Astrocyte-specific druggable protein as PET-ligand target for early detection of Alzheimer’s disease

(1)

INOM

EXAMENSARBETE BIOTEKNIK, AVANCERAD NIVÅ, 30 HP

STOCKHOLM SVERIGE 2017,

Astrocyte-specific druggable protein as PET-ligand target for early detection of Alzheimer’s disease

REBECCA SJÖBERG

KTH

SKOLAN FÖR BIOTEKNOLOGI

(2)

Royal Institute of Technology

Astrocyte-specific druggable protein as PET-ligand target for early detection of Alzheimer’s disease

Rebecca Sjöberg

Degree project in biotechnology, second cycle, 30 hp Stockholm, Sweden 2017

(3)

2

Abstract

Still no cure for Alzheimer’s disease (AD) exist and number of individuals with AD will increase with an aging population. Astrocytes have recently been associated with the early pathological changes that occur in the development of AD. Identifying astrocyte-specific proteins with altered expression patterns in the presymptomatic phase of AD and targeting these proteins with available FDA approved drugs could be a strategy to early detect the disease. These drugs will serve as templates for the development of novel PET-ligands and be more specific than the ones already in use. The aim was to find druggable proteins expressed in astrocytes in human frontal cortex (FCX) whose genes are either up- or downregulated in AD. By studying expression pattern in astrocytes and identifying druggable proteins targeted by FDA approved drugs, potential PET ligands could be developed that enable detection and diagnosis of AD in a presymptotic phase. Four potential druggable astrocyte- specific proteins with altered expression in AD were selected based on proteomic and transcriptomic data from human cerebral cortex. Proteins were studied using a sensitive fluorescence

histochemistry approach based on thyramide signal amplification method. Cellular distribution and disease associated changes were investigated using tissue micro array (TMA) containing FCX tissue of AD subjects and non-demented age-matched controls. Creatine kinase B (CKB) was considered as being astrocyte-specific with an upregulation on transcript level but reduced protein quantity seen with the immunofluorescence analysis in the AD brain. The reduced protein level was confirmed using western blot analysis. Variable levels of CKB were observed in the control subjects where individuals with signs of ongoing disease processes had reduced levels of CKB. Decreased levels of CKB seem to be affected by ongoing disease processes and might therefore be used as an early marker for AD.

(4)

3

Abstrakt

Det finns ännu inget botemedel för Alzheimer’s sjukdom (AS) och antalet individer med AS fortsätter att öka i takt med att befolkningen blir äldre. Astrocyter har nyligen sammankopplats med de tidiga patologiska förändringarna som sker i utvecklingen av AS. Genom att identifiera astrocytspecifika proteiner med förändrade uttrycksmönster i den pre-symtomatiska fasen av AS kan det vara möjligt att upptäcka sjukdomen i ett tidigt skede genom att använda FDA godkända läkemedel som kan binda till proteinerna och därigenom fungera som mall för utvecklingen av nya PET-ligander. Syftet var att hitta upp- eller nedreglerade astrocyte-specifika proteiner i den frontala hjärnbarken som FDA godkända läkemedel kan binda till. Genom att studera proteinuttryck i astrocyter kan potentiella PET-ligander utvecklas som kan göra att AS kan upptäckas i en pre-symtomatisk fas. Fyra potentiella astrocyte-specifika proteinkandidater med förändrat uttryck i AS valdes ut baserat på proteomik and transkripomik data från mänsklig hjärnbark. Proteinerna studerades genom att använda en

fluorescerande histokemisk metod baserat på thyramide signal amplifiering. Den cellulära

distributionen och sjukdomsassocierade förändringarna undersöktes i mänsklig vävnad av frontala hjärnbarken från AS och icke-dementa åldersmatchande kontroller. Kreatinkinas B (CKB) ansågs astrocytspecifikt och var uppreglerad på transkriptnivå men en reducerad proteinnivå noterades vid den immunofluorescerande analysen i AS hjärnan. Den reducerade proteinnivån bekräftades genom western blot analys. Det noterades även att nivån av CKB varierade i kontrollgruppen, där kontroller med tecken på en pågående sjukdomsprocess hade lägre nivåer av CKB. En minskad proteinnivå av CKB verkar påverkas av pågående sjukdomsprocesser och proteinet har därför potential att kunna användas som en tidig sjukdomsmarkör.

(5)

1. Introduction

1.1 Alzheimer’s disease

Alzheimer’s disease (AD) is a progressive neurodegenerative disease and the risk of developing AD increases with age [1]. It is estimated that around 100 000 individuals in Sweden live with the disease today and this number is expected to increase with an aging population [2, 3].

The AD brain is characterized by a decrease in cerebral glucose metabolism [4] as well as accumulation of extracellular amyloid plaques and intraneuronal neurofibrillary tangles (NFTs) [5, 6, 7]. Amyloid plaques are formed from amyloid-beta (Aβ) peptides that aggregate together [2] whereas NFTs are found inside neurons and consist of hyperphosphorylated and misfolded tau proteins [8]. Accumulation of amyloid plaques and NFTs begins in temporal cortex (TCX) and hippocampus respectively, then gradually propagate to other brain regions in a characteristic pattern (figure 1) [5, 8, 9, 10].

Degree of amyloid plaques and NFTs in the brain is classified into different stages

(Appendix 8.1.1, table 3 and 4). As the disease progresses damage and loss of neurons occur [2, 5, 6, 7]. Loss of neurons lead to cortical atrophy (brain shrinkage) and results in remaining neurons being less well-connected to each other (figure 2), which contributes to the clinical symptoms associated with AD [7, 11]. Early symptoms of AD are memory decline and learning disabilities. Later on physical disabilities and difficulties swallowing occur, which result in death due to

malnutrition or pneumonia [1, 12].

1.2 A quest for better PET-ligands

Reduction in glucose metabolism and accumulation of amyloid plaques and NFTs begin approximately ten years before clinical presentation of symptoms [2, 4, 5, 8] at which stage irreversible brain damage has already occurred [13]. At present, no cure for AD exist.

Drugs available are only temporarily relieving symptoms by increasing the effect of the neurotransmitter acetylcholine [2, 5, 14]. If these drugs are administered before

symptoms appear (in a preclinical stage) they are more effective in reducing

neurodegeneration, thereby delaying disease progression [15].

Several methods are used to diagnose AD, one is Positron Emission Tomography (PET)

imaging [16]. This is a non-invasive, cerebral imaging method to study biochemical functions for early disease detection [15, 17, 18]. PET-imaging is based on radiolabelled

Figure 1. The abnormal protein aggregates are classified as amyloid load ranging from A-C respectively Braak stages I-VI depending on degree of spread in the brain. Increased density of shading indicates an increased deposition of the protein aggregates [10].

Figure 2. Comparison between healthy brain (left) and AD brain (right). Accumulation of amyloid plaques and NFTs causes neuronal loss and contribute to brain shrinkage [11].

(8)

7 molecules (PET-ligands) administered into

subjects, where they target proteins

associated with a particular disease. Detection of target binding is made by the emission of β radiation [16, 17]. Finding novel PET-ligands is however associated with expensive and time consuming research, and the success rate is usually low [17]. At present PET-imaging is used to diagnose AD by examining glucose metabolism (FDG-PET) or depositions of Aβ (amyloid-PET) [19, 20], but these pathological alterations are not specific for AD. An option is to use available FDA approved drugs to serve as templates for the development of novel PET-ligands that can target proteins (druggable proteins) associated with AD pathology [15].

1.3 Astrocytes as PET-ligand targets

Astrocytes are glial cells found in the central nervous system (CNS) and has recently been associated with the pathological development of AD [5]. Different sub-types of astrocytes exist, but all have a characteristic branched cell morphology that enables physical connection with blood vessels and neurons.

Astrocytes express numerous receptors and signaling molecules to maintain homeostasis, supply neurons with metabolites and regulate availability of ions and neurotransmitters [21, 22]. The many receptors and signalling molecules enable astrocytes to respond fast if alterations, such as pathological events, occur in the surrounding environment. In the presence of such changes, astrocytes become reactive and alter gene expression [23, 24].

1.4 The druggable proteome

Human Protein Atlas (HPA) is an open access database and uses an antibody-based approach of both commercial and non- commercial antibodies for immunostaining of cells and tissues. It containing both

immunofluorescence (IF) and

immunohistochemistry (IHC) images together with an application-specific validation for each antibody [27]. Human tissue proteome

consists of seven different subcategories of

protein-coding genes with information about RNA expression as well as expression and localization of their corresponding proteins.

One subcategory is The druggable proteome, that consists of proteins targeted by FDA approved drugs. Currently (2017-05-17) 646 different proteins are targeted by FDA approved drugs [28].

1.5 Immunofluorescence visualizes proteins in tissue

Proteins in tissues can be visualized using immunofluorescence and two approaches exist (direct and indirect) for detection. Direct uses a primary antibody conjugated with a fluorophore and indirect uses both a primary and a secondary antibody. The secondary antibody, conjugated either to a fluorophore or an enzyme, binds to the primary unlabelled antibody attached to the target protein [29, 30]. Tyramide Signal Amplification (TSA) is a method that increases sensitivity of

fluorescence signals to detect low abundant target proteins by using an enzyme driven accumulation of fluorescence signal. When a secondary antibody conjugated with the enzyme horseradish peroxidase (HRP) is incubated with the TSA reagent, HRP catalyses covalent depositions of fluorophores and results in an increased fluorescence signal [30]

which reflects protein quantity [31]. Accuracy of the immunofluorescence staining depends on antibody specificity and its selectivity toward target protein is confirmed using western blotting (WB) which also quantifies proteins [32, 33].

1.6 Aim

The aim was to find druggable proteins expressed in astrocytes in human frontal cortex (FCX) whose genes are either up- or downregulated in AD. By studying expression pattern in astrocytes and identifying

druggable proteins targeted by FDA approved drugs, potential PET ligands could be found that enable detection and diagnosis of AD.

(9)

8

2. Materials and methods

This section describes how the selection of candidate proteins was made and how the experiments were performed.

2.1 Data analysis

2.1.1 Proteomic analysis of immunohistochemistry images

Expression profiles of proteins known to be astrocyte specific were studied using IHC images from normal human cerebral cortex found at HPA webpage. Scores were given based on localisation and intensity of the protein staining. IHC images found in The druggable proteome were then analysed using the same approach. Only proteins that were glial specific, dominantly expressed in cerebral cortex and had validated HPA antibodies for IHC were selected.

2.1.2 Transcriptomic analysis of RNA- sequencing data

RNA-sequencing (RNA-seq) data delivered as read counts was provided from Mayo Clinic [34] The data was based on RNA isolated from TCX that had underwent 101 bp paired-end sequencing in IlluminaHiSeq 2000 platform (Illumina, San Diego, CA). Samples derived from 93 post-mortem AD subjects (Braak stage ≥ 4) and 276 controls without

neurodegenerative diagnose (disease stages ranged from I 0 to III A). The data was normalized using Fragments Per Kilobase per Million (FPKM). Mean value and standard deviation (SD) were calculated and a t-test (p<0.05) was used to exclude transcripts not showing any significant differential expression (DE) between AD and control subject

(Appendix 8.2, table 6).

2.2 Human brain tissue samples

Post-mortem brain tissue samples from human FCX was obtained from the Netherlands Brain Bank (NBB). The brain tissues were used in Tissue Micro Arrays (TMAs) for immunofluorescence staining and as larger brain tissue sections for western blot (PMID: 16957166). Cerebrospinal fluid (CSF),

blood plasma and medical records from subjects were included. Each TMA had 29 tissues arrayed in duplicates (10 AD, 10 Lewy body dementia (DLB) and 9 non-demented age-matched controls (hereafter denoted control subjects)). DLB was not included in this project. Disease stages of AD subjects ranged from V C to VI C and for control subjects I 0 to I B. More information of the subjects is found in Appendix 8.1.2, table 5.

2.3 Immunoflourescence analysis

Immunofluorescence histochemistry was performed on tissue from FCX. Co-existence of proteins was initially made with the astrocytic marker S100β. The analysis was performed on control subjects, since protein expression patterns usually are altered in

neurodegenerative conditions and thus not mirror the actual physiology of the cells. If co- existence occurred, the protein was

investigated further to evaluate if the protein was astrocyte-specific by counting number of overlaps with cell type markers for astrocytes (S100β, GFAP and ALDH1L1), oligodentrocytes (CNP), neurons (NeuN) and micro glia cells (Iba-1). Different protocols were used for the immunofluorescence immunohistochemistry and below is the general workflow described.

More thorough descriptions are found in Appendix 8.3 together with antibodies (primary and secondary) and their dilution factors (Appendix 8.4.1, table 8 and 9).

Recipes of bufferts and solutions are found in Appendix 8.5.

Paraffin embedded TMAs were dewaxed (Bond Dewax Solution, AR9222, Leica

Biosystems), treated with citric acid based pH 6 solution (Bond Epitope Retrieval Solution 1, AR9640, Leica Biosystems), blocked in peroxidase blocking buffer (Novocastra Peroxidase Block, RE7101, Leica Biosystems) and rehydrated in PBS using fully automated Leica BOND-RX IHC/ISH Stainer. IHC was then performed manually. TMAs were incubated with primary antibodies in Primary Antibody Diluent for 16 h at 4°C, washed 3x 5 min in TBS-Tween®20 and incubated for 30 min in

(10)

9 room temperature (RT) with TNB Blocking

Buffer. Secondary antibodies and nuclei staining (HOECHST 33342, 62249,

ThermoFisher) were diluted in TNB, incubated with TMAs for 60 minutes in RT, then washed 3x 5 min in TBS-Tween®20. Each TMA was incubated with TSA reagent FITC, Cy3.5 or Cy5 (Fluorescein Amplification Reagent, Perkin Elmer) diluted in TSA amplification diluent (1x Plus Amplification Diluent, FP1135, Perkin Elmer) for 15 min in RT, then washed 3x 5 min in TBS-Tween®20. TMAs were dipped in 70 % EtOH, incubated with Sudan Black solution for 30 min in RT then again dipped in 70 % EtOH.

TMAs were put in PBS for 1 min before mounted with 50 µl PVA-DABCO mounting medium (10981, Sigma Aldrich).

2.4 Image acquisition and analysis

Fluorescence microscope images were scanned using a scanning microscope (Meta Systems, Alltlussheim, Germany) and software Metafer 5 (V 3.12.6) was used to generate four-channel fluorescence images. Whole slides were pre-scanned using 2.5x objective to create a field of view position map that was scanned using 10x objective. Filters were set for DAPI, FITC, Cy3.5 and Cy5. Images were stitched using VSlide 1.1.107. Co-existence and cell counting were analysed by manual inspection using VSViewer V 2.1.112. Protein quantity in fluorescence images was measured using the script macro in ImageJ50i as average pixel intensity over whole tissue area

(Appendix 8.6).

Only tissues that met following criterion were analysed: (i) the IHC staining was equally distributed throughout the tissue; (ii) the cell type markers that were used reflected the characteristic morphology of the cells; (iii) cell nucleus were present in the IHC stained cells;

and (iv) the studied tissue was at least 80 % intact.

2.5 Western blot analysis

Frozen brain tissue samples of FCX from same subjects as were used for IHC were

homogenized with 500 µl Lysis Buffer in a

Tissue Grinder Potter, poured into separate Eppendorf tubes and centrifuged in VWR Micro Star 17R at 16xG for 30 minutes at 4°C.

Supernatants were collected and diluted 1x, 10x and 100x with distilled water (dH2O). BSA (Quick Start™ Bovine Serum Albumine (BSA) Standard 2mg/ml, Bio-Rad, 5000206) was diluted 0.5, 1, 2 and 4 µg/ml in dH2O for Bradford Standard Curve (Appendix 8.7.1, table 12, figure 9). 50 µl supernatant dilution and 150 µl dye reagent (Quick Start™ Bradford 1x Dye Reagent, Bio-Rad, 5000205) were added in duplicates on a 96-well plate.

Absorbance was measured at 595 nm using spectrophotometer SpectraMax 250 Microplate Reader (Marshall Scientific, Hampton). Protein concentrations were determined and samples prepared for WB (Appendix 8.7.2, table 13). Samples were heat treated for 5 min at 95°C, centrifuged in VWR Micro Star 17R at 16xG for 1 minutes at 20°C and loaded in Mini-PROTEAN® TGX™ Pre Cast Gels 4-20% (4561096 , Bio-Rad). Loading Buffer (1x TNE and 5x Laemmli Sample Buffer;

diluted 5:1) was used in the two outer wells on each side and Standard Buffer (Loading Buffer and Protein Standard Marker (Precision Plus Protein™ WesternC™ Standards,

1610376, Bio-Rad); diluted 1:1) was used as ladder. The gel was run in Mini-PROTEAN®

Tetra System (Bio-Rad, United States) with 100 ml 10x TGS buffer (1610772, Bio-Rad) diluted 1:10 in dH2O and 40 V was applied for 30 minutes, then increased to 200 V for 30 min.

Proteins were transferred to 0.2 µm Trans- Blot® Turbo™ Mini PVDF Transfer Membranes (1704156, Bio-Rad) in Trans-Blot® Turbo™

Transfer System (Bio-Rad, United States) using protocol Turbo TGX (2.5 A, 25 V, 7 min), then incubated with BSA Blocking Buffer on a rotating plate over night at 4°C. Antibodies used in western blot analysis are found in Appendix 8.7.3, table 14 and 15. Primary antibodies were diluted in BSA Blocking Buffer and incubated with membranes on a rotating wheel for 2 h in RT, then washed 3x 5 min in

(11)

10 TBS-Tween®20. HRP-conjugated secondary

antibodies were diluted in Milk Blocking Buffer and incubated with membranes for 1 h in RT, washed 3x 5 min in TBS-Tween®20, and put in a chemilumiscence solution (Clarity™ Western ECL Substrate, Luminol/enhancer solution, Bio-Rad; Clarity™ Western ECL Substrate, Peroxide solution, Bio-Rad; diluted 1:1) for 1 min in RT. Blots were scanned in ChemiDoc™

MP Imaging System (Bio-Rad, United States) and quantified by measuring band pixel intensities using software Image™ Lab.

2.6 Databases and search strategy

Information about AD, astrocytes and

methods were gathered from scientific papers and were found by utilise PubMed search engine. All papers have been peer-reviewed and are published in known scientific journals.

Webpages from manufactures were used to find information of the technical equipment used in the experiments.

The web based database Gene Ontology (GO) was used to study GO-annotations (molecular function, biological processes and cellular compartments) of the proteins by utilizing the search and browsing tool AmiGO. Information of proteins was provided from Uniprot.

DrugBank database was utilized to identify the drugs that targets the astrocyte-specific druggable proteins found.

2.7 Ethical considerations

The tissues used in the project derived from post-mortem donors in Netherlands that had provided informed consent. All subjects were anonymized and had different ID numbers.

(12)

11

3. Result

In the following section are the results from data analysis, immunofluorescence

histochemistry and western blot analysis presented.

3.1 Data analysis

3.1.1 Proteomic analysis of immunohistochemistry images

Based on protein expression profiles in IHC images from normal human cerebral cortex (figure 3) and scoring of protein staining (table

1), the proteomic analysis resulted in four druggable proteins (MAP4 had two transcripts, marked #1 and #2) that were expected to be astrocyte-specific (full names

found in Appendix 8.2, table 6 together with their GO-notations). Cell type markers known to be astrocyte specific stained the

characteristic morphology associated with astrocytes (upper row in figure 1). Based on similar staining pattern, proteins (lower row in figure 1) were selected. CKB had a staining pattern similar to the astrocyte specific markers, whereas the other proteins mainly were localised to cell nucleus. Level of protein expression among the selected proteins varied in cerebral cortex and is reflected in the IHC

images below. CKB had the highest protein expression whereas ALOX5 was only expressed in low amount.

Table 1. Scores were given based on localisation and intensity of protein staining. Proteins known to be expressed in astrocytes are in blue text and proteins in black are the proteins assumed to be astrocyte specific. Scoring were based on following criterion: Staining exist: + ; Very strong staining: ++ ; No staining: -.

Protein Nuclei staining Cell body staining Processing staining End feet staining

GFAP - ++ ++ ++

S100β + + + -

ALDH1L1 ++ ++ - -

GLT-1 - + + ++

AQP4 - ++ + -

CKB - ++ + -

MAP4 #1 ++ + + -

MAP4 #2 ++ ++ + -

CACNA1A - - ++ ++

ALOX5 + - - -

Figure 3. IHC images from normal human cerebral cortex available at HPA website was used to study expression and distribution of proteins. The upper row visualizes proteins known to be expressed in astrocytes. The lower row visualizes images of druggable proteins (two transcripts available for MAP4) that were selected.

(13)

12 3.1.2 Transcriptomic analysis of RNA-

sequencing data

Mean FPKM values from RNA-seq data revealed that all genes except CACNA1A were upregulated in AD subjects (figure 4). Based on t-test (p<0.05) ALOX5 had no significant DE between AD and control subjects and was therefore excluded. CKB had the highest mean FPKM values both in AD and control group and the difference between the groups was also the most significant (p=1.72·10^-5). Largest spread of mean FPKM was seen for the CKB transcripts among AD subjects (432.159

±110.347), whereas for the control subjects varied mean FPKM most in MAP4 #1 (237.019

±140.467). More information about the transcripts is found in Appendix 8.2, table 6.

3.2 Immunofluorescence analysis

3.2.1 Co-existence analysis

Pre-selected HPA antibodies were used to visualise the localisation of druggable proteins in FCX tissue and examine their co-existence with astrocytic marker S100β, which was seen as an overlap. Overlap indicated that the protein was localised in astrocytes and seen as yellow colour when images of the druggable

protein and cell type marker were merged together (figure 5). All images are from the same control tissue (C3, 2011-028) and x100 magnification was used. Images shows, from the left (A) astrocytic marker S100β with DAPI, in the middle (B) druggable protein with DAPI and to the right (C) images of A and B merged together.

Co-existence with S100β occurred for CKB, MAP4 #1 and MAP4 #2 and these druggable proteins were therefore considered as being expressed in astrocytes. CACNA1A did not overlap with S100β and was not investigated further. The localisation of CKB, MAP4 #1 and MAP4 #2 were mainly in the cell body, but CKB was also present in end feets (seen as branched staining in figure 5). The two MAP4 proteins had similar protein distribution, but MAP4 #1 was more localised around cell nucleus than MAP4 #2.

3.2.2 Cell-counting

To examine which brain cell types the druggable proteins were localised in co-

Figure 5. Co-existence analysis of candidate proteins indicated that all candidate proteins except CACNA1A were localised in astrocytes. Arrows highlights some of the overlaps.

Figure 4 The diagram compares the mean FPKM values of control and AD subjects based on data from Mayo Clinic.

(14)

13 existence and cell-counting were made with

different cell type markers (table 2, figure 6).

Cell-counting was measured as the ratio of numbers of protein overlap that occurred over total amount counted cell type markers.

CKB had 100 % overlap with S100β and overlapped almost entirely with ALDH1L1 and GFAP, all of which are cell type markers for astrocytes. Cells that were positively stained for S100β and ALDH1L1 overlapped with nearly all proteins stained for CKB. For the GFAP positive stained cells overlapped less

than with CKB. No overlap was seen with neuronal, oligodendrocyte and microglial markers. The two MAP4 proteins did mainly overlap with cells positively stained for the three astrocytic cell markers, but overlap also occurred with neuronal cell type marker and, although not significant, oligodendrocytic cell type marker. MAP4 #2 overlapped more with the astrocytic cell type markers than MAP4 #1.

Almost all GFAP positive stained cells overlapped with MAP4 #1, but only half the MAP4 #1 stained proteins overlapped with GFAP.

Table 2. Cell counting was used to identify which cell types the druggable proteins were expressed in and is presented below in % as the ratio between number of overlaps that were seen between a druggable protein with a cell type marker and total cells that were counted.

CKB S100β

(astrocyte s)

ALDH1L1

(astrocytes)

GFAP

(astrocytes)

Iba-1

(microglial cells)

NeuN

(neuron) CNP

(oligodendro cytes)

CKB 100.0 % 100.0 % 95.0 % 92.2 % 0.0 % 0.0 % 0.0 %

S100β 100.0 % 100.0 % - - - - -

ALDH1L1 99.2 % - 100.0 % - - - -

GFAP 39.4 % - - 100.0 % - - -

Iba-1 0.0 % - - - 100.0 % - -

NeuN 0.0 % - - - - 100.0 % -

CNP 0.0 % - - - 100.0 %

MAP4 #1 S100β ALDH1L1 GFAP Iba-1 NeuN CNP MAP4 #1 100.0 % 61.1 % 30.2 % 54.4 % 0.0 % 24.0 % 0.9 %

S100β 68.9 % 100.0 % - - - - -

ALDH1L1 52.5 % - 100.0 % - - - -

GFAP 98.9 % - - 100.0 % - - -

Iba-1 0.0 % - - - 100.0 % - -

NeuN 16.0 % - - - - 100.0 % -

CNP 0.9 % - - - 100.0 %

MAP4 #2 S100β ALDH1L1 GFAP Iba-1 NeuN CNP MAP4 #2 100.0 % 77.3 % 29.0 % 97.8 % 0.0 % 15.0 % 2.2 %

S100β 89.3 % 100.0 % - - - - -

ALDH1L1 50.0 % - 100.0 % - - - -

GFAP 82.2 % - - 100.0 % - - -

Iba-1 0.0 % - - - 100.0 % - -

NeuN 10.0 % - - - - 100.0 % -

CNP 1.1 % - - - 100.0 %

(15)

14

Figure 6. Co-existence analysis of candidate proteins with different cell type markers.

(16)

15

3.3 Top astrocyte-specific druggable protein

The result from co-existence analysis and cell- counting revealed that CKB was the druggable protein that was the most astrocyte-specific of the proteins analysed. The FDA approved drug and that targets CKB is creatine. Creatine is a peptide synthesized from L-arginine, glycine and L-methionine in the kidneys, liver, pancreas and, to a certain extent in the brain.

Uptake of creatine to the brain from blood occur through an active transport system over the blood brain barrier (BBB) [35].

3.4 Protein quantification

Protein quantification of fluorescence images (Appendix 8.4.2, table 10 and 11) showed that CKB was low in all AD subjects (16.42 ±4.29).

For the control subjects was the average quantity higher, but a larger spread was noted (40.25 ±29.90). Five control subjects had similar quantity of CKB as the AD subjects (figure 7, graph A). A t-test (p<0.05) showed no statistical significant difference (p=0.056) between the AD and control group.

Possible correlations between protein quantity of CKB and presence of either amyloid plaques or NFTs were studied. Graph B in figure 7 shows how mean intensity of CKB degreases with degree of amyloid load.

Highest protein quantity was measured in the two control subjects without any visible signs of ongoing disease process (amyloid load 0).

Lowest protein quantity was detected in AD subjects and three of the control subjects with amyloid load B. Graph C in figure 7 shows that all but one control subject were classified with Braak stage I and these have different

quantities of CKB. The AD subjects that were classified with Braak stage of V or VI had reduced protein quantity. Most AD subjects with Braak stage VI had a higher level of CKB than those in V.

Images in D, figure 7, are from three selected subjects (their ID number in left corner of the images) and shows how the protein quantity of CKB varies depending on amount of Aβ

present. Left images visualising CKB with DAPI and right images the presence of Aβ with DAPI. The images denoted C4 belong to a control subject without neurodegenerative signs (classified as stage I 0) and was the control subject with highest measured protein quantity of CKB of all subjects included in this project. As is seen in the image of CKB, the protein is localised to both cell body and end feet. In the right image is small amount of Aβ proteins seen close to the cell nucleus and it is considered as being intracellular. In control subject C5 (classified as stage I B) is the protein not distributed as extensive throughout the cell as it was for C4. The amount of Aβ in C5 is higher compared to C4 and the protein is seen as larger, extracellular aggregates of amyloid plaques, but

Figure 7. Mean intensity of CKB was compared between AD and control subjects (graph A). A correlation between CKB level and presence of amyloid plaques and NFTs was made (graph B and C). Images of subjects with different degree of amyloid load and Braak scores were analysed (D).

(17)

16 intracellular Aβ peptides are also present. In

AD subject A7 (classified as stage VI C) is almost no CKB visualised and the protein is only localised near cell nucleus. This subject has a substantial amount of aggregated extracellular amyloid plaques. No intracellular Aβ peptides are seen.

3.5 Western blot analysis

Presence of CKB in FCX was confirmed using western blot. The mean observed molecular weight of CKB for all subjects was 42.7 kDa, almost identical with the predicted molecular weight of 42.6 kDa (Appendix 8.7.4, table 16).

GAPDH was used as loading control (figure 8, image A).

The protein quantity of CKB using WB approach (Appendix 8.7.4, table 16 and 17) showed that control group in general had a higher protein quantity of CKB but with more variance among the subjects (52.48 ±33.86) than in the AD group (5.18 ±8.17), see figure 8, image B. Two control subjects (C2 and C9) displayed no visible bands on the membrane and these subjects had similar level of CKB as the AD subjects (figure 8, image B). All AD subjects had weakly or no visible bands on the membrane. A t-test (p<0.05) indicated that there was a significant difference between the groups (p=0.0020).

Possible correlations between protein quantity of CKB and presence of amyloid plaques and NFTs were made. In figure 8, image C is the measured CKB intensity

compared with level of amyloid load. The two control subjects with amyloid load 0 had different quantities of CKB. One subject was classified with amyloid load B but had almost the same level of CKB as the control subject with amyloid load 0. Control subject with amyloid load B had varied quantities of CKB and for the AD subjects was the general quantity low. The graph in figure 8, image D, shows that nearly all control subject were classified with Braak stage I with different quantities of CKB measured. AD subjects were classified as either Braak stage V or VI and had

reduced protein quantities compared with the control subjects. Most AD subjects with Braak stage VI had higher measured intensities of CKB than those with Braak stage V.

4. Discussion

In this section are the results from previous section analysed and discussed.

4.1 Data analysis and experimental approach

Cerebral cortex was selected as brain tissue to study since this region predominantly is affected in AD [10]. Since the aim was to find altered protein expression pattern associated with early disease process, the fluorescence histochemistry was performed on FCX tissue.

This region is, as discussed by Braak & Braak [10], affected by patological alterations later than other brain areas (eg temporal cortex and hippocampus) and hence assumed to better illustrate early disease events. Even though the transcriptomic data was based on

Figure 8. Western blot analysis was made to confirm presence of CKB in the samples and quantify the amount of protein.

(18)

17 TCX, it gave a guidance of genes expected to

be up or downregulated in FCX.

As the focus was to find druggable proteins that could be used as targets for available FDA approved drugs, this limited the number of proteins to analyse. However, discovering novel PET-ligands are associated with high costs and usually a low success rate. The PET- imaging methods that currently are used to diagnose AD are not disease specific, and a quest of better protein targets unique for AD exists. Finding druggable proteins expressed in astrocytes were therefore a short-cut to examine if available FDA approved drugs could be used as templates for novel PET-ligands and thus enable early detection of AD.

The immunofluorescence histochemistry was performed manually since gradient

differences in tissues were seen when Leica BOND-RX IHC/ISH Stainer was used. This issue was caused by the hydrophobic glass areas between the tissue cores which prevented a homogenous distribution of the fluids.

Performing immunofluorescence

histochemistry manually increased the risk of random errors but since criterion regarding both staining and tissue quality were made before the immunofluorescence analyses were performed, the error was reduced.

4.2 Immunofluorescence analysis

As was mentioned in the introduction, there are many subclasses of astrocytes in the brain and no cell marker stains all types of

astrocytes [36]. Three different astrocyte markers were therefore used in the immunofluorence histochemistry to identifying many of the sub-types. Co- existence analysis and cell counting revealed that CKB was the only druggable protein of those selected that was specifically associated with astrocytes, since it overlapped almost 100 % with all astrocyte markers and showed no overlap the other cell type markers used.

This finding indicated that astrocytes are a cell type in the brain with high CKB expression.

Using CKB as a novel cell type marker for

astrocytes could therefore be a complement to the already existing cell type markers. The CKB specificity of astrocytes also designates them as being the energy storage cells for phosphocreatine which underpin the important role of astrocytes in the brain.

When the fluorescence images of CKB were analysed an unexpected significant decrease in fluorescence intensity and an altered staining pattern were seen for the protein in all AD subjects. This observation was the opposite as the expected after studying the transcriptomic data of CKB, since it showed an upregulation of the gene in AD. This

assumption was based on the general understanding that one FPKM equals one mRNA molecule per average cell in a sample [37] and that transcript level and protein level are related [38]. Among the control subjects varied the intensity of CKB and three of these had a similar fluorescence staining profile as the AD subjects. The differences in

fluorescence intensity were initially

considered to be a staining issue, but repeated experiments gave similar result. The subjects medical record were analysed and a probable correlation between fluorescence intensity of CKB and presence of amyloid plaques was found. It is already reported that Aβ peptides contributing to cellular oxidative stress in the brain [39] which causes oxidation of proteins into stable, non-functional carbonylated proteins with aldehyde- and ketone sidechains [40, 41, 42]. This explanation clarifies the decreased intensity of CKB observed in the fluorescence images and demonstrates that transcriptomic and proteomic data not necessarily are related. One reason for the CKB gene being upregulated in AD might be that astrocytes are sensing a reduced energy metabolism as consequence of the

carbonylation, and thus increasing gene expression of CKB to maintain a functional energy level. If the decreased

immunofluorescence intensity was caused by inability of the antibody to target and bind

(19)

18 carbonylated CKB or was a result of protein

degradation is however not clear.

Two control subjects had no signs of amyloid plaques in the neuropathological examination (amyloid load 0), but presence of intracellular Aβ protein were noted in the fluorescence images. In the other control subjects were different amounts of intracellular and extracellular amyloid plaques visualized and these control subjects had decreased intensities of compared with the control subjects without extracellular amyloid plaques. This observation supports the claim made in [43], that accumulation of

intraneuronal Aβ proteins eventually results in cell lysis, causing the proteins to disperse in the extracellular space where they forming larger aggregates. Further, the decreased intensity of CKB that was observed among the control subjects as a probable consequence of the accumulated extracellular amyloid plaques justifies that pathological alterations occur before onset of symptoms. This finding highlights that a decreased level of CKB might be an early marker in the pathological

development of AD. Probably would the control subjects with amyloid load B developed AD if they had lived longer.

It was observed that CKB was more localised to cell nucleus the more extracellular amyloid plaques that were present. Why this pattern occurred can only be speculated. It might be that CKB is more abundant close to cell nucleus also in normal conditions due to the many energy demanding functions and the higher number of mitochondia, which might be the reason for why CKB is still detectable in this region. Based on co-existence analysis of CKB and astrocyte specific cell type markers it was noted that the number of astrocytes remained the same even if CKB was

decreased. This emphasizes that the reduced CKB not is the result of cell death but rather a consequence of alternated conditions within the cells.

No clear correlation was seen between level of CKB and Braak stage. It was however noted that some of the AD subjects with Braak stage VI had higher level CKB than AD subjects in stage V. If this pattern is a coincidence or is the result of increased gene expression of CKB is not clarifyed.

4.3 Association between CKB and AD

CKB was found to be an astrocyte-specific druggable protein with altered expression pattern in AD. This enzyme phosphorylates a majority (60-70%) of the freely available creatine in the brain into phosphocreatine, an energy storage form in the cells. CKB is, together with creatine and phosphocreatine an energy buffering and transport system in the cells that connects the generated energy produced in mitochondria with brain regions with high energy consumption, for instance the neurons. When increased energy demand occurs, CKB converts phosphocreatine back to creatine and phosphate where the phosphate group binds ADP to generate ATP [35].

Astrocytes are physically connected with neurons for the provision of essential metabolites, among them ATP. Decreasing levels of CKB will however result in an insufficient amount of ATP for the energy demanding functions that occur in neurons which disrupts their capacity to release as well as to remove neurotransmitters in the

synaptic clefts (excitotoxicity) [44, 45], causing damage and death of neurons. A decreased quantity of CKB in astrocytes might therefore contribute to the damage and neuronal loss that is seen in AD pathology.

4.4 PET ligand for CKB

The FDA approved drug that targets CKB and might be used as template for a novel PET- ligand is creatine. It shall however be

emphasized that creatine only will function to detect reductions in CKB and has no role as medical drug. One issue with creatine is its limited ability to diffuse over BBB [46], and this might be the reason for why the brain has an endogenous creatine synthesis.

(20)

19

4.5 Protein quantification

The decreased level of CKB that was observed in the fluorescence images was confirmed and quantified by measuring fluorescence mean intensity and performing western blot.

Although their general result was similar they used different approaches in measuring intensity. Protein quantity in the fluorescence images was measured as average pixel

intensity over whole tissue area. However, the intensity is affected by background intensity and tissue quality that has to be adjusted without influencing the result. Using WB as quantification method was therefore

considered more accurate since the measured pixel intensity was restricted to the bands on the membrane that exclusively should contain the specific protein.

The difference in how the intensity was measured between the methods became obvious when their t-tests (p<0.05) were compared. When the t-test was based on measured intensities from WB the difference between AD and control group was significant (p=0.002). However, when t-test was applied on the measured mean intensities from the fluorescence images there was no significant difference (p=0.056) between the groups.

Explanations for the varied result in p-value could be due to dynamic range and signal-to- noice ratio. The differences in the measured intensities could also have been influenced by which hemisphere that was used in the experiments. The right hemispheres were used in immunofluorescence histochemistry whereas the left for WB and the pathological alterations might not have progressed equally in both hemispheres.

WB was also used as a validation method that confirmed specificity of the antibody to target CKB, since only one band with a molecular weight similar to the predicted value was displayed on the membrane. The consistent band of control load GAPDH throughout all samples confirmed that all samples contained proteins and indicated that the reduced

staining of CKB in AD not was based on bias or sample load.

5. Conclusions

Only CKB was considered as an astrocyte druggable proteins of the proteins analysed with a significant different expression between AD and control subjects. Although transcriptomic data showed an upregulation of CKB in AD the immunofluorescence analysis showed a reduction due to protein

carbonylation. The observed decrease in protein level seem to be affected by ongoing disease processes and CKB might therefore be used as an early marker for AD. Creatine was found as PET-ligand template to target CKB with the potential to detect the protein reductions in a presymptomatic stage.

6. Acknowledgements

I would like to thank my external supervisor Jan Mulder who gave me the opportunity to work with this project and for his guidance and help throughout the project. A big thank to Nicholas Mitsios for his support and assistance in the lab. I also want to thank my main supervisor Cristina Al-Khalili Szigyarto for her supportive role in the beginning of the project. At last I will thank the whole Tissue Profiling group for their gentleness during my weeks at SciLife Laboratories.

7. Future perspectives

One astrocyte druggable protein was found in this project whose expression was altered in AD. Further investigation is needed on both gene level and protein level to better understand the functions and regulations of CKB that occur in AD as well as to determine what actually causes the decreased protein level seen in the presence of ongoing disease process. The decrease in CKB was observed both in immunofluorescence and western blot analysis. It is known that CKB undergoes

(21)

20 carbonylation in presence of oxidative stress [42], but it is currently unclear if this

carbonylation leads to reduced levels of CKB due to protein degradation or that the antibody used did not bind carbonylated CKB.

By examining the molecular structure of carbonylated CKB the reason for the decreased protein level can be revealed.

Based on the results and unexpected findings that were made in this project new questions arose that need to be investigated further.

The transcriptomic data indicated an upregulation of CKB in AD, but the immunofluorescence analysis showed a reduced level of CKB. It was also seen that the control subjects with signs of ongoing disease process had varied levels of CKB. What drives the cell to upregulate gene expression of CKB in AD and what causes the decreased protein level that was observed in the presence of pathological alterations are challenging questions that need to be analysed further.

Due to the limited amount of subjects used in this project, a larger study based on more subjects is required for a more reliable result.

More controls subjects without signs of any neurological disease is necessary to

investigate the association of decreased CKB level and presence of a disease processes. It must also be investigated if other neurological diseases display similar decrease in CKB as was

associated with AD. This project was limited to study protein expression pattern in astrocytes in FCX, but investigating other brain regions and other cell types might find other interesting druggable proteins to targetith PET-ligands.

The chemical properties of creatine as functional PET-ligand template must be investigated, since it must be permeable through the BBB efficiently to be useful.

According to [47], the permeability for

creatine over BBB is low but perhaps this issue can be solved with chemical modifications of the structure without disturbing the target binding to CKB. It must also be investigated if the ligand can be linked to either [11C] or [18F] [17] to enable detection in PET imaging.

Using cerebrospinal fluid (CSF) as detection method might be an option if difficulties with using creatine as PET ligand template occur, since it is known that carbonylated proteins are detectable in CSF [46]. A ratio between carbonylated CKB and normal CKB level could be one approach to set a threshold for individuals with risk of developing AD.

Analysing blood serum is another approach, but studies [47, 48] implies that detection of brain specific proteins in serum is limited due to BBB, and the protein level will therefore not correlate with the level found in CSF.

(22)

21

References

[1] Niedowicz, D. M., Nelson, P. T., & Murphy, M. P. Alzheimer’s disease: Pathological Mechanisms and Recent Insights. Current Neuropharmacology. 2011; 9(4): 674–684.

http://doi.org/10.2174/157015911798376181

[2] Van Cauwenberghe, C., Van Broeckhoven, C., & Sleegers, K. The genetic landscape of Alzheimer disease: clinical implications and perspectives. Genetics in Medicine. 2016; 18(5): 421–430.

http://doi.org/10.1038/gim.2015.117

[3] Alzheimer’s sjukdom och övrig demens [Internet]. Hjärnfonden; 2015. [Cited 22 February 2017].

Available from: http://www.hjarnfonden.se/om-hjarnan/diagnoser/alzheimers/

[4] Costantini, L. C., Barr, L. J., Vogel, J. L., & Henderson, S. T. Hypometabolism as a therapeutic target in Alzheimer’s disease. BMC Neuroscience. 2008; 9(Suppl 2): 1-9. http://doi.org/10.1186/1471-2202- 9-S2-S16

[5] Ben Haim, L., Carrillo-de Sauvage, M.A., Ceyzériat, K., & Escartin, C. (2015). Elusive roles for reactive astrocytes in neurodegenerative diseases. Frontiers in Cellular Neuroscience. 2015; 9, 278.

http://doi.org/10.3389/fncel.2015.00278

[6] Bloom, G.S. Amyloid-β and Tau - The Trigger and Bullet in Alzheimer Disease Pathogenesis. JAMA Neurology. 2014;71(4): 505-508. http://doi.org/10.1001/jamaneurol.2013.5847

[7] Crews, L., & Masliah, E. Molecular mechanisms of neurodegeneration in Alzheimer’s disease.

Human Molecular Genetics. 2010;19(R1): 12–20. http://doi.org/10.1093/hmg/ddq160

[8] Sadigh-Eteghad, S., Sabermarouf. B., Majdi, A., Talebi, M., Farhoudi, M., & Mahmoudi, J. Amyloid- Beta: A Crucial Factor in Alzheimer's Disease. Medical Principles and Practices. 2015; 24: 1-10.

[9] Kapitein, L.C., & Hoogenraad, C.C. (2015). Buildning the Neuronal Microtubule Cytoskeleton.

Neuron. 2015; 87(3): 492-506. http://doi.org/10.1016/j.neuron.2015.05.046

[10] Braak, H., & E. Braak, E. (1991). Neuropathological stageing of Alzheimer-related changes. Acta Neuropathologica. 1991; 82(4): 239-259.

[11] Kandel, E. R., Schwartz, J. H. 1., & Jessell, T. M. Principles of neural science. 5th ed.: McGraw-Hill;

2012. ISBN 0-07-139011-1.

[12] Alzheimer’s disease fact sheet [Internet]. National Institute of aging; 2012. [Cited 8 February 2017]. Available from:

https://www.nia.nih.gov/sites/default/files/alzheimers_disease_fact_sheet_1.pdf

[13] Hampel, H., O’Bryant, S. E., Castrillo, J. I., Ritchie, C., Rojkova, K., Broich, K, et al. PRECISION MEDICINE - The Golden Gate for Detection, Treatment and Prevention of Alzheimer’s Disease. The Journal of Prevention of Alzheimer’s Disease. 2016; 3(4): 243–259.

http://doi.org/10.14283/jpad.2016.112

[14] Blennow, K., Mattsson, N., Schöll, M., Hansson, O. & Zetterberg, H. Amyloid Markers in Alzheimer’s Disease. Trends in Pharmacological Science. 2015 ;36(5): 297-309.

http://doi.org/10.1016/j.tips.2015.03.002

(23)

22

[15] Forlenza, O.V., Diniz, B.S. & Gattaz, W.F. Diagnosis and biomarkers of predementia in

Alzheimer’s disease. Biomolecular Medicine. 2010; 8(89): 1-14. http://doi.org/10.1186/1741-7015-8- 89

[16] Schmidt, M.E. & Andrés, J.I. The value of PET ligand discovery to CNS drug development. Future Medicinal Chemistry. 2017; 9(4): 351-356. http://doi.org/10.4155/fmc-2017-0018

[17] Zhang, L. & Villalobos, A. Strategies to facilitate the discovery of novel CNS PET ligands. EJNMMI Radiopharmacy and Chemistry. 2016; 1:13. http://doi.org/10.1186/s41181-016-0016-2

[18] Vanltallie, T.B. Biomarkers, ketone bodies and the prevention of Alzheimer’s disease.

Metabolism. 2015; 64(3): 51-57. http://doi.org/10.1016/j.metabol.2014.10.033

[19] Ossenkoppele, Rik. Impact of molecular imaging on the diagnostic process in a memory clinic.

Alzheimer's & Dementia: The Journal of the Alzheimer's Association. 2013; 9(4): 414 – 421.

http://doi.org/10.1016/j.jalz.2012.07.003

[20] Marcus, C., Mena, E., & Subramaniam, R. M. Brain PET in the Diagnosis of Alzheimer’s Disease.

Clinical Nuclear Medicine. 2014; 39(10): 413–426. http://doi.org/10.1097/RLU.0000000000000547 [21] Ugbode, C., Hu, Y., Whalley, B., Peers, C., Rattray, M., & Dallas, M.L. Astrocytic transporters in Alzheimer's disease. Biochemical Journal. 2017; 474(3): 333-355.

http://doi.org/10.1042/BCJ20160505

[22] Oberheim, N. A., Goldman, S. A., & Nedergaard, M. Heterogeneity of Astrocytic Form and Function. Methods in Molecular Biology. 2012; 814:23–45. http://doi.org/10.1007/978-1-61779-452- 0_3

[23] Ben Haim, L., Carrillo-de Sauvage, M.-A., Ceyzériat, K., & Escartin, C. Elusive roles for reactive astrocytes in neurodegenerative diseases. Frontiers in Cellular Neuroscience. 2015; 9, 278.

[24] Heneka, Michael T. Neuroinflammation in Alzheimer’s disease. The Lancet Neurology. 2015;

4(14): 388 – 405. http://dx.doi.org/10.1016/S1474-4422(15)70016-5

[25] Morales, I., Guzmán-Martínez, L., Cerda-Troncoso, C., Farías, G. A., & Maccioni, R. B.

Neuroinflammation in the pathogenesis of Alzheimer’s disease. A rational framework for the search of novel therapeutic approaches. Frontiers in Cellular Neuroscience. 2014; 8, 112.

[26] Sofroniew, M. V., & Vinters, H. V. Astrocytes: biology and pathology. Acta Neuropathologica.

2010; 119(1): 7–35. http://doi.org/10.1007/s00401-009-0619-8

[27] Human Protein Atlas [Internet]. Introduction; 2017. [Cited 31 Mars 2017]. Available from:

http://www.proteinatlas.org/about

[28] Human Protein Atlas [Internet]. The druggable proteome, 2017. [Cited 26 January 2017].

Available from http://www.proteinatlas.org/humanproteome/druggable

[29] How to prepare your specimen for immunofluorescence microscopy [Internet]. Leica Microsystems; 2017. [Cited 25 January 2017]. Available from: http://www.leica-

microsystems.com/science-lab/how-to-prepare-your-specimen-for-immunofluorescence- microscopy/

(24)

23

[30] TSA Signal Amplification [Internet]. Perkin Elmer; 2017. [Cited 31 Mars 201u]. Available from https://www.perkinelmer.com/lab-solutions/resources/docs/BRO_TSAsignalamplification.pdf [31] Mahmood, T., & Yang, P.-C. Western Blot: Technique, Theory, and Trouble Shooting. North American Journal of Medical Sciences. 2012; 4(9): 429–434. http://doi.org/10.4103/1947- 2714.100998

[32] Signore, M., & Reeder, K.A. Antibody validation by Western blotting. Methods in Molecular Biology. 2012; 823:139-55. http://doi.org/10.1007/978-1-60327-216-2_10

[33] Bordeaux, J., Welsh, A. W., Agarwal, S., Killiam, E., Baquero, M. T., Hanna, J. A., … Rimm, D. L.

Antibody validation. BioTechniques. 2010; 48(3): 197–209. http://doi.org/10.2144/000113382 [34] Allen M, Carrasquillo MM, Funk C, et al. Human whole genome genotype and transcriptome data for Alzheimer’s and other neurodegenerative diseases. Scientific Data. 2016; 3:160089.

http://doi.org/10.1038/sdata.2016.89

[35] Creatine [Internet]. Drugbank; 2017. [Cited 17 May 2017]. Available from:

https://www.drugbank.ca/drugs/DB00148

[36] Sekar, S., McDonald, J., Cuyugan, L., Aldrich, J., Kurdoglu, A., Adkins, J., et al. Alzheimer’s disease is associated with altered expression of genes involved in immune response and mitochondrial processes in astrocytes. Neurobiology of Aging. 2015; 36(2): 583–591.

http://doi.org/10.1016/j.neurobiolaging.2014.09.027

[37] Hebenstreit D, Fang M, Gu M, Charoensawan V, van Oudenaarden A, et al. RNA sequencing reveals two major classes of gene expression levels in metazoan cells. Molecular systems biology.

2011; 7:497. http://doi.10.1038/msb.2011.28

[38] Schwanhausser B, Busse D, Li N, Dittmar G, Sch uchhardt J, Wolf J, et al. Global quantification of mammalian gene expression control. Nature. 2011; 473: 337–342.

http://www.doi.10.1038/nature10098

[39] Huang, W.J., Zhang, X., & Chen, W.W. Role of oxidative stress in Alzheimer’s disease. Biomedical Reports. 2016; 4(5): 519–522. http://doi.org/10.3892/br.2016.630

[40] Aksenov, M., Aksenova, M., Butter, A. & Markesbery, W.R. Oxidative modifcation of creatine kinase BB in Alzheimer's disease brain. Journal of Neurochemistry. 2000; 74(6): 2520-2527 [41] Dalle-Donne, I., Rossi, R., Giustianni, D., Milzani, A. & Colombo, R. Protein carbonyl groups as biomarker of oxidative stress. Clinical Chimica Acta. 2003; 329(1-2): 23-38.

[42] Irani, D. N. Cerebrospinal fluid protein carbonylation identifies oxidative damage in autoimmune demyelination. Annals of Clinical and Translational Neurology. 2017; 4(2): 145–150.

http://doi.org/10.1002/acn3.379

[43] Laferla, F., Green, K.N., & Oddo, S. Intracellular amyloid-β in Alzheimer's disease. Nature. 2007;

8(7): 499-509. http://doi.org/10.1038/nrn2168

[44] Butterfield, D. A., Swomley, A. M., & Sultana, R. Amyloid β-Peptide (1–42)-Induced Oxidative Stress in Alzheimer Disease: Importance in Disease Pathogenesis and Progression. Antioxidants &

Redox Signaling. 2013; 19(8): 823–835. http://doi.org/10.1089/ars.2012.5027

[45] De Lores Arnaiz GR, Ordieres MGL. Brain Na⁺, K⁺-ATPase Activity In Aging and Disease.

International Journal of Biomedical Science : IJBS. 2014;10(2):85-102.

(25)

24

[46] Beard, E. & Braissant, O. Synthesis and transport of creatine in the CNS: importance for cerebral functions. Journal of Neurochemistry. 2010; 115(2):297-313. http://doi.org/10.1111/j.1471-

4159.2010.06935.x

[47] Ågren, H. & Niklasson, F. Creatine and creatinine in CSF: indices of brain energy metabolism in depression. Journal of Neural Transmission. 1988; 74(1):55-59.

[48] Swahn, C.G. & Sedvall, G. CSF creatine in schizophrenia. Biological Psychiatry. 1988; 23(6): 586- 594.

(26)

25

8. Appendix

8.1 Information of disease stages

8.1.1 Disease stage definition

Table 3. Braak stages I-VI describes the distribution of NFTs in different brain regions in AD.

Braak stage Description

I-II

Transentorhinal stages.

Neurons in the transentorhinal Pre-α develop NFTs are affected.

Mild involvement of hippocampus.

Absence of isocortical changes.

III-IV Limbic stages.

Both transentorhinal Pre-α and entorhinal layer are severly affected by NFTs.

Moderate hippocampal involvement.

Low or absence of changes in isocortex.

V-VI

Isocortical stages.

Severe changes and considerable loss of neurons in transentorhinal Pre-α.

All components of hippocampal formation are involved.

Isocrotex is severly affected.

Table 4. Amyloid load describes the distribution of extracellular amyloid plaques in different brain regions in AD.

Amyloid load Description

0 No plaque deposition.

A Low densities of amyloid deposits in basal regions of isocortex (particularly in frontal, temporal and occipital lobe).

Absence of amyloid deposits in hippocampus.

B Medium densities of amyloid deposits present in all regions of isocortex.

Mild involvement of amyloid deposits in hippocampus.

C Densely packed deposits of amyloids in all isocortical regions.

Hippocampus have same pattern of deposits as in stage B.

Gradual involvement of subcortical structures.

8.1.2 Subject information

Table 5. Age, gender, post-mortem interval, disease stage and neuropathological information from autopsy of the AD and control subjects.

Case Gender

(F/M)

Age PMI (h) Disease stage

Neuropathological information A1

2006-060

F 83 04:55 VI C Moderate number senile plaques and NFTs.

Many plaques, part of them “classic”.

A2 2001-044

M 85 04:25 V C Substantial number of senile plaques and NFTs.

Many plaques, largely “diffuse” part “classic”.

A3 2010-051

M 74 07:40 V C Many senile plaques and “classic” plaques.

Numerous NFTs.

A4 2012-060

F 80 04:00 VI C Many senile plaques.

Large number “diffuse” plaques and moderate number “classic” plaques.

A5 2013-021

F 96 05:05 VI C Few dispensed senile plaques.

Few plaques, moderate number of “diffuse” plaques. Many NFTs.

A6 2007-068

F 70 05:20 VI C Moderate number senile plaques.

Many plaques, most “diffuse” but also many “classic”. Many NFTs.

A7 2012-068

M 88 05:30 VI C Many senile plaques.

Moderate number NFTs.

A8 2012-125

M 71 06:35 VI C Single senile plaque.

Many diffuse plaques, few classical plaques. Large number NFTs.

A9 2008-047

M 77 06:35 VI C Many senile plaques.

Many plaques, many “classic”. Many NFTs A10

2002-056

F 85 03:45 V C Many amyloid plaques.

Many plaques, few “classic”. Few NFTs.

C1 2009-022

F 77 02:55 I B Moderate number of senile plaques, many “classical” shape.

No NFTs.

C2 2000-137

F 92 07:15 I B Few senile plaques.

Substantial number of “classical” plaques. No NFTs.

C3 2011-028

F 81 04:25 I 0 One single senile plaque.

No NFTs.

C4 2013-016

M 83 05:15 I 0 Only a single “classic” plaque. Some “diffuse” plaques.

No NFTs.

C5 M 88 07:25 I C Many senile plaques.

Astrocyte-specific druggable protein as PET-ligand target for early detection of Alzheimer’s disease

Astrocyte-specific druggable protein as PET-ligand target for early detection of Alzheimer’s disease

REBECCA SJÖBERG

Astrocyte-specific druggable protein as PET-ligand target for early detection of Alzheimer’s disease

Rebecca Sjöberg

Abstract

Abstrakt

Table of contents

1. Introduction

1.1 Alzheimer’s disease

1.2 A quest for better PET-ligands

1.3 Astrocytes as PET-ligand targets

1.4 The druggable proteome

1.5 Immunofluorescence visualizes proteins in tissue

1.6 Aim

2. Materials and methods

2.1 Data analysis

2.2 Human brain tissue samples

2.3 Immunoflourescence analysis

2.4 Image acquisition and analysis

2.5 Western blot analysis

2.6 Databases and search strategy

2.7 Ethical considerations

3. Result

3.1 Data analysis

3.2 Immunofluorescence analysis

3.3 Top astrocyte-specific druggable protein

3.4 Protein quantification

3.5 Western blot analysis

4. Discussion

4.1 Data analysis and experimental approach

4.2 Immunofluorescence analysis

4.3 Association between CKB and AD

4.4 PET ligand for CKB

4.5 Protein quantification

5. Conclusions

6. Acknowledgements

7. Future perspectives

References

8. Appendix

8.1 Information of disease stages