Binding of Human Proteins to Amyloid-β Protofibrils - ACS Chemical

Romain Salza , Claire Lethias , Sylvie Ricard-Blum ... Niels R. Reinders , Yvonne Pao , Maria C. Renner , Carla M. da Silva-Matos , Tessa R. Lodder , ...
0 downloads 0 Views 4MB Size
Subscriber access provided by - Access paid by the | UCSF Library

Article

Binding of human proteins to amyloid-# protofibrils M. Mahafuzur Rahman, Henrik Zetterberg, Christofer Lendel, and Torleif Härd ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/cb5008663 • Publication Date (Web): 03 Dec 2014 Downloaded from http://pubs.acs.org on December 12, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Chemical Biology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

Binding of human proteins to amyloid-β protofibrils

M. Mahafuzur Rahman1, Henrik Zetterberg2,3, Christofer Lendel1,4,5, Torleif Härd1 1

Department of Chemistry and Biotechnology, Swedish University of Agricultural Sciences (SLU), Box 7015, SE-750 07 Uppsala, Sweden. 2

Department of Psychiatry and Neurochemistry, University of Gothenburg, SE-413 45 Göteborg, Sweden. 3

4

UCL Institute of Neurology, Queen Square, London WC1N 3BG, United Kingdom.

Present address: Department of Chemistry, Royal Institute of Technology (KTH), SE-100 44 Stockholm, Sweden.

5

To whom correspondence should be addressed: Christofer Lendel, Department of Chemistry and

Biotechnology, Swedish University of Agricultural Sciences (SLU), Box 7017, SE-750 07 Uppsala, Sweden Tel.: +46-(0)18-673183; Email: [email protected] or [email protected];

Abbreviations used are: Aβ, amyloid-β; AD, Alzheimer’s disease; ANS, 1-anilinonaphtalene 8sulfonic acid; ApoE, apolipoprotein E; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; CSF, cerebrospinal fluid; LC, liquid chromatography; MS, mass spectrometry; PAI, peptide abundance index; SPR, surface plasmon resonance

ACS Paragon Plus Environment

1

ACS Chemical Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

The progressive neurodegeneration in Alzheimer’s disease is believed to be linked to the presence of prefibrillar aggregates of the amyloid-β (Aβ) peptide in the brain. The exact role of these aggregates in the disease pathology is, however, still an open question. Any mechanism by which oligomeric Aβ may cause damage to neuronal cells must, in one way or another, involve interactions with other molecules. Here, we identify proteins in human serum and cerebrospinal fluid that bind to stable protofibrils formed by an engineered variant of Aβ42 (Aβ42CC). We find that the protofibrils attract a substantial number of protein binding partners. Many of the 101 identified proteins are involved in lipid transport and metabolism, the complement system or in hemostasis. Binding of representative proteins from all these groups with micromolar affinity was confirmed using surface plasmon resonance. In addition, binding of apolipoprotein E to the protofibrils with nanomolar affinity was demonstrated. We also find that aggregation of Aβ enhances protein binding, as lower amounts of proteins bind monomeric Aβ. Proteins that bind to Aβ protofibrils might contribute to biological effects that these aggregates are involved in. Our results therefore suggest that improved understanding of the mechanisms by which Aβ causes cytotoxicity and neurodegeneration might be gained from studies carried out in biologically relevant matrices in which Aβ β-binding proteins are present.

ACS Paragon Plus Environment

2

Page 3 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

Alzheimer’s disease (AD) is the most common form of dementia. The pathological hallmarks of AD are the assembly of amyloid-β (Aβ) into amyloid fibrils in extracellular plaques and the intracellular aggregation of hyperphosphorylated protein tau. The details of these processes and how they are related to neurodegeneration are, however, still only vaguely understood. The amyloidcascade hypothesis1 states that AD is initiated by an imbalance in the production and clearance of Aβ that eventually leads to the accumulation and deposition of Aβ aggregates. Furthermore, several lines of evidence support the role of soluble prefibrillar Aβ aggregates, so-called oligomers or protofibrils, as the causative agents of the disease2. Investigations of such species are difficult because of their transient nature and the understanding of their properties and disease-associated mechanisms is therefore limited. However, any mechanism or pathway leading to neurodegeneration must involve interactions between the protein aggregates and other biomolecules. For example, a frequently discussed hypothesis is that interaction with, and disruption of, cell membranes could lead to neurodegeneration. Several different proteins have been suggested to play important roles in AD pathology via direct or indirect interactions with Aβ. From a general perspective, it was recently demonstrated that many essential proteins can co-aggregate with amyloidogenic peptides and thereby give rise to cytotoxicity via loss-of-function mechanisms3. It is also well known that senile plaque include a range of different proteins in addition to Aβ4,5. The knowledge about which proteins that interacts with soluble prefibrillar aggregates is, however, limited. Such interactions could provide direct links to the processes involved in neurodegeneration (Figure 1). Furthermore, decoration of Aβ aggregates with proteins might significantly alter their processing in vivo, in line with what has been shown for artificial nano-particles6,7. Hence, characterization of the interactome of soluble neurotoxic Aβ aggregates may be important for the understanding of the AD pathology. Biochemical studies of prefibrillar protein aggregates are challenging because of heterogeneous sample compositions and the transient nature of oligomeric species. We have developed an engineered Aβ variant (AβCC) in which two alanines (in positions 21 and 30) are replaced by cysteines8. The resulting intramolecular disulfide bridge prohibits the peptide to form amyloid fibrils and AβCC instead forms stable protofibrils with properties that are indistinguishable to those reported for wild

ACS Paragon Plus Environment

3

ACS Chemical Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 29

type Aβ. In atomic force microscopy, the protofibrils appear as rod-like particles that are ca 3 nm in diameter and 60-200 nm in length with a distinct morphology compared to amyloid fibrils9 and a recent model based on solid state NMR spectroscopy suggest that the protofibrils are built from hexameric barrel-like units10. We have also shown that they are rich in β-sheet secondary structure8, bind the dye 1-anilinonaphtalene 8-sulfonic acid (ANS) and that they are recognized by antibodies that are conformation-specific for wild type Aβ aggregates9. Finally, the Aβ42CC protofibrils are cytotoxic in cell culture8 and they attenuate synaptic activity in mouse hippocampal neurons9. In the present study we use Aβ42CC protofibrils to investigate which proteins in human serum and cerebrospinal fluid (CSF) that interact with prefibrillar, soluble Aβ aggregates. RESULTS AND DISCUSSION Binding of serum proteins. A pull-down approach was used to identify proteins that bind to Aβ protofibrils. Aβ42CC protofibrils were immobilized on tosyl-activated magnetic beads and incubated with human serum. After repeated washing steps, the samples were analyzed using SDS-PAGE. Several gel bands were observed for the protofibril samples while the control (beads coated with glycine) only displayed a few weak bands (Figure 2a). We also observed gel bands corresponding to monomer, dimer and trimer of Aβ42CC, as expected for protofibrils that dissociate in SDS8. The strong band at 34 kDa was previously identified as apolipoprotein E (ApoE)9. The results demonstrate that the methodology works as expected and that Aβ42CC protofibrils associate with several human proteins. To identify the binding proteins, whole gel lanes, except the regions containing Aβ42CC, were extracted and investigated by mass spectrometry (MS). After proteolytic digestion, the peptide mixtures were analyzed using liquid chromatography-MS/MS and the originating proteins were identified using the MASCOT search engine (www.matrixscience.com). We found that some proteins bind to both protofibrils and control beads, although the SDS-PAGE shows that the levels were much lower for the control (Figure 2a). The higher level binding to the protofibrils was also obvious when the peptide abundance indices (PAI) of proteins that were found in both Aβ and the control experiments were compared (Figure 2b) (we define PAI as the number of identified peptides of a

ACS Paragon Plus Environment

4

Page 5 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

protein divided by its molecular weight11). Although the difference in surface area accessible for binding between the protofibril-coated beads and the control might affect the outcome of the experiments it is clear that background binding to the beads does not compromise the results. The protein with the overall highest PAI was ApoE in the Aβ sample, which is in agreement with the strong band in figure 2a. The list of binding proteins also contains proteins involved in the complement system and hemostasis, as well as apolipoproteins and immunoglobulins (Figure 3 and supporting information S1 and S2). However, the set of binding proteins is very specific in comparison with what is in general expected from proteomics analysis of serum samples, and the same is true for the CSF samples (vide infra). Binding of CSF proteins. Serum is a high protein concentration environment. The protein concentration in CSF is approximately 100 times lower. However, CSF is perhaps more relevant for AD pathology as it communicates freely with the brain interstitial fluid and represents the environment where senile plaque is formed12. SDS-PAGE analysis of protein binding in CSF showed that more proteins were extracted by the protofibrils than by the glycine-coated control beads, similar to the serum results (Figure 4a). In addition, control experiments with tryptophan-coated beads, that should present a more hydrophobic surface than the glycine beads, were carried out on the CSF samples. SDS-PAGE analysis of the experiments with these beads showed the same low level binding as the glycine controls (Figure 4b). We analyzed twelve CSF samples (six samples from AD patients and six samples from non-AD individuals). The SDS-PAGE analysis did not reveal major differences between the two classes of samples although different levels of binding proteins can be observed (Figure 4a). All twelve CSF samples were analyzed by MS for binding to Aβ42CC protofibrils and two CSF samples (one AD and one non-AD) were analyzed for binding to glycine and tryptophan controls. The identified proteins are summarized in figure 3 and supporting information S1 and S2. Notably, only two proteins were found in the control samples (ApoE and serum albumin), which supports that the identified proteins in the Aβ samples do interact with the protofibrils. The average PAI values for these two proteins were substantially higher for the protofibril samples (1.06±0.14 and 0.30±0.11 for ApoE for serum albumin, respectively) than for the controls (0.22 and 0.092 for ApoE and serum albumin, respectively, in the glycine experiments and 0.069 and 0.022 for ApoE and serum albumin,

ACS Paragon Plus Environment

5

ACS Chemical Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

respectively, in the tryptophan experiments). Hence, identification of these proteins in the protofibril samples is not only due to background binding to the beads. The list of proteins from CSF shares some similarities with the serum results (Figure 3 and supporting information S1) but there are also examples of proteins that were not identified in serum, for instance prostaglandin-H2 D-isomerase and mimecan. The majority of the frequently occurring proteins (identified in at least six CSF samples) have previously been associated with Aβ deposits in vivo (e.g. Refs. 4,5). Moreover, we observed substantial variations between the CSF samples with certain proteins found in only one single sample while others were detected in all twelve samples (Figure 3 and supporting information S1). Such variations are expected as the total protein concentrations as well as the relative levels of each protein should differ between the investigated samples. The total protein concentrations in the CSF samples were estimated using the Bradford assay but no correlation between the total protein level and the number of identified proteins was found (data not shown). Comparisons between CSF samples from AD and non-AD subjects did not show any significant differences. There were, however, indications that some interacting proteins were overrepresented in one of the sample groups, e.g. apolipoprotein A-IV and prostaglandin-H2 Disomerase. An extended study would be required to confirm these differences. Binding of proteins to monomeric Aβ. As a comparison, we investigated the binding of CSF proteins to monomeric Aβ. This was a more challenging task, as the monomeric state is not as stable as the Aβ42CC protofibrils and easily aggregates. We could not use any variant of Aβ42 as it would aggregate more rapidly than the ligation to the beads. Instead we used wild type Aβ40 and performed the ligation at low temperature (4 °C). We carried out the binding experiments using two sets of beads. One prepared with the same peptide concentration as previously used for Aβ42CC protofibrils and the second with a 1:10 dilution, to reduce peptide aggregation during the ligation. SDS-PAGE analysis of the binding experiments show that in both cases, substantially more protein binds to the Aβ42CC protofibrils than to Aβ40 monomer (Figure 5). Hence, we conclude that aggregation of Aβ significantly enhances binding of proteins in human body fluids. However, a few weak gel bands were also visible in the Aβ40 samples

ACS Paragon Plus Environment

6

Page 7 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

suggesting that some of the proteins also recognize monomeric Aβ. These proteins were identified by MS and only 9 different proteins were found compared to in average 30 proteins for the protofibrils (Figure 3 and supporting information S1 and S2). A previous study identified serum albumin, clusterin, fibulin, serum amyloid P component, vitronectin, and apolipoproteins AI, AIV and E as protein ligands of monomeric Aβ40 in human serum13. The differences compared to our results are most likely explained by the different protein compositions of serum and CSF. Binding of proteins to a non disease-related protein nanofiber. To further assess the specificity of the identified interactions we investigated the binding to protein nanofibers that are not associated with human disease. We immobilized nanofibers of the yeast prion protein Sup35 on magnetic beads, which were then incubated with two CSF samples. SDS-PAGE analysis shows that several proteins indeed bind the nanofibers. However, the pattern of gel bands is different compared to Aβ protofibrils (Figure 4b). MS analysis identified fewer binding proteins (18 and 20, respectively) than for the protofibrils (30 on average). The majority of the proteins were the same as those that bind to protofibrils but there were also a number of different proteins, indicating distinct interaction patterns of the two types of protein nanostructures. Interestingly, the Sup35 fibrils seem to bind a number of amyloid related proteins, including the prion protein and Aβ itself, which is not observed for the protofibrils. Confirmation of the interactions using surface plasmon resonance. To confirm that the identified binding partners interact with Aβ protofibrils we immobilized Aβ42CC protofibrils on a carboxymethylated dextran surface and investigated the binding of selected proteins using surface plasmon resonance (SPR). We selected four proteins from different functional categories for these tests. ApoE was selected as it is a recognized genetic risk factor. ApoE also serves as a positive control since it was identified as a binder in all experiments (with or without Aβ). Indeed, the SPR experiments confirmed strong binding of ApoE to the protofibrils (Figure 6d). We also investigated the binding of apolipoprotein A-I, complement C3 and antithrombin III from the lipid transport/metabolism, complement/inflammation and hemostasis groups, respectively (Figure 6a-c). All three proteins were found to bind to the immobilized Aβ protofibrils. The values of

ACS Paragon Plus Environment

7

ACS Chemical Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 29

the dissociation constants were estimated using the procedure described in the methods section. All proteins except ApoE showed binding with micromolar affinity (KD = 3 µM, 0.6 µM and 0.3 µM for apolipoprotein A-I, complement C3 and antithrombin III, respectively). ApoE had a higher affinity with KD = 3 nM. Aβ protofibrils are decorated by human proteins. Our data suggest that Aβ protofibrils have the ability to interact with a range of different human proteins. The list includes several proteins that are associated with Aβ in senile plaques, which indicates that our experiments are of relevance for the in vivo situation. A direct consequence of these findings is that the biological and neurotoxic effects of Aβ aggregates might not be fully understood from experimental models that lack these binding partners. As a comparison, it is established that synthetic nanoparticles bind a number of proteins in human plasma and that the combined nanoparticle-protein entity constitutes the biologically active unit7. We hypothesize that this could be true also for nanosize protein aggregates. Indeed, some of the proteins identified in our experiments, including apolipoproteins, albumin and fibrinogen, are also frequently adsorbed to synthetic nanoparticles6. These similarities, together with the notion that it is unlikely that all the identified proteins could have specific epitopes of recognition on the Aβ42CC protofibrils, suggest that protein binding to Aβ aggregates mainly relies on non-specific interactions. Several studies have shown that the cytotoxic effects of Aβ are modulated by other proteins, e.g. Refs 14–18. Hence, further investigations of the protein binding patterns might lead to new insights about the pathways of neurodegeneration that are triggered by protein aggregates. One can think of several hypothetical mechanisms of toxicity that involve the binding of human proteins to protein aggregates: 1) the sequestered proteins might be prohibited from carrying out their functions resulting in loss-of-function effects; 2) the colocalization of different proteins to the same aggregate might catalyze chemical reactions or initiate signalling processes that eventually lead to cell death; 3) binding of specific proteins might target the protofibrils for cellular internalization where they may interfere with intracellular components; and 4) the protofibrils can, upon internalization, bring extracellular proteins with them which then end up in a cellular compartment where they should not be. In addition, protein binding might disguise the protofibrils and thereby interfere with therapeutic approaches such as immunotherapy.

ACS Paragon Plus Environment

8

Page 9 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

The identified binding partners could also reflect protective mechanisms that are active against protein aggregates. Such chaperone functions are associated with two major genetic risk factors associated with AD, i.e. ApoE and clusterin, which both can modulate Aβ aggregation18,19 and promote Aβ clearance20 and transport over the blood-brain barrier21. In addition, clusterin is involved in several additional AD-associated processes such as cholesterol metabolism and transport (that also involve ApoE)22, apoptosis23 and complement activation24. Clusterin and ApoE, were found to bind Aβ protofibrils in all samples of this study. Pathways of neurodegeneration. Among the identified binding partners, a few main groups of proteins can be distinguished: 1) complement-related inflammatory response proteins; 2) proteins involved in hemostasis, 3) proteins related to lipid transport and metabolism and 4) proteins involved in metabolism. At least the first three groups are involved in biological processes that have been implicated in AD. Aβ has been shown to activate the complement pathway in the absence of immunoglobulin binding25, a process that appears to be common for several neurodegenerative protein aggregation disorders26. Such activation is unlikely to originate from the native, monomeric form of Aβ, since it is always present. More likely, aggregated Aβ is the origin of this process and it is known that complement factors colocalize with senile plaque25,27. Our results show that Aβ protofibrils can bind complement C1q as well as C3 and thereby potentially activate the complement pathway. Unlike the amyloid in the senile plaque, the protofibrils are mobile and could propagate a state of chronic inflammation to new areas in the brain. Vascular abnormalities are hallmarks of AD pathology. Aggregated Aβ has been shown to bind to and activate factor XII, which represents an initial step in the blood coagulation process28. Although we did not observe binding to factor XII, several other proteins related to the coagulation process were identified. Among those is factor V that has a central role in the hemostasis pathway and of which deficiency leads to increased risk of hemorrhage29. Hence, a loss of function effect of factor V could be related to microhemorrages that are part of the AD pathology. Furthermore, fibrinogen appears as binding partner in several samples. Aβ has been reported to bind strongly to fibrinogen and stabilize

ACS Paragon Plus Environment

9

ACS Chemical Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 29

fibrin networks30, which might lead to decreased cerebral blood flow and eventually neurodegeneration. There is also a close connection between hemostasis and inflammation, which might contribute to the AD pathology31. Lipids are intimately linked to AD and Aβ. The identified apolipoproteins, including clusterin, all seem to be associated with high-density lipoparticles and cholesterol. The cholesterol metabolism is a process that is frequently discussed in relation to AD pathology and the serum cholesterol level is associated with an increased risk to develop AD32. Cholesterol is, together with sphingolipids, essential for the neuronal membrane architecture and the formation of lipid rafts. Such variations in the composition of the lipid environment regulate Aβ precursor protein processing, Aβ production and Aβ aggregation22. Therefore, interference with the lipid metabolism and transport by protofibrils could cause local disturbances in Aβ production as well as propagation of Aβ accumulation. Among the metabolic enzymes, we identified glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in all samples. Besides its role in glycolysis, this protein has been assigned a plethora of different functions33, e.g. regulation of apoptosis and membrane fusion. Hence, the observed interaction with Aβ protofibrils provides a linkage to these processes that are central in AD pathology. Indeed, GAPDH has been suggested to play a key role in neurodegeneration34. It has been shown to bind monomeric35 and fibrillar Aβ36 and Aβ has been suggested to promote GAPDH aggregation and nuclear accumulation in neurons37. Moreover, GAPDH has been identified as a component of neurofibrillar tangles38 and it thereby connects oligomeric Aβ and tau pathology. In this context, the ability of GAPDH to translocate between different cellular compartments33 might be an important piece of the AD puzzle. Interestingly, the protein mimecan, found in all twelve CSF samples, has not been linked to AD previously (according to PubMed, October 2014). However, it belongs to the proteoglycan family, which is strongly associated with AD. For example, heparan sulfate proteoglycans are found in senile plaque39 and they have been shown to promote the assembly of amyloidogenic proteins and to have active roles in the internalization and toxicity of protein aggregates40. Potential implications for AD therapy. The difficulties encountered in direct targeting of protein aggregation41 call for alternative approaches in drug discovery for neurodegenerative

ACS Paragon Plus Environment

10

Page 11 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

disorders. Several studies have shown that Aβ toxicity can be modulated by other proteins14-18. Hence, manipulating the binding of human proteins to oligomeric Aβ might turn out to be a promising therapeutic avenue. We here present a list of proteins that interact with one, well-defined, prefibrillar Aβ aggregate. Further studies of other classes of aggregates are necessary to obtain a complete picture of the biological processes. Moreover, the binding pattern in CSF might be different compared to brain parenchyma although CSF communicates freely with the brain interstitial fluid which is the matrix in which Aβ deposits12 and CSF Aβ42 levels have been shown to reflect brain amyloid load42. Still, based on the present results we hypothesize that selected interactions could be directly targeted by inhibitors of protein-protein binding. The identified biological pathways also open for targeting up- or downstream components. METHODS Materials. Human serum (single male donor, 51 years old) was purchased from 3H Biomedical and CSF samples were obtained from the Clinical Neurochemical Laboratory (Sahlgrenska University Hospital, Gothenburg, Sweden). CSF samples were from patients who sought medical advice because of cognitive impairment. Patients were designated as normal or AD according to CSF biomarker levels using cut-offs that are >90% specific for AD43: total tau >350 ng l-1, phosphor-tau >80 ng l-1 and Aβ42