 04:00 NA V 47 Hyper tension related SVD F 92 A1 B2 C0 07:25 NA Alzheimer’s disease: AD, cerebral amyloid angiopathy: CAA, M: male, F: female, post mortem delay: PMD, not available/not applicable: NA, used for mass a b spectrometry analysis: MS, used for validation: V. ( Aβ only present as dysphoric CAA) ( Focal tau accumulation around blood vessels with prp-amyloid deposits) ( only dyshoric angiopathy in gallyas staining) file 1: Figure S1 ). Each gel lane was sliced into 12 equal modifications. Mass deviation tolerance was set to sized parts to reduce sample complexity during later mass 20 ppm for monoisotopic precursor ions and 0.5 Da for spectrometry analysis and each part was cut into blocks of MS/MS peaks. False-discovery rate cut-offs for peptide approximately 1 mm and collected in an Eppendorf tube. and protein identifications were set to 1% for both. The Gel fragments were destained in ultrapure water with minimum peptide length was seven amino acids. Identi- 50 mM NH HCO and 50% (v/v) acetonitrile overnight. fied proteins that had the same set of peptides or a sub- 4 3 Gel fragments were dehydrated using acetonitrile for set of peptides compared to another protein, were 20 min and dried for 30 min using a SpeedVac. The merged into one protein group. Peptides that were shared gel parts were rehydrated in 70 μl of ultra-pure water between different proteins were assigned to the protein containing 50 mM NH HCO and 10 μg/ ml trypsin with most peptide evidence (so-called ‘Razor’ peptides). 4 3 (sequence grade; Promega) and incubated overnight at Only protein groups with at least a single unique and a 37 °C to facilitate digestion of the proteins. Peptides single Razor peptide were included. For relative protein were extracted twice with a solution containing 0.1% quantification MaxQuant LFQ intensities based on at least (v/v) trifluoric acid and 50% (v/v) acetonitrile for a single shared peptide ratio were used [19]. 20 min. The samples were dried using a SpeedVac and stored at − 20 °C until further analysis. Statistical analysis of differential protein expression To identify proteins that differ in abundance between Mass spectrometry analysis the different experimental groups an ANOVA (Kruskal– The peptides of the individual sample fractions were dis- Wallis test) was performed using the Perseus software solved in 15 μL of 0.1% (v/v) acetic acid of which 10 μL platform [20], adhering to a significance cut-off of p ≤ 0.05. was loaded onto a nano-liquid chromatography (nano- The p values were not corrected for multiple testing to in- LC) system (Eksigent). The peptides were separated cludemoreproteinsand providea broadimpressionofthe using a capillary reversed phase C18 column that had differences in the proteome. been equilibrated with 0.1% (v/v) acetic acid at a flow Conditions that were set for inclusion of CAA selective rate of 400 nL/min. The peptides were eluted by increas- proteins comprise of three approaches (A, B and C) that ing the acetonitrile concentration linearly from 5 to 40% are visualized in Fig. 2. Approach A: T-tests (two-sided, in 80 min and to 90% in 10 min, using the same flow rate. assuming unequal variances, performed using Excel Eluted peptides were transferred into the LTQ/Orbitrap (Microsoft)) were performed contrasting the three ex- MS (Thermo Scientific) by Electro Spray Ionisation (ESI). perimental groups. When there was a significant differ- The Orbitrap was operated in the range of m/z 350–2000 ence (p < 0.05) between both the control group versus at a full width at half maximum resolution of 30,000 after CAA, andthe AD groupversusCAA,aproteinwas la- accumulation to 500,000 in the LTQ with one microscan. belled as CAA specific. Approach B: If the number of The five most abundant precursor ions were selected for quantitative values in the control group was zero or fragmentation by collision-induced dissociation (CID) with one while the AD and CAA groups both had two or an isolation width of 2 Da. more quantitative values, than a t-test was performed between theAD and theCAA group. When theAD Protein inference and relative protein quantification group had zero or one quantitative values while the MaxQuant software was used for spectrum annotation, control and CAA groups both had two or more quanti- protein inference, and relative protein quantification tative values a t-test was performed between the CAA [18]. Spectra were annotated against the Uniprot human and control group. Approach C: In the case of zero or reference proteome database (version 2016_04). Enzyme single quantitative values in both the control and AD specificity was set to Trypsin/P, allowing at most two groups, proteins were included based exclusively on a missed cleavages. Carbamido-methylation of cysteine minimum of four quantitative values in the CAA group. was set as a fixed modification, and N-acetylation and Also, we included proteins with zero or single quantitative methionine oxidation were set as variable Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 5 of 19 values in the CAA group and four or more values in both staining, the sections were fixed in 100% acetone for the AD and control groups. 10 min. For paraffin sections the paraffin was removed by ANOVA (Kruskal–Wallis test) and posthoc Dunn’s washing in xylene. Next, the sections were washed in multiple comparison tests on immunoblot data and im- decreasing concentrations of ethanol (100%, 96% and 70% munohistochemical data was performed using Graphpad (v/v)). Endogenous peroxidase activity was quenched by Prism (GraphPad Software). incubating in methanol with 0.3% H O for 30 min at RT. 2 2 Next, antigen retrieval was performed by submerging the Immunoblot analysis slides in citrate buffer (pH 6) and heating in an autoclave. Protein extracts were prepared by lysis of whole occipital Primary antibodies were diluted in antibody diluent lobe tissue in reducing SDS sample buffer using a 1:20 (VWR) and incubation was performed overnight at 4 °C. tissue weight to lysis buffer ratio. Proteins were dena- All primary antibodies and corresponding dilutions used tured at 95 °C for 5 min and separated by SDS-PAGE are listed in Table 2. After incubation the sections were using precast Stain Free gradient gels (Bio-Rad) and thoroughly washed in PBS (pH 7.4) for 30 min followed transferred (40 V overnight at 4 °C) onto a 0.45 μm by incubation of an HRP-labelled secondary antibody, PVDF membrane (Merck Millipore), which was pre- Envision (DAKO) for 30 min. Again, the sections were incubated in 100% methanol. The PVDF membrane was thoroughly washed in PBS (pH 7.4) for 30 min and then incubated in Odyssey blocking buffer for 1 h and subse- incubated with DAB to visualize antibody binding. quently incubated with the primary antibody overnight. Counterstaining of the nuclei was performed by incuba- After washing in Tris-buffered saline (pH 7.5) with 0.1% tion in hematoxylin for 3 min followed by extensive (v/v) Tween-20 (TBST) for 3 × 10 min, the membrane washing in running tab water for 5 min. Next, the slides was incubated for 3 h with the secondary antibody. were dehydrated by incubation in increasing concentra- Visualization was achieved using an Odyssey imaging tions of ethanol consisting of 70% (v/v), 96% (v/v) and system using excitation wavelengths of 700 nm and 100% (v/v) ethanol. The slides were then incubated in 800 nm. Total protein load was visualized using a chemidoc xylene and mounted using Quick-D mounting medium. EZ (Bio-Rad) after electro blotting (Additional file 2:Figure A negative control was made by omission of the pri- S2) and the protein densitometric values were then used to mary antibody. Quantification of the staining was done normalize for total protein input. Primary antibodies and using ImageJ using the threshold colour plugin. dilutions are shown in Table 2. Secondary antibodies used were IRDye 800 CW Goat anti-Rabbit (LI-COR) and IRDye Results 680 conjugated Goat anti-Mouse (LI-COR) both were Selection of cases, controls and analysis of brain tissue used at a 1:7.000 dilution. All anti-bodies were diluted Three groups with a total of 20 cases were assembled in Odyssey blocking buffer (LI-COR). Quantification based on careful neuro-pathological inspection: 1) cogni- was performed using ImageJ software. tively healthy control cases (n =6) without any Aβ path- ology or tau pathology, 2) AD cases with severe Aβ plaque Immunohistochemical analysis pathology but no vascular deposits (no CAA) (n =7) and Fresh frozen or paraffin embedded human occipital tissue 3) AD cases with severe nearly pure CAA type-1 path- was cut (5 μm). For frozen tissue the sections were placed ology and a negligible amount of plaque pathology (n =7). on a SuperFrost Microscope Slide (VWR, PA, USA) and From here, these groups will be mentioned as “control”, air-dried overnight at room temperature (RT). Prior to “AD” and “CAA”, respectively. Inclusion of these cases Table 2 Antibodies used in this study Antibody Source Species Ordernr. Clone Dilution (IHC) Amyloid-beta Kind gift of Prof. Dr. Korth, Heinrich Heine Mouse IC16 1:200 University, Düsseldorf, Germany APOE Abcam Mouse ab1907 E6D7 1:3200 APOE Santa Cruz Biotechnology Mouse sc-13521 A1.4 Used for immunoblot NDP Novus Biologicals Rabbit NBP1–84769 polyclonal 1:400 NDP R&D systems Mouse MAB3014 #343711 1:800 HTRA1 R&D systems Mouse MAB2916 #275615 1:6400 APCS Statens Serum Institut, SSI Antibodies Mouse #56585 HYB281–05 1:1600 COL6A2 Abnova Mouse H00001292-M01 2C5-F2 1:3200 COL6A2 Santa Cruz Biotechnology Rabbit SC-83607 polyclonal 1:1600 Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 6 of 19 was done based on histochemical analysis using Congo- quantification of 2427 proteins in total and approximately red and additional IHC for Aβ on the occipital frozen tis- 1500 proteins identified per individual case (Additional sue intended for LC-MS-MS analysis. file 3:FigureS3),withaminimumofone tryptic We focussed our analysis on the occipital lobe as this peptide detected. All quantified proteins are listed in region is the most frequently and severely affected by Additional file 4: Table S1. CAA pathology. Tissue sections of human occipital lobe To gain insight into the global similarities and differ- from all selected cases were mounted on PEN-foil slides ences between the three groups and the individual cases and Aβ pathology was visualized using fast immunohis- an ANOVA (Kruskall Wallis) was performed. This yielded tochemistry. Grey matter tissue was isolated using LCM. 309 proteins that have a significant difference (p <0.05) in Tissue isolation from the AD cases and CAA cases was abundance between any of the experimental groups. Using focused on occipital lobe grey matter areas with severe these proteins in an unsupervised clustering analysis, three Aβ pathology, i.e. high plaque load or high CAA type-1 different expression signatures were obtained. The protein burden, respectively. This was done to selectively enrich expression signatures of the AD and CAA groups ap- the input material for the proteomics analysis for these peared largely similar, whereas both were different from types of Aβ pathology. For control cases occipital lobe the control group (Additional file 5: Figure S4A). Un- grey matter areas from the same anatomical region were supervised clustering analysis of the individual cases using selected for isolation. LCM-collected tissue samples were the 309 ANOVA-identified proteins, separated the con- lysed and proteins were separated using SDS-PAGE. trols from the disease cases (Additional file 5:FigureS4B). Each PAGE sample lane was divided into 12 fractions The CAA and AD cases were not separated on the basis and subjected to in-gel trypsin digestion (Fig. 1). of the full set of differentially expressed proteins indicating that overall their protein expression profile is largely simi- Protein quantification and global protein expression profiles lar. One CAA case (case #5) clustered with the control To identify and quantify proteins, liquid chromatography cases indicating that the protein expression profile of this followed by mass spectrometry (LC-MS-MS) was performed sample is more similar to the control cases than to other on the 20 laser-dissected tissue samples. This allowed CAA or AD cases. Visualizing the expression profile of Fig. 1 Workflow used in this study. Amyloid Beta pathology was visualized in human postmortem occipital lobe tissue. Unaffected grey matter was isolated from healthy control cases. Grey matter with high burden of Aβ pathology was isolated from the AD and CAA cases thereby isolating tissue with high plaque load or high CAA type-1 burden, respectivelyTissue was lysed and the proteins were separated using SDS-PAGE and subjected to in-gel trypsin digestion. Peptides were analysed using LC-MS-MS. A database search for protein identification and protein quantification was performed using MaxQuant software. ANOVA (Kruskall Wallis) and t-tests were performed to identify significantly regulated proteins Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 7 of 19 case #5 next to the average expression profiles of the three Approach C (Fig. 2) resulted in the identification of groups confirmed the resemblance of case #5 to the con- HLA-DRA, HLA-DQA2, HTRA1, APCS, COL6A2, trol group, but also showed several proteins that are simi- MOB2, POTEI, KIAA1468, TMF1 and SGIP1 (Fig. 3c lar in expression to the AD and or CAA groups and Table 3c) as CAA specific proteins. (Additional file 6: Figure S5A). Earlier, Case #5 was identified as having an expression profile resembling a control case. Case #5 was found positive for Alzheimer type 2 astrocytes, possibly related Identification of proteins selectively altered in CAA type-1 to high alcohol intake, and exhibited relatively low tau To identify proteins that have a significantly different pathology. Otherwise, this case showed no pathological abundance in CAA type-1 compared to both the control abnormalities when compared to the rest of the CAA and the AD group, and therefore represent unique fea- type-1 group. However, the expression of several CAA se- tures of CAA type-1, we performed student t-tests (two- lective markers that we identified was inspected for case #5. sided, assuming unequal variances) for those proteins The levels of these markers correspond well with the other where at least two quantitative values per groups were cases of the CAA group (Additional file 6:FigureS5B), in- available. When there was a significant difference (p <0. dicating that these proteins are inseparably linked to the 05) between both the control group versus CAA, and pathology of CAA type-1. In addition, although the number the AD group versus CAA, a protein was designated as of cases is too small to do valid statistics, we observed CAA-specific (Fig. 3a and Table 3a). CLU, APOE, SUCLG2, no clear relation between gender and expression of the PPP2R4, KTN1, ACTG1, TNR, COL6A3 and NFASC met markers (Additional file 7: Figure S6). these criteria. In addition, levels of CLU, APOE, SUCLG2, To determine whether the above-described approaches PPP2R4 and ACTG1 were also significantly different were indeed appropriate in selecting CAA specific pro- (p < 0.05) when comparing the AD group with the teins, we performed additional immunoblotting and im- control group. munohistochemical (IHC) analysis. After calculating the multiple testing corrected false discovery rate (FDR) only CLU was considered signifi- Confirmation of MS data using immunoblotting and cant. This is likely due to the relatively low sample size immunohistochemical analysis of this exploratory study and the high inter-individual Of the proteins described in Table 3 we selected APOE variance that is inevitably associated with the use of hu- (approach A), NDP (approach B), HTRA1, APSC and man tissue. Given the explorative nature of this study we COL6A2 (approach C), based on the fold change or spe- relaxed criteria and adhered to the uncorrected p-values cific expression in the CAA type-1 group compared to for protein inclusion. the AD and control groups, to confirm our mass spectrom- Importantly, using label-free mass spectrometry to etry results. Immunoblotting was performed on whole identify and quantify proteins, the absence of data for a tissue lysates of the same cases as used for the mass spec- number of proteins is observed. Despite great improve- trometry analysis. When comparing the CAA group with ments in the speed and sensitivity of MS analysers miss- the control group we found significant differences in NDP ing data is almost unavoidable. When quantitative data expression (Fig. 4). For APOE, APCS and COL6A2, the are absent in one group while being present in the other data showed the same trend of increased abundance in the group(s), this likely indicates differences in abundance, CAA group as the proteomics data, but the differences did which might represent interesting candidate marker pro- not reach significance. A likely explanation for this is the teins. Therefore, absence of data in one or more patient higher variation of expression of these proteins in the tissue groups required 2 additional approaches to also consider used for immunoblotting, which in contrast to the mass these proteins in this study. An overview of the 3 comple- spectrometry exploratory analysis, was not selectively menting strategies for protein inclusion is shown in Fig. 2 enriched for pathological burden using LCM, and in- and a complete description is present in the methods sec- stead included white matter, leptomeningeal vessels and tion. Note that any given protein is only considered using a grey matter with a lower pathological burden. To un- single approach as these approaches are mutually exclusive. equivocally demonstrate CAA related expression, we Using approach B (Fig. 2), proteins with a significant turned to IHC analysis of these same proteins, which, difference were included, and APP, UBLCP1, SRI, NDP, in contrast to immunoblotting, allows region specific PNP, C1orf123, DHX15, SYNPO, TPM1, CADPS2 and analysis similar to the LCM-LC-MS-MS analysis. For SERPINA3 (Fig. 3b and Table 3b) were identified as pro- this a separate cohort was used consisting of cognitively teins selectively present in CAA type-1. Peptide data on healthy control cases (n = 8) without any Aβ or tau APP indicates that quantification was based on two pep- pathology, 2) AD cases with severe Aβ plaque path- tides in which the most abundantly detected peptide ology but no vascular deposits (no CAA) (n =8) and 3) (LVFFAEDVGSNK) is part of Aβ. AD cases with severe CAA type-1 pathology (n =6). Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 8 of 19 Table 3 A, B and C Proteins identified as selectively altered in CAA type-1 Gene Protein P-val C vs P-val AD vs FDR C vs FDR AD vs FC C vs FC AD vs # detections # detections # detections CAA CAA CAA CAA CAA CAA Control AD CAA A. Significant CAA versus control and CAA versus Alzheimer's disease CLU Clusterin;Clusterin beta chain; 0.000 0.000 0.001 0.007 4.47 2.33 6 7 7 Clusterin alpha chain APOE Apolipoprotein E 0.001 0.001 ns ns 4.97 2.11 6 7 7 SUCLG2 Succinyl-CoA ligase [GDP- 0.002 0.002 ns ns 2.17 0.61 5 7 7 forming] subunit beta, mitochondrial PPP2R4 Serine/threonine-protein 0.026 0.010 ns ns 2.18 0.80 6 7 7 phosphatase 2A activator KTN1 Kinectin 0.015 0.021 ns ns 0.34 0.40 2 2 3 ACTG1 Actin, cytoplasmic 2;Actin, 0.000 0.035 ns ns 0.80 0.90 6 7 7 cytoplasmic 2, N-terminally processed TNR Tenascin-R 0.007 0.017 ns ns 0.82 0.82 6 7 7 COL6A3 Collagen alpha-3(VI) chain 0.016 0.024 ns ns 7.79 4.95 3 5 7 NFASC Neurofascin 0.028 0.040 ns ns 0.86 0.87 6 7 7 B. Significant CAA versus control or Alzheimer’s disease and ≤1 detection in other group APP Amyloid beta A4 protein;N- NA 0.000 NA ns NA 6.65 1 7 7 APP;Soluble APP-alpha;Soluble APP-beta; UBLCP1 Ubiquitin-like domain- NA 0.007 NA ns NA 1.96 1 2 4 containing CTD phosphatase 1 SRI Sorcin NA 0.011 NA ns NA 0.25 1 3 5 NDP Norrin NA 0.020 NA ns NA 5.16 0 4 7 PNP Purine nucleoside NA 0.030 NA ns NA 1.72 0 3 3 phosphorylase C1orf123 UPF0587 protein C1orf123 NA 0.047 NA ns NA 0.71 0 6 6 DHX15 Putative pre-mRNA-splicing 0.005 NA ns NA 0.44 NA 3 0 5 factor ATP- dependent RNA helicase DHX15 SYNPO Synaptopodin 0.015 NA ns NA 0.44 NA 4 1 4 TPM1 Tropomyosin alpha-1 chain 0.021 NA ns NA 0.37 NA 3 1 3 CADPS2 Calcium-dependent secretion 0.042 NA ns NA 1.22 NA 2 1 3 activator 2 Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 9 of 19 Table 3 A, B and C Proteins identified as selectively altered in CAA type-1 (Continued) Gene Protein P-val C vs P-val AD vs FDR C vs FDR AD vs FC C vs FC AD vs # detections # detections # detections CAA CAA CAA CAA CAA CAA Control AD CAA SERPINA3 Alpha-1- 0.042 NA ns NA −0.86 NA 2 1 3 antichymotrypsin;Alpha-1- antichymotrypsin His-Pro-less C. ≤1 detection in control and Alzheimer's disease and ≥4 in CAA OR ≤1 in CAA and ≥4 in Alzheimer’s disease and control HLA-DRA;HLA- HLA class II histocompatibility NA NA NA NA NA NA 0 1 7 DQA2 antigen, DR alpha chain;HLA class II histocompatibility antigen, DQ alpha 2 chain HTRA1 Serine protease HTRA1 NA NA NA NA NA NA 0 1 7 APCS Serum amyloid P- NA NA NA NA NA NA 0 1 6 component;Serum amyloid P-component(1–203) COL6A2 Collagen alpha-2(VI) chain NA NA NA NA NA NA 0 1 5 MOB2 MOB kinase activator 2 NA NA NA NA NA NA 1 1 5 POTEI POTE ankyrin domain family NA NA NA NA NA NA 1 0 4 member I KIAA1468 LisH domain and HEAT NA NA NA NA NA NA 1 0 4 repeat-containing protein KIAA1468 TMF1 TATA element modulatory NA NA NA NA NA NA 0 1 4 factor SGIP1 SH3-containing GRB2-like NA NA NA NA NA NA 4 4 1 protein 3-interacting protein 1 Proteins were found using the three different selection methods as described in Fig. 2 AD Alzheimer’s disease, CAA cerebral amyloid angiopathy, FDR false discovery rate, NA not applicable, NS not significant, FC fold change Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 10 of 19 Fig. 2 Three strategies used to select proteins that are differentially expressed in CAA type-1 compared to control and AD brains. Criteria of each of the selection strategies are specified, numbers of resulted proteins indicated, and selected proteins are listed in tables and figures as indicated First Aβ pathology was visualized and its presence was vessels in the CAA cases and showed plaque pathology in confirmed in AD and CAA type-1 cases showing plaques the AD cases without CAA. Control cases were all nega- and vascular Aβ pathology, respectively (Fig. 5b and c). tive for HTRA1 (Fig. 5j-l). IHC for APOE resulted in pro- Then, IHC analysis was performed to gain information on nounced staining of the vasculature in CAA cases and the localization of the selected proteins. IHC for NDP appeared related to compact deposits as well as more dif- showed pronounced immunoreactivity in CAA type-1 fuse dyshoric deposits. Also immunoreactivity of APOE cases that appeared associated to the vasculature. NDP was observed in the AD cases related to the Aβ plaques, immunostaining in CAA, appeared to be associated with although the staining was less intense than that related to both compact Aβ depositions as well as more diffuse the vascular amyloid in the CAA cases (Fig. 5p-r). APCS staining in the parenchyma in cases that exhibit dyshoric IHC illustrated the presence of this protein in relation Aβ deposits. Staining was more pronounced related to ca- with both diffuse and compact Aβ pathology in both the pillaries compared to larger vessels. The AD cases with CAA and AD group. However, staining related to the plaques were nearly devoid of immunoreactivity, controls plaque pathology was less intense than that related to the did not show any immunoreactivity for NDP (Fig. 5d-f). vascular Aβ pathology (Fig. 5m-o). Different antibodies against NDP showed similar results For quantification of the IHC, images were obtained at (data not shown). sites that, for the AD and CAA cases, had high Aβ COL6A2 IHC showed some immunoreactivity in con- pathological burden in nearby sections of the same tis- trol and AD cases which was restricted to leptomeningeal sue block. The percentage of positively stained pixels vessels (Additional file 8: Figure S7) and a few large vessels over a total of 5 images from each case was determined. in the brain tissue. In CAA type-1, immunoreactivity for Although this method is semi-quantitative, it allowed a COL6A2 was highly increased and includes brain capillar- region-specific analysis in line with the tissue obtained ies and larger vessels. Immunoreactivity was mostly by laser capture dissection that was at the basis of the associated with the endothelium and/or the adventitia original mass spectrometry analysis. Using IHC quantifi- (Fig. 5g-i). Similar results were obtained using two differ- cation we observed strongly increased immunoreactivity ent antibodies for COL6A2 (data not shown). HTRA1 in CAA type-1 cases for NDP and COL6A2, and moder- IHC showed clear overlap with Aβ in both AD and CAA. ate increases for APCS, HTRA1 and APOE, confirming Compact and diffuse staining was observed related to the the MS results (Fig. 6). Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 11 of 19 Fig. 3 Relative abundance of proteins with altered expression in CAA type-1 compared to Alzheimer’s disease cases and controls as determined by MS. Three groups of selected proteins (panels a-c), with altered levels (MS-derived, log2 LFQ intensity values) in CAA type-1 compared to the control groups and the AD groups. Selection criteria are specified in Fig. 2. Gene names are indicated Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 12 of 19 Fig. 4 Immunoblotting analysis of proteins with altered expression in CAA type-1. a Immunoblotting was performed for, NDP, COL6A2, APCS and APOE on occipital lobe tissue lysates of non-demented controls (N) AD cases (A) and CAA cases (C). Immunoreactivity was observed at the correct molecular weight for each protein. b A significant difference (p < 0.05) was found in the expression of NDP in the CAA group when compared to control but not when compared to the AD group. For COL6A2, APCS and APOE significance was not reached between any of the groups using this technique. Data are expressed as mean ± SEM Specificity in relation to other small vessel diseases vessels (Fig. 7g). IHC for APOE and APCS presented a To assess the specificity of Aβ, NDP, COL6A2, APCS highly positive, staining that appeared similar to the Aβ and APOE in relation to other small vessel diseases we staining. performed additional IHC on cases that present various The PrP-CAA tissue was confirmed negative for Aβ types of vascular defects, i.e., cotton wool plaque path- pathology (Fig. 7k) and showed positive for Prp (data ology, prion CAA, cerebral autosomal dominant arterio- not shown). NDP immunoreactivity was highly increased pathy with subcortical infarcts and leukoencephalopathy and localised to the affected vessels. NDP staining was (CADASIL), hypertension related small vessel disease more pronounced around affected capillaries than at af- and Cathepsin A-related arteriopathy with strokes and fected larger vessels. COL6A2 immunoreactivity was leukoencephalopathy (CARASAL). IHC was performed clearly present around capillaries and larger affected ves- on sections that exhibited the relevant pathological char- sels, of which the vascular pathology was clearly ob- acteristics of each disease including an additional CAA served using haematoxylin. Although co-occurring in type 1 case. Immunoreactivity for Aβ, NDP, COL6A2, largely the same vessels COL6A2 did not generally co- APOE and APCS was confirmed in the CAA type-1 case localize with the deposits, but instead localized more in- (Fig. 7a-e). Immunoreactivity for these proteins was also ternally as a component of the basal membrane. APOE assessed and confirmed in case exhibiting hereditary and APCS were also highly present and co-localized with cerebral haemorrhage with amyloidosis Dutch type the Prp deposits. (HCHWA-D), which is a heredity form of CAA-type-1 In the white matter no immunoreactivity was seen for (Additional file 9: Figure S8). Aβ or any of the marker proteins (Fig. 7p-t) in the con- In tissue with cotton wool plaques the pathology was trol case. The CADASIL case showed characteristic confirmed using IHC for Aβ displaying pathology pathology in the white matter, thickened vessel walls, in around capillaries and larger vessels with dyshoric the white matter as is common with this condition. No changes extending deep into the parenchyma (Fig. 7f). Aβ pathology was present (Fig. 7u) Mild immunoreactiv- Intense NDP immunoreactivity was observed related to ity for NDP was present and of the assessed proteins capillaries and the dyshoric changes and to a lesser ex- COL62A staining was most pronounced. APOE and tend directly lining the larger vessels. COL6A2 showed APCS also displayed mild immunoreactivity related to intense immunoreactivity lining the capillaries and larger the affected vessels (Fig. 7v-y). Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 13 of 19 Fig. 5 Immunohistochemical analysis of selected proteins. Representative images were taken. Aβ pathology was visualized. a The control case does not have any Aβ pathology. b Plaque pathology is confirmed in the AD case and (c) CAA type-1 pathology is confirmed in the CAA type-1 case. d-f Extensive NDP immunoreactivity is observed in the CAA type-1 cases whereas absent in both control and AD cases without CAA. g-i COL6A2 immunoreactivity is hardly observed in the control and AD cases, however, extensive immunoreactivity is observed in the CAA type cases and includes both capillaries and large vessels. j-l Immunoreactivity for HTRA1 is absent in control tissue, however, is observed both related to plaque pathology and CAA at comparable intensity. m-o Immunoreactivity for APCS is absent in control cases but is observed both related to plaque and CAA type-1 pathology. However, the intensity of the staining observed in the AD cases is considerably less. p-r Also, APOE immunoreactivity is observed related to both plaque and CAA type-1 pathology, yet its intensity in CAA type-1 is far greater. Scale bar, 100 μm in images A to R. Scale bar in image (C`) 10 μm and in all zoomed images, which are marked with a grave accent (`) In hypertension related small vessel disease the presence in CAA (type-1 and cotton wool) and Prp-CAA. Involve- of COLA6A2 was most prominent while immunoreactivity ment of NDP, APOE and APCS in other small vessel dis- for NDP, APOE and APCS was low but present (Fig. 7z-ad). eases is varying from non (CARASAL) to mild Interestingly IHC analysis of the CARASAL cases showed (CADASIL). Importantly, NDP is explicitly suitable to evi- only the prominent presence of COL6A2 in affected vessels dently separate CAA from Aβ plaque pathology (Table 4). while NDP, APOE and APCS were absent. Taken our data together, from the tested panel of pro- Discussion teins, we recognize COL6A2 as a general small vessel dis- One of the most prevalent cerebro-vascular diseases in ease marker. NDP, APOE and APCS are most prominent the elderly is sporadic CAA, characterized by vascular Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 14 of 19 Fig. 6 Semi-quantitative analysis of immunohistochemical data of proteins with altered expression in CAA type-1. Immunohistochemical stainings were quantified by measuring the percentage of pixels that showed positive immunoreactivity. Significance was calculated using a one-way ANOVA (Kruskal-Wallis test) and posthoc Dunn’s Multiple Comparison Test. A significant increase (p < 0.05) in immunoreactivity in the CAA group compared to both control and AD groups was observed for NDP and COL6A2. For APOE, APCS and HTRA1 significant differences were only found when comparing CAA with control, but not with the AD group. Data are expressed as mean ± SEM deposition of amyloid-beta protein. CAA can occur as plaque pathology, in accordance with previous findings an isolated disease or as part of the pathology in AD. [24, 25]. Interestingly, increased levels of CLU have been Several studies have indicated CAA as an important reported in the plasma of CAA patients diagnosed ac- cause of cognitive decline [10, 21, 22]. Currently, there is cording the modified Boston criteria [26]. no treatment for CAA and its presence cannot be diag- Two recent proteomics studies focussed on CAA ana- nosed pre-mortem. Therefore insight in the pathogenic lysing leptomeningeal vessels [27] and leptomeningeal mechanisms and the need for biomarkers are urgent. vessel combined with neocortical arterioles [28]. Some We performed an exploratory laser dissection-assisted similarities were found with these studies, such as the LC-MS-MS analysis of AD brain tissue exhibiting severe increase in CLU and APOE. As expected also several dif- CAA type-1 pathology, AD brain tissue without appar- ferences exist as our analysis focussed on grey matter of ent involvement of CAA, and control brain tissue with- CAA type-1 using micro dissected tissue that is enriched out AD related pathology. We show that the proteome for areas with very high capillary associated Aβ path- of CAA type-1 is different from that of parenchymal ology. These differences might indicate that other mech- plaque pathology in AD, which led to the identification anisms are involved in the pathogenesis in CAA related of proteins selectively associated with CAA. to capillaries compared to larger vessels, e.g., the find- Next to identification of new CAA selective proteins, ings of NDP and COL6A2. In contrast to these previous this study also confirmed the presence proteins already studies we compared CAA cases with both control and known to be involved in CAA pathology, e.g., CLU, AD cases with plaque pathology and without CAA. APOE and APCS [23, 24]. Interestingly, CLU was de- There are many similarities in the response to CAA and tected in all samples included in the study and its abun- AD and these proteins and processes can be cancelled dance was sufficient to completely separate the CAA out against each other, allowing CAA selective proteins group from both the control group and the AD group. to become apparent. Moreover, CLU, APOE and APCS were markedly in- Importantly, we identified potential new key players in creased in AD compared to controls and APOE and the development of CAA. NDP is found highly upregu- APCS showed moderate immunoreactivity related to lated in CAA type-1, cotton wool Aβ pathology and Prp- Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 15 of 19 Fig. 7 Expression of CAA type-1 markers in other small vessel diseases. IHC for Aβ, NDP, COL6A2, APCS and APOE was performed. In a CAA type- 1 case immunoreactivity for all marker proteins is confirmed (a-e). Also immunoreactivity is seen for all markers in the cotton wool case and Aβ pathology was confirmed (f). For NDP, APOE and APCS immunoreactivity is also seen localizing to severe dyshoric angiopathy (g, i and j). COL6A2 immunoreactivity is restricted to the vessel wall (h). In de Prp-CAA case Aβ pathology was absent (k). Extensive immunoreactivity was observed for NDP, COL6A2, APOE and APCS (l-o). No immunoreactivity was observed for Aβ, NDP, COL6A2, APOE and APCS in the white matter of control tissue (p-t). In the CADASIL case no Aβ pathology was present (u). Mild immunoreactivity for NDP (v) COL62A staining was most pronounced (w). APOE and APCS also displayed mild immunoreactivity related to the affected vessels (x, y). In hypertension related small vessel disease no Aβ was detected (z) immunoreactivity of COLA6A2 was moderate (ab) while immunoreactivity for NDP, APOE and APCS were low but present (aa, ac and ad). In the CARASAL case only prominent immunoreactivity of COL6A2 was seen in affected vessels (ag) while NDP, APOE and APCS were absent (af, ah and ai). Scale bar in (a) indicates 100 μm Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 16 of 19 Table 4 Scoring of relative immunoreactivity of 5 CAA-1 positive highly increased in CAA type-1. COL6A2 immunoblot- markers in different small vessel diseases and AD plaque ting did not show a significant difference between the pathology experimental groups. This can be explained by the inclu- Aβ NDP COL6A2 APOE APCS sion of leptomeningeal vessels, that express high levels CAA type-1 3 3 3 3 3 of COL6A2 in all cases in varying amounts in the tissue lysates used for immunoblotting, as shown using IHC. Cotton wool 3 3 3 3 3 COL6A2 was also found increased in the other small Prp-CAA 0 3 3 3 3 vessel diseases that were included for IHC analysis. This Control white matter 0 0 0 0 0 indicates that COL6A2 might be used as valuable indica- CADASIL 0 2 2 1 1 tor of vascular pathology in general, but not specific for Hypertension 0 1 2 1 1 CAA type pathology. COL6A2 is a non-fibril collagen CARASAL 0 0 2 0 0 and COL6 isoforms are present in various tissues includ- ing the vasculature [37]. COL6A2 encodes one of the AD plaques 3 0 0 2 2 three alpha chains of type VI collagen, which is found in 0: no immunoreactivity, 1: mild immunoreactivity, 2: moderate immunoreactivity, 3: extensive immunoreactivity most connective tissues. Type VI collagen anchors endo- thelial basement membranes by interacting with type IV CAA, and localizes around the affected vasculature. collagen [38]. In the brain, collagen VI was shown neu- NDP immunoreactivity was only mildly increased in roprotective and its expression increased in animal CADASIL and hypertension related small vessel disease. models of AD [39]. In addition, these diseases affect different anatomical re- HTRA1 is a trypsin-like serine protease which was de- gions, clearly identifiable with imaging studies, and tected in all CAA samples, with a single value in the AD present a distinct clinical picture, leaving NDP a promis- group and zero quantitative values in the control group. ing biomarker for CAA. Using immunohistochemistry we found a significant dif- NDP is a small, secreted protein with a molecular ference with the control groups but not with the AD weight of approximately 15 kD. It has important func- group as HTRA1 marks normal plaque pathology as tion in the formation of the brain vasculature during de- well. HTRA1 is relevant in neurodegeneration as this velopment and in maintenance of a proper functioning protease is involved in the degradation of APP and Aβ BBB [29]. In the adult brain the NDP gene is primarily [40]. In addition, HTRA1 was found to degrade APOE4 expressed by astrocytes [30]. NDP activates the canon- more efficiently than APOE3 and the presence of ical Wnt/β-catenin signalling pathway via the frizzled APOE4 reduces digestion of MAPT by HTRA1 [41]. (Fzd)4/low-density lipoprotein receptor-related protein Moreover, mutations in HTRA1 are the cause of the her- (Lrp)5/6 receptor complex [31]. editary small vessel disease CARASIL (cerebral auto- In neural progenitor cells (NPCs) derived from FAD somal recessive arteriopathy with subcortical infarcts mutant PSEN1 subjects it was found that NDP mRNA is and leukoencephalopathy) [42, 43]. upregulated, but no increase in mRNA was found in AD As part of our LC-MS-MS exploratory study we iden- human temporal lobe [32]. In the retina NDP was found tified other proteins that are potentially interesting for to promote regrowth of capillaries and formation of additional research in relation to CAA, but were not intra-retinal vessels after oxygen-induced retinal damage specifically followed up in this study. For instance, we [33]. Mice overexpressing NDP, had significantly less observed significant high levels of HLA-DR/HLA-DQ in vascular loss following oxygen exposure. Mutations in CAA type-1. This protein is associated with inflamma- the NDP gene result in Norrie disease, which is primar- tion and high numbers of activated microglia [4]. ily an eye disease that leads to blindness. Interestingly, PNP for which a single nucleotide polymorphism was 30–50% of these patients display developmental delay, found to be associated with faster progression of AD intellectual disability, behavioural abnormalities, or [44]. Its relation to CAA is yet unknown. psychotic-like features [34, 35]. In addition, NDP has SUCLG2 which is involved in clearance of Aβ1–42 been shown to protect neurons against excitotoxicity in- [45] was found increased in CAA compared to control duced by NMDA [36]. NDP seems to have protective and even higher levels were observed in the AD cases. properties for both endothelial cells and neurons, but APP was identified with increased levels in CAA. Pep- whether NDP upregulation is beneficial in the context of tide data indicates that the most abundantly detected CAA pathology is unknown. peptide is a part of Aβ, however this analysis cannot dis- Our proteomics data show COL6A2 expression in criminate between Aβ or APP. CAA cases, which is supported by a strong increase in The proteins in this study, that show selective associ- COL6A2 immunoreactivity in the affected brain paren- ation with CAA type-1 pathology might serve as poten- chyma. In addition to COL6A2 we also found COL6A3 tial CAA type-1 biomarkers in patients. As larger vessels Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 17 of 19 are also positive for the markers that were assessed using of the control group (2nd row), AD group (3rd row) and the CAA group IHC, these markers might also be relevant in CAA type- (4th row). Green, expression below the overall mean; red, above the overall mean. The expression profile of case #5 is largely similar to that of 2, although in the case of NDP the intensity of immuno- the control groups but some proteins show a similar expression as in the reactivity is less in larger vessels compared to affected AD and/or CAA groups. (B) Expression values (LFQ values) of several CAA capillaries in CAA. specific proteins identified in this study with case #5 indicated as empty triangle pointing down. Case #5 does not differ from the CAA group in CAA selective markers might be used for pathological these markers. (TIF 1835 kb) assessment of the severity of CAA. The association of Additional file 7: Figure S6. Protein expression of males versus Aβ with the vasculature, and in particular capillaries, is females. Quantitative data on several CAA selective data was plotted with not always obvious in thin microscopic sections. Also, males represented as triangles and females as dots. No clear relationship between gender and protein abundance was observed. (TIF 24739 kb) the use of these proteins as potential diagnostic markers Additional file 8: Figure S7. Immunoreactivity for COL6A2 is equally should be explored. present in leptomeningeal vessels in control, AD and CAA tissue. (TIF 3794 kb) The need for a biomarker for CAA is urgent, in part Additional file 9: Figure S8. Immunohistochemistry of Amyloid-beta, for (early) diagnosis of CAA, but also for stratification of NDP, COL6A2, APOE and APCS on multiple brain regions of a HCHWA-D CAA patients involved in clinical trials for AD. For instance, type-1 case. Brain tissue of a case exhibiting a hereditary form of CAA type-1 was analyzed by immunohistochemistry of Amyloid-beta, NDP, COL6A2, APOE anti-amyloid immunotherapies in development may war- and APCS. Aβ pathology was confirmed and immunoreactivity associated with rant separation of AD patients with or without CAA be- CAA type-1 pathology was found present for all markers. Scale bar in upper cause of expected side effects associated with CAA, left picture represents 100 μm. (TIF 34988 kb) including vasogenic oedema and cerebral microhemor- rhages [46, 47]. In addition, these markers would help to Acknowledgements The authors thank the Netherlands Brain Bank (Amsterdam, the Netherlands) improve the assessment of the safety of anticoagulation for supplying human brain tissue. The authors want to thank Will Hermsen, therapy in patients with CAA as they increase the risk of University Medical Center Utrecht, for performing the immunohistochemistry intracerebral haemorrhage [48]. on the prion tissue. This work was financially supported by Amsterdam Neuroscience and Alzheimer Nederland, grant number AN-16054. David Hondius was supported by the CAVIA project (nr. 733050202), which has Conclusion been made possible by ZonMW, part of the Dutch national ‘Deltaplan for Dementia’: zonmw.nl/dementiaresearch”. In conclusion, we present a set of marker proteins con- taining known and new markers representing valuable Authors’ contributions tools for both clinical and neuropathological diagnosis DCH, KWL, ABS and AJMR designed the experiments. DCH, KNE, THJM, RCvdS performed the experiments. DCH, KWL, JJMH, ABS and AJMR which can contribute to studies investigating the role of interpreted the results. AJMR and MB provided samples and performed the CAA in AD pathology. In addition to their use as bio- pathological characterization. DCH was responsible for writing of the markers, the newly found proteins might be further in- manuscript. KWL, JJMH, MB, PvN, ABS, AJMR made intellectual contributions and contributed to the writing of the manuscript. All authors read and vestigated to increase our understanding of etiology and approved the final manuscript. disease mechanism related to CAA, and ultimately may be used as therapeutic targets. Competing interests A selection of proteins including, but not limited to, NDP, CLU, APOE, HTRA1, APCS, COL6A2 and COL6A3 are part of the patent application P113281EP00. Additional files Publisher’sNote Additional file 1: Figure S1. Coomassie blue staining of the SDS PAGE Springer Nature remains neutral with regard to jurisdictional claims in gels containing the microdissected tissue lysates. (TIF 478 kb) published maps and institutional affiliations. Additional file 2: Figure S2. Total protein fluorescent signal from blots Received: 22 March 2018 Accepted: 23 April 2018 used for immunoblot analysis. Total protein load was visualized using a chemidoc EZ (Bio-Rad) after electroblotting and used to obtain densitometric values which were then used to normalize for total protein References input. (TIF 553 kb) 1. Attems J, Jellinger K, Thal DR, Van Nostrand W (2011) Review: sporadic Additional file 3: Figure S3. Number of proteins detected per individual cerebral amyloid angiopathy. Neuropathol Appl Neurobiol 37:75–93. https:// case. Proteins were quantified based on a minimum of one peptide and doi.org/10.1111/j.1365-2990.2010.01137.x adhering to an FDR of < 0.01. (TIF 19662 kb) 2. Rudolf Thal D, Sue GriYn WT, I de Vos RA, Ghebremedhin E (2008) Cerebral Additional file 4: Table S1. Complete dataset, containing log2 amyloid angiopathy and its relationship to Alzheimer’s disease. Acta transformed quantitative values (LFQ values) of all quantified proteins Neuropathol 115:599–609. https://doi.org/10.1007/s00401-008-0366-2 per individual case. (XLSX 665 kb) 3. Attems J (2005) Sporadic cerebral amyloid angiopathy: pathology, clinical implications, and possible pathomechanisms. Acta Neuropathol 110:345–359. Additional file 5: Figure S4. Clustering analysis of experimental groups https://doi.org/10.1007/s00401-005-1074-9 and individual cases. Clustering analysis and heat maps of the different 4. Richard E, Carrano A, Hoozemans JJ, Van Horssen J, Van Haastert ES, experimental groups (A) and individual cases (B) based on proteins with Eurelings LS, De Vries HE, Thal DR, Eikelenboom P, Van Gool WA, a significant difference (ANOVA, p < 0.05) in expression between any of Rozemuller AJM (2010) Characteristics of dyshoric capillary cerebral the groups. (TIF 709 kb) amyloid angiopathy. J Neuropathol Exp Neurol 69:1158–1167. https:// Additional file 6: Figure S5. Protein expression of CAA case #5 relative doi.org/10.1097/NEN.0b013e3181fab558 to the experimental groups and individual cases. (A) On the left the 5. van Veluw SJ, Kuijf HJ, Charidimou A, Viswanathan A, Biessels GJ, Rozemuller expression profile of case #5 compared to the average expression profile AJM, Frosch MP, Greenberg SM (2016) Reduced vascular amyloid burden at Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 18 of 19 microhemorrhage sites in cerebral amyloid angiopathy. Acta Neuropathol: 24. Verbeek MM, Otte-Höller I, Veerhuis R, Ruiter DJ, De Waal RMW (1998) 1–7. https://doi.org/10.1007/s00401-016-1635-0 Distribution of Aβ-associated proteins in cerebrovascular amyloid of 6. Weller RO, Nicoll J a R (2003) Cerebral amyloid angiopathy: pathogenesis Alzheimer’s disease. Acta Neuropathol 96:628–636. https://doi.org/10. and effects on the ageing and Alzheimer brain. Neurol Res 25:611–616. 1007/s004010050944 https://doi.org/10.1179/016164103101202057 25. Zhan SS, Veerhuis R, Kamphorst W, Eikelenboom P (1995) Distribution of beta amyloid associated proteins in plaques in Alzheimer’s disease and in 7. Bakker ENTP, Bacskai BJ, Arbel-Ornath M, Aldea R, Bedussi B, Morris AWJ, the non-demented elderly. Neurodegeneration 4:291–297 Weller RO, Carare RO (2016) Lymphatic clearance of the brain: perivascular, Paravascular and significance for neurodegenerative diseases. Cell Mol 26. Montañola A, de Retana SF, López-Rueda A, Merino-Zamorano C, Penalba A, Neurobiol 36:181–194. https://doi.org/10.1007/s10571-015-0273-8 Fernández-Álvarez P, Rodríguez-Luna D, Malagelada A, Pujadas F, Montaner 8. Weller RO, Subash M, Preston SD, Mazanti I, Carare RO (2008) Perivascular J, Hernández-Guillamon M (2016) ApoA1, ApoJ and ApoE plasma levels and drainage of amyloid-?? Peptides from the brain and its failure in cerebral genotype frequencies in cerebral amyloid Angiopathy. NeuroMolecular Med amyloid angiopathy and Alzheimer’s disease. In: Brain Pathol, pp 253–266 18:99–108. https://doi.org/10.1007/s12017-015-8381-7 9. Attems J, Jellinger KA (2004) Only cerebral capillary amyloid angiopathy 27. Manousopoulou A, Gatherer M, Smith C, Nicoll JAR, Woelk CH, Johnson M, correlates with Alzheimer pathology?A pilot study. Acta Neuropathol 107: Kalaria R, Attems J, Garbis SD, Carare RO (2017) Systems proteomic analysis 83–90. https://doi.org/10.1007/s00401-003-0796-9 reveals that clusterin and tissue inhibitor of metalloproteinases 3 increase in 10. Eurelings LSM, Richard E, Carrano A, Eikelenboom P, van Gool WA, leptomeningeal arteries affected by cerebral amyloid angiopathy. Rozemuller AJM (2010) Dyshoric capillary cerebral amyloid angiopathy Neuropathol Appl Neurobiol 43:492–504. https://doi.org/10.1111/nan.12342 mimicking Creutzfeldt–Jakob disease. J Neurol Sci 295:131–134. https://doi. 28. Inoue Y, Ueda M, Tasaki M, Takeshima A, Nagatoshi A, Masuda T, Misumi Y, Kosaka T, Nomura T, Mizukami M, Matsumoto S, Yamashita T, Takahashi H, org/10.1016/j.jns.2010.04.020 Kakita A, Ando Y (2017) Sushi repeat-containing protein 1: a novel disease- 11. Thal DR, Ghebremedhin E, Orantes M, Wiestler OD (2003) Vascular pathology associated molecule in cerebral amyloid angiopathy. Acta Neuropathol 134: in Alzheimer disease: correlation of cerebral amyloid Angiopathy and 605–617. https://doi.org/10.1007/s00401-017-1720-z arteriosclerosis/Lipohyalinosis with cognitive decline. J Neuropathol Exp Neurol 29. Engelhardt B, Liebner S (2014) Novel insights into the development and 62:1287–1301. https://doi.org/10.1093/jnen/62.12.1287 maintenance of the blood–brain barrier. Cell Tissue Res 355:687–699. 12. Braak H, Braak E (1991) Neuropathological stageing of Alzheimer-related https://doi.org/10.1007/s00441-014-1811-2 changes. Acta Neuropathol 82:239–259 13. Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, Vogel FS, 30. Ye X, Smallwood P, Nathans J (2011) Expression of the Norrie disease gene Hughes JP, van Belle G, Berg L (1991) The consortium to establish a registry (Ndp) in developing and adult mouse eye, ear, and brain. Gene Expr for Alzheimer’s disease (CERAD): part II. Standardization of the Patterns. https://doi.org/10.1016/j.gep.2010.10.007 neuropathologic assessment of Alzheimer’s disease. Neurology 41:479–479. 31. Xu Q, Wang Y, Dabdoub A, Smallwood PM, Williams J, Woods C, Kelley MW, https://doi.org/10.1212/WNL.41.4.479 Jiang L, Tasman W, Zhang K, Nathans J (2004) Vascular development in the retina and inner ear: control by Norrin and Frizzled-4, a high-affinity ligand- 14. Thal DR, Rüb U, Orantes M, Braak H (2002) Phases of Aβ-deposition in the receptor pair. Cell 116:883–895. https://doi.org/10.1016/S0092-8674(04)00216-8 human brain and its relevance for the development of AD. Neurology 58: 1791–1800. https://doi.org/10.1212/WNL.58.12.1791 32. Sproul AA, Jacob S, Pre D, Kim SH, Nestor MW, Navarro-Sobrino M, Santa- 15. Montine TJ, Phelps CH, Beach TG, Bigio EH, Cairns NJ, Dickson DW, Maria I, Zimmer M, Aubry S, Steele JW, Kahler DJ, Dranovsky A, Arancio O, Duyckaerts C, Frosch MP, Masliah E, Mirra SS, Nelson PT, Schneider JA, Thal Crary JF, Gandy S, Noggle SA (2014) Characterization and molecular DR, Trojanowski JQ, Vinters HV, Hyman BT (2012) National institute on profiling of PSEN1 familial alzheimer’s disease iPSC-derived neural aging-Alzheimer’s association guidelines for the neuropathologic assessment progenitors. PLoS One. https://doi.org/10.1371/journal.pone.0084547 of Alzheimer’s disease: a practical approach. Acta Neuropathol 123:1–11. 33. Ohlmann A, Seitz R, Braunger B, Seitz D, Bösl MR, Tamm ER (2010) Norrin https://doi.org/10.1007/s00401-011-0910-3 promotes vascular regrowth after oxygen-induced retinal vessel loss and 16. Verwey NA, Hoozemans JJM, Korth C, van Royen MR, Prikulis I, Wouters D, suppresses retinopathy in mice. J Neurosci 30 HAM T, van Haastert ES, Schenk D, Scheltens P, Rozemuller AJM, Blankenstein 34. Braunger BM, Tamm ER (2012) The different functions of Norrin. Adv Exp MA, Veerhuis R (2013) Immunohistochemical characterization of novel med biol. https://doi.org/10.1007/978-1-4614-0631-0_86 monoclonal antibodies against the N-terminus of amyloid β-peptide. Amyloid 35. Sims KB (1993) NDP-related retinopathies. University of Washington, Seattle 20:179–187. https://doi.org/10.3109/13506129.2013.797389 36. Seitz R, Hackl S, Seibuchner T, Tamm ER, Ohlmann A (2010) Norrin mediates 17. Hondius DC, Van Nierop P, Li KW, Hoozemans JJM, Van Der Schors RC, neuroprotective effects on retinal ganglion cells via activation of the Wnt/- Van Haastert ES, Van Der Vies SM, Rozemuller AJM, Smit AB (2016) catenin signaling pathway and the induction of neuroprotective growth Profiling the human hippocampal proteome at all pathologic stages of factors in Muller cells. J Neurosci 30:5998–6010. https://doi.org/10.1523/ Alzheimer’s disease. Alzheimers Dement 12:654–668. https://doi.org/10. JNEUROSCI.0730-10.2010 1016/j.jalz.2015.11.002 37. Ricard-Blum S (2011) The collagen family. Cold Spring Harb Perspect 18. Cox J, Mann M (2008) MaxQuant enables high peptide identification rates, Biol 3:a004978–a004978. https://doi.org/10.1101/cshperspect.a004978 individualized p.P.B.-range mass accuracies and proteome-wide protein 38. Kuo HJ, Maslen CL, Keene DR, Glanville RW (1997) Type VI collagen anchors quantification. Nat Biotechnol 26:1367–1372. https://doi.org/10.1038/nbt.1511 endothelial basement membranes by interacting with type IV collagen. 19. Cox J, Hein MY, Luber CA, Paron I, Nagaraj N, Mann M (2014) Accurate J Biol Chem 272:26522–26529 proteome-wide label-free quantification by delayed normalization and 39. Cheng JS, Dubal DB, Kim DH, Legleiter J, Cheng IH, Yu G-Q, Tesseur I, maximal peptide ratio extraction, termed MaxLFQ. Mol Cell Proteomics 13: Wyss-Coray T, Bonaldo P, Mucke L (2009) Collagen VI protects neurons 2513–2526. https://doi.org/10.1074/mcp.M113.031591 against Abeta toxicity. Nat Neurosci 12:119–121. https://doi.org/10.1038/ 20. Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, Mann M, Cox J nn.2240 (2016) The Perseus computational platform for comprehensive analysis of 40. Grau S, Baldi A, Bussani R, Tian X, Stefanescu R, Przybylski M, Richards P, (prote)omics data. Nat Methods 13:731–740. https://doi.org/10.1038/nmeth.3901 Jones SA, Shridhar V, Clausen T, Ehrmann M (2005) Implications of the 21. Arvanitakis Z, Leurgans SE, Wang Z, Wilson RS, Bennett DA, Schneider JA serine protease HtrA1 in amyloid precursor protein processing. Proc Natl (2011) Cerebral amyloid angiopathy pathology and cognitive domains in Acad Sci 102:6021–6026. https://doi.org/10.1073/pnas.0501823102 older persons. Ann Neurol 69:320–327. https://doi.org/10.1002/ana.22112 41. Chu Q, Diedrich JK, Vaughan JM, Donaldson CJ, Nunn MF, Lee K-F, 22. Boyle PA, Yu L, Nag S, Leurgans S, Wilson RS, Bennett DA, Schneider JA Saghatelian A (2016) HtrA1 proteolysis of ApoE in vitro is allele selective. (2015) Cerebral amyloid angiopathy and cognitive outcomes in community- J Am Chem Soc 138:9473–9478. https://doi.org/10.1021/jacs.6b03463 based older persons. Neurology 85:1930–1936. https://doi.org/10.1212/WNL. 42. Hara K, Shiga A, Fukutake T, Nozaki H, Miyashita A, Yokoseki A, Kawata H, Koyama A, Arima K, Takahashi T, Ikeda M, Shiota H, Tamura M, Shimoe Y, 23. Manousopoulou A, Gatherer M, Smith C, Nicoll JAR, Woelk CH, Johnson M, Hirayama M, Arisato T, Yanagawa S, Tanaka A, Nakano I, Ikeda S, Yoshida Kalaria R, Attems J, Garbis SD, Carare RO (2016) Systems proteomic analysis Y,Yamamoto T,Ikeuchi T,KuwanoR,Nishizawa M,Tsuji S,OnoderaO reveals that clusterin and tissue inhibitor of metalloproteinases 3 increase in (2009) Association of HTRA1 mutations and familial ischemic cerebral leptomeningeal arteries affected by cerebral amyloid angiopathy. small-vessel disease. N Engl J Med 360:1729–1739. https://doi.org/10.1056/ Neuropathol Appl Neurobiol. https://doi.org/10.1111/nan.12342 NEJMoa0801560 Hondius et al. Acta Neuropathologica Communications (2018) 6:46 Page 19 of 19 43. Tikka S, Baumann M, Siitonen M, Pasanen P, Pöyhönen M, Myllykangas L, Viitanen M, Fukutake T, Cognat E, Joutel A, Kalimo H (2014) CADASIL and CARASIL. Brain Pathol 24:525–544. https://doi.org/10.1111/bpa.12181 44. Tumini E, Porcellini E, Chiappelli M, Conti CM, Beraudi A, Poli A, Caciagli F, Doyle R, Conti P, Licastro F (2007) The G51S purine nucleoside phosphorylase polymorphism is associated with cognitive decline in Alzheimer’s disease patients. Hum Psychopharmacol Clin Exp 22:75–80. https://doi.org/10.1002/hup.823 45. Ramirez A, van der Flier WM, Herold C, Ramonet D, Heilmann S, Lewczuk P, Popp J, Lacour A, Drichel D, Louwersheimer E, Kummer MP, Cruchaga C, Hoffmann P, Teunissen C, Holstege H, Kornhuber J, Peters O, Naj AC, Chouraki V, Bellenguez C, GerrishA, HeunR,Frolich L,Hull M,Buscemi L, Herms S, KolschH, ScheltensP, Breteler MM, Ruther E, Wiltfang J, Goate A, Jessen F, Maier W, Heneka MT, Becker T, Nothen MM (2014) SUCLG2 identified as both a determinator of CSF a 1-42 levels and an attenuator of cognitive decline in Alzheimer’sdisease. Hum Mol Genet 23:6644–6658. https://doi.org/10.1093/hmg/ddu372 46. Boche D, Zotova E, Weller RO, Love S, Neal JW, Pickering RM, Wilkinson D, Holmes C, Nicoll JAR (2008) Consequence of Abeta immunization on the vasculature of human Alzheimer’s disease brain. Brain 131:3299–3310. https://doi.org/10.1093/brain/awn261 47. Sperling R, Salloway S, Brooks DJ, Tampieri D, Barakos J, Fox NC, Raskind M, Sabbagh M, Honig LS, Porsteinsson AP, Lieberburg I, Arrighi HM, Morris KA, Lu Y, Liu E, Gregg KM, Brashear HR, Kinney GG, Black R, Grundman M (2012) Amyloid-related imaging abnormalities in patients with Alzheimer’s disease treated with bapineuzumab: a retrospective analysis. Lancet Neurol 11:241–249. https://doi.org/10.1016/S1474-4422(12)70015-7 48. Banerjee G, Carare R, Cordonnier C, Greenberg SM, Schneider JA, Smith EE, Van Buchem M, Van Der Grond J, Verbeek MM, Werring DJ (2017) The increasing impact of cerebral amyloid angiopathy: essential new insights for clinical practice. J Neurol Neurosurg Psychiatry 88:982–994. https://doi.org/ 10.1136/jnnp-2016-314697

Journal

Acta Neuropathologica CommunicationsSpringer Journals

Published: Jun 4, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off