The Role of the unfolded protein response (UPR) in tau
phosphorylation and aggregation
Michael Whitehead
Supervisors
Dr Anna Nolle and Dr Wiep Scheper
Acknowledgements
I would like to thank everyone in the neurogenetics lab for being extremely friendly and helpful throughout my project. My supervisors Dr Anna Nolle and Dr Wiep Scheper have worked extremely hard to both teach me the techniques which I have used as well as help and guide me throughout the process, which has been invaluable. Rob Zwart, Judith van der Harg and Hyung Elfrink also advised me throughout my time in the lab, which I greatly appreciated. I would also like to thank Dr Huib Ovaa, at the Netherlands Cancer Institute, for the proteasome activators and Professor Tiago Outeiro, Director of the Department of Neurodegeneration and Restorative Research, University of Gottingen, for the BiFC assay that we used in this project.
Abstract ... 4
Abbreviations ... 5
1 Introduction ... 6
1.1 Tauopathies ... 6
1.2 Microtubule binding protein ... 7
1.3 Tau phosphorylation ... 9
1.4 Tau oligomers not aggregates are toxic ... 10
1.5 Activation of the unfolded protein response is an early event in tauopathies ... 11
1.8 Removing tau: Macro-autophagy ... 12
1.7 Removing tau: ... 13
Aim ... 15
2 Materials and methods ... 16
2.1 Cell culture ... 16
2.3 Transformation ... 16
2.4 Cell transfection ... 17
2.5 Western blot ... 17
2.6 Flurorescent microscopy ... 18
2.7 FACS ... 18
2.8 Statistical analysis ... 18
3 Results ... 19
3.1 Optimisation of transfection protocol for BiFC assay ... 19
3.2 Validation of BiFC assay using Okadaic acid ... 19
3.3 UPR activation using tunicamycin does not result in tau oligomerisation ... 20
3.6 Activation of the UPR using 2-Deoxygucose increased tau oligomerisation ... 21
3.8 UPR activators and positive controls ... 22
4 Discussion ... 32
4.1 Future research ... 36
5 References ... 32
6 Supplementary data ... 43
Abstract
Tau is a microtubule associated protein which has been implicated in a number of diseases termed tauopathies, the most common being Alzheimer’s disease (AD). One of the main pathological hallmarks of tauopathies is tau aggregates which contain hyperphosphorylated, insoluble tau. Using brain material from AD and other tauopathies (without Aβ pathology) our group has shown that activation of the unfolded protein response (UPR) precedes tau aggregation. Cell models indicate that UPR activation leads to phosphorylation of tau. The UPR is primarily a protective mechanism however we hypothesise that chronic activation contributes to tau pathology. Interestingly changes in glucose metabolism, which is a high risk factor for AD, leads to UPR activation. This led to the question; does UPR activation lead to tau aggregation? In this study we used a bimolecular fluorescent complementation assay (BiFC) to visualise tau oligomers in a cell model. This showed that activation of the unfolded protein response (UPR) using 2-deoxyglucose, induces tau oligomerisation. In another line of study we looked at the role of the proteolytic system. Inducing lysosome dysfunction using bafilomycin, in this cell model leads to an increase in tau oligomers, supporting it as a therapeutic target. We also screened 14 proteasome activators for their ability to reduce tau. We identified one potential candidate. In conclusion we have used a cell model for visualising tau oligomerisation, which is intrinsically difficult to achieve and we have shown both UPR activation and lysosome dysfunction significantly increase tau dimers/oligomers. We have also identified a proteasome activator which has therapeutic potential and can be tested in this system.
Key words
Abbreviations
2-DG – 2-deoxyglucose AD – Alzheimer’s disease
ATF6 - activating transcription factor 6 Baf – bafilomycin
BiFC – Bimolecular fluorescent complementation assay DAPI - 4',6-diamidino-2-phenylindole
ER – endoplasmic reticulum
ERAD –endoplasmic reticulum associated degradation FACS – fluorescent activated cell sorting
GF – green fluorescence
GSK-3β – glycogen synthase kinase 3β IRE1 - inositol requiring enzyme 1 MT - microtubules
OA – okadaic acid
O-GlucNAc – O-linked acetyl-glucosamine PBS – phosphate buffered saline
PBST – phosphate buffered saline with tween PERK - pancreatic ER kinase
PM – plasma membrane Sal – salubrinal
SDS-PAGE - sodium dodecyl sulfate polyacrylamide gel electrophoresis
Tg – thapsigarginTm – tunicamycin
UPR – unfolded protein response wat – wortmannin
1 Introduction
1 in 8 people in the US have alzheimer’s disease (AD), making it the most common tauopathy. AD is the sixth leading cause of death in the world and is predicted to rise as healthcare and average life expectancy improves. Current therapeutics are ineffective and clinical trials for new treatments have had only limited success (Karakaya, Fusser et al. 2012). Consequently there is a need to understand the early mechanisms which cause tauopathies, to identify more effective therapeutics. We have found that activation of the unfolded protein response (UPR) is an early mechanism which is strongly associated with tau phosphorylation. Because metabolic stress activates the UPR and is considered a strong epidemiological risk factor (Accardi, Caruso et al. 2012) we hypothesise that the UPR is important process in the pathogenesis of tauopathies. This suggests that the UPR represents a potential therapeutic target however there are a number of unanswered questions as to how it contributes to these diseases.
1.1 Tauopathies
The formation of hyperphosphorylated, insoluble tau aggregates into neuronal inclusions is the characteristic hallmark of all tauopathies. Alzheimer’s disease (AD) is the most common and is associated with both Aβ plaques and NFT (Figure 1B). There are a number of associated genetic risk factors (Figure 1a) however the highest risk factor is aging. Many different brain regions are affected in AD, such as the temporal lobe, which leads to the characteristic memory loss. Figure 1A shows a number of other tauopathies which are distinct form Aβ pathology. Interestingly although tau is associated with all of these diseases there are some characteristic differences between them. One of the most striking differences is that in AD only neurons are affected while in other tauopathies, such as corticobasal degeneration, astrocytes are as well (Figure 1D). There is a major overlap in the symptoms of these diseases, which makes early differential diagnosis very challenging. 38 mutations have been identified in frontotemporal lobar dementia with Parkinsonism linked to chromosome 17 (FTLD-17) which affects tau (Brunden, Trojanowski et al. 2009) and tau neuronal inclusions are consistently indentified in these diseases. All of this data supports the fundamental role of tau in the pathogenesis of tauopathies. The reason for the differences between tauopathies is still uncertain however distinct differences in the tau isoforms can be seen between these diseases (Figure 1A).
1.2 Microtubule binding protein
Tau is mainly expressed in both the peripheral and central nervous system (Gu, Oyama et al. 1996) where it is predominantly found in neuronal axons (Lee et al 2001). In the adult, human brain, alternative splicing produces six different isoforms (Goedert, Spillantini et al. 1989) Figure2) which are defined by the number of microtubule binding repeats (3R or 4R) and N-terminal inclusion or exclusion of exons 2 and 3 (0N, 1N, 2N). Tau is a microtubule associated protein which is important for maintaining the stability of microtubules. Knocking out tau in primary neurons and also mice is not toxic and even helps to protect against Aβ pathology (Roberson, Scearce-Levie et al. 2007; Vossel, Zhang et al. 2010; Roberson, Halabisky et al. 2011). This may be due to the functions of tau overlapping with another microtubule associated protein called MAP1B. When MAP1B and tau are knocked out, this is severely detrimental, compared with MAP1B knockout alone (Takei, Teng et al. 2000). Tau has also been implicated in a number of other processes such as in regulating signaling pathways. In dendrites tau binds to the inner leaflet of the plasma membrane (PM) and acts as a scaffolding protein for the transport of fyn kinase (Morris, Maeda et al. 2011). Tau is regulated by a number of different post-translational modifications. The most important is phosphorylation and the common occurrence of hyperphosphorylated tau in NFT suggests its importance for their formation.
Figure 2 Tau protein isoforms. There are six different isoforms of tau. Half of which have 4 MT binding repeats and the other half have 3 MT binding repeats due to alternative splicing of exon ten. Between isoforms with the same repeats alternative splicing at exons 2 and 3 results in different N-termini. All 38 mutations, associated with FTLD-17 are indicated. These predominantly affect exons 9-12 and intron mutations between exon 10 and 11
Disease
Average age
onset
Causative
agent(s)
Neuronal
inclusions
Early
symptoms
Pathology
Affected brain
regions
Genetic risk
factors
AD 65 Tau (6 isoforms) Aβ plaques Neurofibrillary tangles (NFT) Memory loss NFT – hyperphosphorylated tau Temperal lobe, parietal lobe, parts of the frontal cortex and cingulated gyrus Mutations in presenilins 1 and 2 as well as amyloid precursor protein (APP). APOEε4 is the highest genetic risk factor Picks disease 54 Tau (3R isoforms –no exon 10)
Pick bodies Personality change
Pick bodies and cells CA1 of hippocampus, dendate gyrus, neocortex Apolipoprotein polymorphism Corticobasal degeneration
60 Tau (with exon 10) NFT Movement and
cognitive disfunction
Neuronal and Astroglial inclusions
Cerebral cortex and basal ganglia
Overlap with progressive supranuclear palsy Progressive supranuclear palsy
63 Tau (with exon 10) NFT (sometimes lewy bodies)
Motor defects Neuronal and astroglia tau tangles
Subcortical regions Halotype H1 (dinucleotide polymorphism) PERK genetic variation FTLD-17 30-50 Tau (isoforms depend on mutation) NFT Personality and behavioural changes. Parkinson’s + syndrome Gliosis, spongiform change Frontotemperal neuron loss Tau mutations (38) and progranulin mutations
AD
Picks disease
Corticobasal degeneration
A
Figure 1 Tauopathies A) summary of the most common tauopathies. These are all associated with NFT formed by tau, which is disparate from Aβ in all diseases except AD. Furthermore the brain areas which are affected are also different. However the majority of symptoms are homogeneous and therefore differential diagnosis, especially at early stages, is difficult. A number of genetic risk factors have been identified, the majority of which directly implicate tau (Baker, Mackenzie et al. 2006), OMIM#172700, OMIM#6011104, (Grover, Houlden et al. 1999; Hoglinger, Melhem et al. 2011) (Sergeant, Wattez et al. 1999). B) AD post mortem brain material, which has been stained for NFT C) Pick bodies in post-mortem brain material of an individual with Pick’s disease (Murayama, Mori et al. 1990). D) Post-mortem analysis of corticobasal degeneration shows astroglia plaques (*) of tau, which are common in most tauopathies except AD (Armstrong and Cairns 2009).
1.3 Tau phosphorylation
Phosphorylation is by far the most common post-translational modification of tau with 85 potential sites. It is still unclear what the physiological function of these phosphorylation sites are; however tau hyperphosphorylation is involved in hamster hibernation, although why this happens is still unclear. (Arendt, Stieler et al. 2003). Tau phosphorylation also negatively regulates its binding to microtubules. It is now commonly accepted that hyperphosphorylation of tau, in tauopathies, leads to microtubule (MT) instability (Figure 2). Microtubules can be thought of as highways between the synapse and the cell body of neurons where both organelles and proteins are transported. This transport is fundamental to the function of synapses and so microtubule instability, due to tau hyperphosphorylation, is thought to cause synaptic loss (Figure 2). Interestingly a number of mutations in tau, associated with FTLD-17 reduces the proteins affinity for MT(Brunden, owski et al. 2009). Pathologically this is thought to be due to an imbalance between phosphatases and kinases. GSK-3β is one of the main kinases which phosphorylates tau at residues associated with tauopathies (Wagner, Utton et al. 1996). A number of in vitro and in vivo studies have shown inhibiting GSK-3β has the potential to alleviate tauopathies (Gong, Park et al. 2011; Onishi, Iwashita et al. 2011; Hurtado, Molina-Porcel et al. 2012). Tau also exhibits a gain of toxic function in tauopathies (Brunden, Trojanowski et al. 2008).
1.4 Tau oligomers not aggregates are toxic
Hyperphosphorylation of tau alters its confirmation which allows for their interaction leading to the formation of tau oligomers which subsequently form paired helical filaments and finally NFT. One study used a transgenic mouse model that allowed for controlling the expression of tau P301L (associated with FTLD-17) (Santacruz, Lewis et al. 2005). Inducing expression caused memory loss, NFT formation and neuronal degeneration. Stopping expression of the mutant tau improved memory and decreased neuronal loss. Interestingly NFT continued to increase, which means they are a protective mechanism (Santacruz, Lewis et al. 2005).In cell cultures tau oligomers, rather than aggregates have been shown to be toxic (Gomez-Ramos, Diaz-Hernandez et al. 2006). This has been further verified in mice models where tau oligomers cause both synaptic and mitochondrial dysfunction, disparate of NFT formation (Brunden, Trojanowski et al. 2008; Lasagna-Reeves, Castillo-Carranza et al. 2011). Finally analysis of AD post-mortem material has shown that tau oligomers are detected before NFT formation and onset of disease. These oligomers are able to spread throughout the brain, as shown by injection of mutant tau into mice (Clavaguera, Bolmont et al. 2009; Chmielnicki 2012) in a prion like fashion (Gousset, Schiff et al. 2009). This means tau oligomers only
Figure 2 Toxic loss of tau function. The top image shows a normal neuron where tau binds to and stabilises MT, which allows for transport of protein and organelles between the synapse and the cell body. In tauopathies tau is hyperphosphorylated which causes its release from MT. This results in microtubule instability and subsequent synaptic loss. Tau then forms into intracellular, insoluble, aggregates. Whether tauopathies are primarily due to a toxic loss of function or a toxic gain of function is controversial (Trojanowski and Lee 2005)
chance of developing the disease far more likely. There is currently strong evidence supporting the hypothesis that tau oligomers show a toxic gain of function. We need to understand the early events in tau phosphorylation and oligomerisation to identify therapeutic targets which will reduce the formation of toxic tau oligomers and subsequently inhibit their formation and spread through anatomically linked brain regions to help alleviate the disease.
1.5 Activation of the unfolded protein response is an early event in tauopathies
We have identified activation of the unfolded protein response (UPR) as an early event in tauopathies (Song, De Sarno et al. 2002; Hoozemans, Veerhuis et al. 2005; Hoozemans, van Haastert et al. 2009; Nijholt, de Graaf et al. 2011).Chronic activation of the UPR is associated with diffuse phosphorylated tau and not NFT which suggests that the UPR causes pathological tau phosphorylation. We have shown that over expression of tau in vivo and in vitro does not activate the UPR (unpublished data), which means that the UPR is activated by a disparate mechanism. Interestingly changes in glucose metabolism have been shown to activate the UPR and AD has been described as a “third type of diabetes” due to the associated epidemiological risk factor (Accardi et al 2012). Consequently we need to understand whether activation of the UPR is sufficient to cause pathological tau phosphorylation and subsequent oligomerisation to further validate the UPR as a therapeutic target and to understand the pathogenesis of tauopathies.
The unfolded protein response (UPR) acts to maintain protein homeostasis in the endoplasmic reticulum (ER) by stopping the accumulation of toxic misfolded proteins. This is done primarily through inhibiting protein translation and increasing the expression of chaperones and proteins involved in endoplasmic reticulum associated degradation (ERAD) (Figure3). Failure to remove misfolded proteins leads to chronic activation of the UPR which in turn results in cell death. The central protein involved in activating the UPR is the Hsp70-class chaperone glucose regulated protein-78, which is also known as BiP. BiP is able to detect unfolded proteins, which causes its release from three proteins, which make up the three “arms” of the UPR (Ron and Walter 2007). These are the pancreatic ER kinase-like ER kinase (PERK), activating transcription factor 6 (ATF6) and the inositol requiring enzyme 1 (IRE1). IRE1 splices XBP1 pre-mRNA to produce the transcription factor XBP1 (Brewster, Linseman et al. 2006), which regulates genes that activate ER associated degradation (ERAD) of misfolded proteins (Yoshida, Matsui et al. 2003). PERK activation leads to a global decrease in protein synthesis via inhibition of the guanine nucleotide exchange factor eIF2B (Harding, Zhang et al. 1999). PERK also activates C/EBP- homologous protein (CHOP) which leads to
cell death through inhibition of BCL-2 (McCullough, Martindale et al. 2001). We have also shown that the UPR leads to prefential activation of the lysosome (Nijholt, de Graaf et al. 2011).
1.8 Removing tau: Macro-autophagy
There are three main types of autophagy however macroautophagy will be focused on (Figure 5). The autophagosome consists of a double membrane structure (phagophore) which engulfs predominantly proteins and organelles, in the cytoplasm. Macro-autophagy leads to the degradation of these complexes through fusion with the lysosome, which leads to the formation of an autophagolysosome. The lysosome is formed through endocytosis, of the plasma membrane (PM). Hydroxylases are targeted to these vesicles, from the golgi, and once the pH reaches 4.6-5 they fuse with the autophagosome and degrade its contents (Nijholt, De Kimpe et al. 2011). Lysosome dysfunction is common in tauopathies due to aging and Aβ in AD (Levine, Kroemer et al. 2009, Silva, Esteves et al. 2011). The current hypothesis is that lysosome dysfunction contributes to tau Figure 3 The unfolded protein response in tauopathies. Activation of the UPR is an early event in tauopathies. Acutely this leads to a protein translational block, ERAD and an increase in chaperones to restore protein homeostasis. Chronic activation leads to cell death, primarily mediated though the mitochondria, results.
neurotoxicity (Zhang, Sheng et al. 2009). However an in vitro model is needed to understand how lysosome dysfunction contributes to tau oligomerisation, as this is poorly understood. More research is needed in this area as there are wide reaching implications for understanding tauopathies and therapeutic interventions. Reducing the concentration of tau before oligomers form also represents a potential therapeutic target.
1.7 Removing tau:
Proteasome dysfunction has been identified as a common event in AD (Keck, Nitsch et al. 2003; Nixon, Wegiel et al. 2005) and therefore is likely to have an important role in pathogenesis. Proteasome dysfunction has been shown to be a result of inhibition by NFT of tau, rather than a cause (Keck, Nitsch et al. 2003) of tauoapathies. Tau is degraded by the 26s proteasome which it is targeted to by polubiquitin tags. The 26s proteasome consists of an ATP dependant regulatory cap called 19s and a cylinder with protease activity called 20s, which degrades the protein into its constituent amino acids for recycling (Sullivan, Shirasu et al. 2003; Marteijn, Jansen et al. 2006) .Under normal circumstances this is the predominant mechanism for tau removal however the relatively small pore sizes of the proteasome means larger tau aggregates cannot be degraded by the proteasome. Therefore it is unlikely that the proteasome plays an important role in the degradation of tau oligomers. Genetic oblation of specific components of this system causes a neurodegenerative Figure 4 Macroautophagy. Schematic representation of macroautophagy. The double membrane phagosome engulfs cytosolic contents and forms the autophagosome. This then fuses with a lysosome and the contents are degraded to their constituents before being recycled.
Proteins
lysosome
phagophore
autophagosome
hydroxylases
phenotype in mice (Bedford, Hay et al. 2008; Romero-Granados, Fontan-Lozano et al. 2011) although NFT of tau are not observed. This supports the hypothesis that proteasome dysfunction is not a primary but secondary cause of tauopathies. Both in vivo and in vitro studies have shown that decreasing tau has no aberrant effects and even helps to relieve Aβ related pathology (Roberson, Scearce-Levie et al. 2007; Vossel, Zhang et al. 2010; Roberson, Halabisky et al. 2011). The proteasome is an important therapeutic target as increasing its activation would decrease total tau levels in the cytoplasm and subsequently help to reduce the formation of tau oligomers and decrease disease progression.
Aim
This project can be split into two sections. In this first we aim to use a bimolecular complementation assay (BiFC) for visualising tau oligomerisation in living cells. We have already shown that activating the UPR leads to tau phosphorylation. Using this system we aim to identify whether this phosphorylation leads to tau oligomerisation. In the other section this projects aims to look at the role of the proteolytic system in tauopathies. We will induce lysosome dysfunction to identify how this effects tau oligomerisation. We will also use a screen of proteasome activators to identify their ability to degrade full length tau.
2 Materials and methods
2.1 Cell culture
Hela cells, an immortal cervival cancer cell line, were cultured in Dulbecco’s modified eagle’s medium (DMEM) with GLUTAMAX (Gibico) supplemented with 10% fetal calf serum (FCS) (BRL) and 100 u/ml penicillin (Yamanouchi Pharma BV) and 100μg/ml streptomycin (sigma). Cells were maintained as a monolayer culture at 37 ۫C in a humidified incubator containing 5% CO2. Cells were
trypsinised, counted using the countess (Invitrogen) and seeded onto 24 well plates (Invitrogen) with a density of 10 x 105. Cells were seeded for 24hrs before treatment. Cells were treated with different concentrations of bafilomycin (baf) to inhibit the lysosome, tunicamycin (Tm), 2-deoxoyglucose (2-DG) and thapsigargin (TG) (all Sigma) to activate the unfolded protein response (UPR) and okadaic acid and wortmannin for a positive control of tau phosphorylation (Sigma). Salubrinal (Sal) (Calbiochem) was used to inhibit the dephosphorylation of elF-2α. All treatments were done for 24 hours unless otherwise stated.
2.3 Transformation
Both of the split-venus constructs were transformed into top10 cells (Invitrogen). Between 50 and 100ng of each plasmid was added to cells before incubation on ice for 10 minutes. The cells were then heat shocked for 30 seconds at 42 ۫C. 450μl of SOC medium (Invitrogen) was added and incubated at 37 ۫C and 600 rpm for 1 hr. The cells were then plated on lysogeny broth (LB). Lysogeny broth contained 1% bacto tryptone (Brunschwig chemie), 0.5 yeast extract (Brunschwig chemie) and 1% NaCl (Merck, KGaA) and H2O, before being autoclaved. Once cool the antibiotic ampicillin (Sigma)
was added at a concentration of 1:1000 and 1.5% agar was added to make plates. Plates were incubated at 37 ۫C overnight (O/N). 200ml of LB was then inoculated with the bacteria overnight at 37 ۫C. A plasmid midi-prep (Qiagen) was then used according to manufacturer’s instructions. DNA concentrations were determined using a nanodrop (ND-1000) spectrometer and corresponding software.
2.4 Cell transfection
Cells were transfected using lipofectamine 2000 (Invitrogen), according to manufacturer’s instructions. Briefly all plasmid constructs used a 1:4 ratio of lipofectamine to DNA, which was added to 60 μl optimem per well and incubated for twenty minutes. The cell culture medium was changed to one without antibiotics before 60 μl of the solution was then added drop wise to each well and gently mixed. After four hours the cell culture medium was replaced with cell culture medium that contained antibiotics. All treatments were done after 24 hours for tau transfection. For the Tau construct pGN2 hTau 40 was previously cloned into a pcDNA4 vector (Invitrogen). This construct was
transfected for 24 hours before any treatment. The venus constructs were tranfected for times indicated in results section.
2.5 Western blot
Cells were kept on ice, washed with PBS (Gibico BRL) and lysed with lysis buffer (1% triton, PBS, phosphates and protease inhibitor Roche Diagnostics). Cells were scraped off and transferred to eppendorf (Invitrogen) tubes and incubated for 15 minutes before being centrifuged at 14,000 rpm for 20 minutes. The supernatant was then removed and the protein concentration was quantified (BioRad) via analysis with the FLUOstar Omega (BMG LABTECH) and analysed using associated Omega software. An equal amount of protein for all samples was then added to SDS and PBS before being boiled at 95 ۫C. Samples were then run on a 10% SDS-PAGE gel. After which wet blot apparatus (Biorad) was used to transfer the protein to a PVDF membrane (Immobilon). The membrane was then washed six times for five minutes with PBST (PBS and 0.05% tween). The membrane was then blocked for one hour in PBST with 5% milk. After which the anti-tau antibody (DAKO: A002401) was added to fresh PBST with 5% milk at a concentration of 1:1000 O/N. The membrane was then washed another six times for five minutes each with PBST. PBST with 5% milk was then added with a secondary antibody (polyclonal goat anti-rabbit, DAKO E0432), at a concentration of 1:2000 for one hour at room temperature. The lumi-light western blot substrates (Roche Diagnostics) were then used to visualise the tau protein with the LAS-3000 luminescent image analyser (Fuji Photo Film). Densitometry analysis was then done using advanced image data analyser (AIDA version 4.26, Raytest).
2.6 Flurorescent microscopy
The Leica DMI3000 B microscope was used in combination with corresponding software (Leica AF6000). All images were taken of live cells using a 1s exposure time, intensity of 10. Images were then analysed in adobe photoshop (CS4).
2.7 FACS
HeLa cells were trypsinised and then suspended in PBS with 5% FCS. Dead cells were stained for using 4',6-diamidino-2-phenylindole (DAPI) (1μl in 500 μl for 30 seconds). For fluorescent activated cell sorting (FACS) analysis the BD LSRFortessa (BD Biosciences) was used and 20,000 events for each sample were obtained using the BD FACSDiva software (BD Biosciences). The signal to noise of fluorescence was calibrated using non-transfected controls. The data was then analysed using FlowJo software (version 7.6.5) where the medium green fluorescence (GF) intensity was calculated.
2.8 Statistical analysis
All statistical analysis was done using SPSS statistics 20 (IBM). For FACS experiments the medium GF was analysed using a one way annova and then the Turkey post hoc test to determine individual sample differences. P > 0.05 was used as the cut off for a statistically significant difference. The Levene’s test of homogeneity was also used to test whether the data could be analysed using a one-way annova test. In all cases results were p < 0.05, hence the data could be analysed in this one-way. Excel 2003 (Microsoft) was used to produce all graphs and calculate standard deviations.
3 Results
3.1 Optimisation of transfection protocol for BiFC assay
To visualise full length tau dimers/oligomerisation in vitro we used a Bimolecular complementation assay (BiFC) (Figure 5A). In this system two tau constructs (Supplementary figures) are co-transfected into HeLa cells. Each tau construct is attached to the N or C terminal half of the fluorescent reporter respectively. When the tau proteins interact with each other, the two halves reconstitute the venus protein and emit green fluorescent light. First we established the optimal transfection protocol. The optimal ratio of DNA to Lipofectamine was 1:4, as this gave the best transfection efficiency (data not shown). We then wanted to reduce the background GF which results in unspecific interaction of the venus protein. 0.05μg of each construct in 24 well plates proved to be most suitable (Figure5B-E).
3.2 Validation of BiFC assay using Okadaic acid
Next we tested whether oligomerisation of tau leads to fluorescent emission in this BiFC assay. To induce oligomerisation of tau we used okadaic acid (OA), which inhibits protein phosphatase 2A (PP2A), and is commonly used in tau aggregation assays. Hela cells were transfected with both of the tau-venus constructs for 20hrs and then treated with increasing concentrations of OA. Using fluorescence microscopy an increase in GF was observed with increasing 5 and 10nM of OA (Figure6A). Due to the toxicity of OA treatment the total number of cells decreased compared to the control (data not shown). These results show that the basal level of GF is relatively high due to unspecific interaction of the venus halves. Tau dimers/oligomers were diffusely distributed throughout the cell. Some cells with OA treatment showed a localisation of tau dimers/oligomers to the plasma membrane (PM), however the morphology of the cells became more spherical (data not shown). Note that dead cells show a high green fluorescent intensity (GF). Consequently we used fluorescence activated cell sorting (FACS) as this enables the separation of dead cells from living cells. Untransfected cells were first analysed using FACS to determine any auto fluorescence. We treated the cells for 24 hours with a high concentration of OA (50mM). This analysis showed a small increase in GF intensity, which was not statistically significant (Figure 7A). Due to the toxicity of treatment a reduction of GF cells by nearly 70%, compared to the control was observed (Figure 7D). To ensure no dead cell were analysed a DAPI stain was also used (Figure7D). In this experiment the transfection efficiency was about 20% (Figure 7C ).
3.3 UPR activation using tunicamycin does not result in tau oligomerisation
We then used the BiFC assay to look at whether tunicamycin (Tm), which is a commonly used inducer of the UPR, would lead to tau dimer/oligomer formation via tau phosphorylation. No significant increase in GF intensity was observed with 24 hrs treatment of tunicamycin , compared to control (Figure 6B). The results indicate the tau does not oligomerise in the presence of Tm. Next we asked whether UPR activation using Tm increases the build-up of tau oligomers induced by another insult. Therefore to initiate oligomerisation we added OA to the Tm treatment. However no further increase in GF intensity was detected with combined OA and tunicamycin treatment compared with OA treatment alone (Figure 6B). Subsequently, we investigated whether UPR activation induces tau oligomerisation when degradational capacity is diminished. Here we used Bafilomycin (Baf) which inhibits the acidification of lysosomes. First transfected cells were treated with increasing concentrations of Baf to test whether lysosomal inhibition of its own leads to oligomerisation. An increase in GF, using fluorescence microscopy, for Baf treatment was observed (Figure 8A). Distribution of tau oligomers within the cells was diffuse throughout the cell in both controls and cells treated with Baf (data not shown). To quantify the fluorescence intensity FACS analysis was used which showed a small but reliable increase in median GF intensity (Figure 10A). Next, we treated the cells with a combination of Tm and Baf. No further increase in GF intensity was observed with Tm and Baf treatment compared to Baf treatment alone (Figure 8B). Taken together, these results show that Tm does not induce or increase tau oligomerisation using the BiFC system.
3.6 Activation of the UPR using 2-Deoxygucose increased tau oligomerisation
2-Deoxyglucose (2-DG) was used to induce the UPR via glucose deprivation due to the inhibition of glycolysis. HeLa cells transfected with the BiFC assay were treated for 24 hrs with 2-DG which led to an increase in GF intensity compared to the control (Figure 9A).Furthermore an even distribution throughout the cell in both the control and the cells treated with 2-DG was observed. Again we tested whether UPR activation increases oligomerisation induced by another insult. Observed by microscopical analysis treatment with a combination of 2-DG and OA led to a further small increase in GF compared to DG treatment alone (Figure 9A). Second Baf was used in combination with 2-DG. Again a small increase in GF was observed compared to 2-DG treatment alone (Figure 9A). FACS analysis was then used to quantify the changes in GF intensity. Figure 10A shows that 2-DG treatment causes about a 25% increase in GF intensity, which is reproducible. However only a small and not significant increase with 2-DG and Baf treatment was measured, compared to 2-DG only treatment. A reduction in GF cells was also observed due to the toxicity of the treatments (Figure 10D). These results demonstrate that 2-DG leads to tau oligomerisation. If 2-DG treatment increases
B
C
D
E
Venus
protein
Tau dimers
A
Tau monomers
0.25µg
0.15µg
0.15µg
0.05µg
Figure 5 Optimising transfection of the BiFC assay A) Tau is fused to the N terminal and C terminal half of the venus protein. Coexpression of these constructs leads to complementation of the venus protein when the tau protein oligomerises. B-E). Fluorescent microscopy was used to identify the best concentration of the BiFC constructs for transfection in HeLa cells. In all cases a ratio of 1:4 of DNA to lipofectamine was used, as this proved to be most effective. All cells were transfected for 20 hours, except for D which was transfected for 6 hours. The indicated concentrations of DNA were used to determine that 0.05µg of each construct was the most suitable.
tau oligomerisation in the presence of another stimulator, has to be determined in further experiments.
3.8 UPR activators and positive controls
We were unable to quantify an increase in GF intensity with OA treatment (Figure 7A). Consequently we used wortmannin (Wat), which is an activator of GSK-3β, to stimulate tau oligomerisation. Figure 11 shows an increase in GF; however this could not be quantified using FACS analysis (Figure 12A). Second we then tested whether thapsigargin (Tg) and salubrinal (Sal) treatment could induce tau dimers/oligomers. An increase was observed using fluorescent miscroscopy but again this could not be quantified using FACS analysis (Figure 11). The distribution of the GF was diffuse throughout the cells for all treatments in this experiment (data not shown). A significant decrease in GF cells was observed especially for Sal treatment where a loss of 60-70% compared to the control was observed (Figure 12D).
Figure 6 Okadaic acid increased GF but Tunicamycin and Tunicamycin and Okadaic combinations did not
increase or exacerbate the GF. A). HeLa cells transfected with both constructs of the venus system for 20hrs
and then treated with indicated concentrations of okadaic acid (OA) for 6hrs. B) HeLa cells transfected with both constructs of the BiFC assay for 20hrs and then treated with tunicamycin (Tm 0.5μg) or Tm and indicated concentrations of OA in combination. N=5 arrows indicate some dead cells.
0 0.2 0.4 0.6 0.8 1 1.2 1.4 control OA 0 0.2 0.4 0.6 0.8 1 1.2 control OA
A
Medi a n G F a s fra cti o n o f co n tr o lA
C
B
B
D
G F cel ls a s f ract ion o f cont rolFigure 7 FACS analysis of UPR activators and positive controls. A) The median GF as a fraction of the control was calculated for HeLa cells transfected with the BiFC assay for 24 hours and then treated for 24hours as with okadaic acid (50nM). A small increase in median GF, compared to the control, can be seen. Bars indicate standard deviation. B) Gate used to isolate living cells. C) Gate used to select living GF cells (also stained for using DAPI to identify dead cells). D) The number of GF cells as a fraction of the control is
B
C
Figure 8 Bafilomycin increased GF but Tunicamycin and Tunicamycin and Bafilomycin combinations did not
increase or exacerbate GF. A) HeLa cells transfected with both constructs of the venus system for 20hrs and
then treated with indicated concentrations of okadaic acid (OA) for 6hrs. B) HeLa cells transfected with both constructs of the BiFC assay for 20hrs and then treated with tunicamycin (Tm)(0.5ug) or Tm and indicated concentrations of OA in combination.
N=5
A
2-DG
Figure 9 2-DG increases Tau oligomers. A) 2-DG treatment (40mM) resulted in an increase in GF intensity. This also increased GF in combination with Baf and OA compared to Baf and OA acid alone. B) GF is distributed evenly throughout cytoplasm of the cell except for with OA treatment where some cells showed GF concentrated at the plasma membrane (PM).
N=3 Baf Baf + 2-DG OA OA + 2-DG B Control 2-DG OA 10nM OA 10nM + 2-DG Baf 50nm Baf 50nm + 2-DG OA 10nM Control
Figure 10 FACS analysis of UPR activation and/or lysosome dysfunction. A) The medium GF for only transfected cells was analysed for the indicated treatments. The concentration of Baf used was 50nM and the concentration of 2-DG was 40mM. A statistically significant increase from control was seen with Baf, 2-DG and Baf + 2-DG treatment. The difference between Baf and 2-DG+Baf treatment is also significant. * p<0.002, ** p<0.001. Bars indicate standard deviation. B) This shows the gate used to select for living cells. C) This graph shows the gate used to select only GF cells. In this experiment the transfection efficiency was about 20%. DAPI intensity allowed to select against dead cells. D) The number of GF cells analysed is indicated (compared to the control). This indicates a loss in cells due to treatment. E) This histogram compares the GF intensity to the number of cells. This clearly indicates an increase in the number of cells with higher GF (with the same profile as seen in A). Representative samples for each treatment were used in this example. (Green - control, orange
Baf 50 0 0.2 0.4 0.6 0.8 1 1.2
control baf 2-DG 2-DG + baf
50 GF cells as fraction of control
A
B
C
D
E
C ount of ce lls3.9 Proteasome activators
An array of 14 compounds which have been identified as proteasome activators were tested in HeLa cells transfected with tau to study their effect on tau protein levels. The cells were transfected with full length tau for 24 hours and then treated for 24 hours before protein levels were analysed using western blot. The activators 1-10 showed either an increase or similar levels of tau compared to the control (Figure 13) Compound 12 also shows an increase in tau levels. However a decrease is observed with compounds 11, 13 and 14 (Figure 13). 15 (IU1) was used as a positive control; however no difference in tau levels from the control were observed. The control triplicates show equal tau levels, which suggests the observed differences in tau are not due to transfection or loading differences. The experiment for compounds 11, 13 and 14 and treatment were done in triplicate. This showed a significant decrease in tau protein with compound 11, compared to the control (Figure 14). However compounds 13 and 14 showed only a small decrease in tau protein compared to the control (not significant) A 48hr treatment (data not shown) was also done for compound 11 however only a small reduction in tau was observed (not significant), a small increase using the positive control (compound 15) was also used.
Figure 11 Positive controls and UPR activators. Fluorescent microscopy of HeLa cells transfected with the BiFC assya for 24 hours and then treated, as indicated, for 21 hours. 40nM od 2-DG shows a small increase in GF. 50nM of OA shows an increase in GF and brightfield microscopy indicates round cells and increased cell death An increase is also seen for Sal (100mM), TG (0.5μM) and Wat (100nM) treatment.
0 0.2 0.4 0.6 0.8 1 1.2
control Sal TG Wat
0 0.2 0.4 0.6 0.8 1 1.2
control sal TG wat
M ed ia n GF a s a f ra ct io n o f co n tr o l G F c el l co u n t as a fr ac ti o n o f co n tr o l
Figure 12 FACS analysis of UPR activators and positive controls. A) The median GF as a fraction of the control was calculated for HeLa cells transfected with the BiFC assay for 24 hours and then treated for 24hours as indicated. A small but not significant increase was seen with OA treatment. No increase was seen with any of the other treatments. Bars indicate standard error. OA (50nM), Tg (0.5μM), Sal (100nM), Wat (100nM). B) Gate used to isolate living cells. C) Gate used to select living GF cells (also stained for using DAPI). D) The number of GF cells as a fraction of the control is indicated. A large reduction in the number of cells from all treatments is seen. Especially with OA treatment where over a 60% decrease, compared to the control, is seen. For treatments N=1 in triplicate. A 4hour treatment was also done (experiment not shown) with the same result – except no increase in OA was observed either. Bars indicate standard error.
A
B
C
D
GF c el l co u n t a s a f ra ct io n o f co n tr o l M e di an G F a s a f ract ion o f cont rol GF ce lls a s f ract ion o f cont rolFigure 13 Screen of proteasome activators for tau protein reduction. A) HeLa cells were transfected with tau and then treated for 24 hours with 2μM of each proteasome activator. A western blot against total tau was then done. The consistent amount of tau in the controls suggests all tau transfections were equal. B) The densitometry analysis from the western blot shows the amount of tau as a fraction of the control. In the majority of cases an increase in tau is actually observed. For compounds 11, 13 and 14 a modest decrease is seen. 15 (positive control) showed no decrease in tau compared to the control. N = 1, C=control
B
C 11 12 13 14 15 Controls 100 kD Anti-tauA
100 kD 1 2 3 4 5 6 7 8 9 10 Control Anti - tau To tal Tau as a fr ac tion o f c o n tr o l0 0.2 0.4 0.6 0.8 1 1.2 control 11 13 14
B
100 KD -
11
13
14
control
A
*Figure 14 Proteasome activators. A) HeLa cells were transfected with tau and then treated for 24 hours with 2μM of proteasome activators 11, 13 and 14. A western blot against total tau was then done. B) The densitometry analysis from the western blot shows the amount of tau as a fraction of the control. A significant decrease in tau protein, compared to the control is seen. Only a small decrease with the other two proteasome activators was observed. N=1 in triplicate. Bars – standard deviation,
*
p<0.005. One way annova used for statistical analysis.*
To tal Tau as a fr ac tion o f c o n tr o l4 Discussion
We previously identified activation of the unfolded protein response (UPR) as an early event in a number of different tauopathies (Hoozemans, Veerhuis et al. 2005; Hoozemans, van Haastert et al. 2009; Nijholt, van Haastert et al. 2012). The UPR can be activated by metabolic stress which is a strong epidemiological risk factor for AD (Accardi, Caruso et al. 2012) and genetic variants in PERK (a central protein in the UPR) increases the risk of developing progressive supranuclear palsy (Hoglinger, Melhem et al. 2011), suggesting UPR activation is important in tauopathies. In vitro UPR activation leads to pathological tau phosphorylation by GSK-3β (Rsense, Ferreiro et al. 2008). However we have shown in hibernating hamsters that activation of the UPR leads to tau hyperphosphorylation but does not cause tauopathy (Van der Harg et al [submitted]). This study aimed to identify whether phosphorylation of tau, mediated by UPR activation, leads to tau oligomerisation in vitro. To achieve this we used a BiFC assay developed by Outeiro et al. Previously this group has published several papers, using this assay, to identify the mechanisms for the aggregation of proteins implicated in neurodegenerative diseases (Outeiro, Putcha et al. 2008; Herrera, Tenreiro et al. 2011). In another line of study we looked at the role of the proteolytic system. Lysosome dysfunction is common in tauopathies and is an important therapeutic target. In this study we aimed to identify the effect of lysosome dysfunction on an in vitro assay for tau oligomerisation. We also aimed to screen 14 proteasome activators for their ability to degrade full length tau. This is important because a compound capable of doing this would reduce total tau and therefore reduce the formation of toxic tau oligomers and help to alleviate the tauopathies
To validate the BiFC assay we tested whether inducing tau oligomerisation led to fluorescent emission. We used Okadaic acid (OA) which inhibits protein phosphatase 2A and hence leads to an increase in tau phosphorylation (Zhang and Simpkins 2010). Microscopical analysis suggested tau oligomerisation, with treatment, but we were not able to quantify this signal using FACS analysis. This could be due to the toxicity reducing the number of cells by 60-70%, compared to the control, resulting in the statistical power not being sufficient to identify an increase in GF. The tau oligomers formed could also be toxic leading to cell death and resulting in an increase in GF intensity not being measured. The highly dividing Hela cells could also reduce the burden of tau dimers/oligomers per cell. The disparity between microscopical analysis and FACS quantification could also be due to the difficulty of identifying a difference in GF with microscopy due to the high GF of dead cells and cell death after treatment. The control also shows a high basal GF. This high GF could be due to tau dimer/oligomer formation because tau is being over expressed and any interaction is irreversible.
This hypothesis is supported by Outeiro et al (2008) whom using the BiFC assay for the protein alpha-synuclein identified high molecular weight species.
We then tested whether phosphorylation of tau, by UPR activation, was sufficient to induce tau dimer/oligomer formation. Previously we have shown that a number of UPR inducers such as tunicamycin and 2-DG lead to phosphorylation of tau, by GSK-3β, at the pathologically relevant site ser396 (unpublished data). 2-DG mediates this process by inducing metabolic stress via glycolysis inhibition, which in turn activates the UPR. In this study we have shown that treatment with 2-DG leads to a significant increase in tau dimer/oligomer formation using the BiFC assay (~25%). Consequently in this model activation of the UPR, by metabolic stress (using 2-DG), leads to tau oligomerisation which supports our hypothesis that chronic activation of the UPR causes tau oligomerisation. This has significance for tauopathies because metabolic stress is an important epidemiological risk factor for developing AD, which has been described as a “third type” of diabetes (Accardi, Caruso et al. 2012). Interestingly PERK increases during metabolic stress (Gomez, Powel et al. 2008) and analysis of post-mortem tissue from AD suggests PERK activates GSK-3β, via an unknown mechanism (Hoozemans, Haastert et al. 2009). PERK is one of the central proteins of the UPR, its main role being inhibition of protein translation. GSK-3β has been strongly connected to pathological tau phosphorylation and is a therapeutic target in its own right. Coupled to which variants in the gene that encodes PERK confers a risk for developing a tauopathy called progressive supranuclear palsy (Hoglinger, Melhem et al. 2011). Taken together this supports out hypothesis that chronic activation of the UPR is important in the pathogenesis of tauopathies.
TM was also used to activate the UPR via inhibition of N-linked glycosylation. No increase in tau oligomers were measured and no further increase was measured in combination with Baf treatment, which induced lysosome dysfunction (as the lysosome could have been degrading the formed oligomers). This could be due to the toxicity of the treatment, as Tm induces cell death, as well as the unspecific activation of the UPR. Thapsigargin and salubrinol were also used to activate the UPR. Salubrinol specifically activates the PERK pathway of the UPR and Thapsigargin through changing calcium levels in the cytoplasm. FACS analysis did not show an increase in tau dimers/oligomers. This could be due to the toxicity of the treatment which lead to a 60% decrease in cells compared to the control, which would result in the statistical power not being sufficient to identify a difference. Further experiments would be interesting as a small increase was observed microscopically however we have already noted that this is not necessarily accurate. The same results were obtained for
wortmanin, which specifically activates GSK-3β. This is only preliminary data and so different treatment times, treatment concentrations or cell types used could yield different results.
In the second part of this study we looked at the role of the proteolytic system in tauopathies. We used Baf to induce lysosome dysfunction, by inhibiting acidification, which lead to a significant increase in tau dimers/oligomers (~15%). When the UPR was also activated tau dimers/oligomers increased, compared to lysosome dysfunction alone. This is noteworthy because lysosome dysfunction is common in tauopathies and leads to a neurodegenerative phenotype in mice (Hara, Nakamura et al. 2006; Komatsu, Waguri et al. 2006). This is partly due to aging (Levine and Kroemer 2009) and is especially pronounced in AD where Aβ also contributes to lysosome dysfunction (Silva, Esteves et al 2011 ). In this study we have already shown that activation of the UPR causes tau oligomerisation. Therefore it is very interesting that the increase in tau dimers/oligomers seen with lysosome dysfunction (which is common in tauopathies), is exacerbated by UPR activation. We do not know whether lysosome dysfunction plays a causative role in forming tau dimers/oligomers or if there is simply an increase because they are no longer being degraded by the lysosome. Baf treatment leads to a reduction in cathepsin D (Pivtoraiko, Harrington et al. 2010) which is seen during lysosome dysfunction and in degenerating neurons during AD (Cataldo, Barnett et al. 1995). This protein is neuroprotective and its reduction has been shown to increase the toxicity of tau; by producing c-terminally truncated tau that act as seeds for oligomerisation (Khurana, Elson-Schwab et al. 2010). It has also been suggested that lysosome dysfunction leads to tau phosphorylation (Zhang, Sheng et al. 2009). Future research will identify whether lysosome dysfunction in this cell model leads to an increase in tau dimers/oligomers through a causal mechanism, which would have significant implications for tauopathies. Our group has also shown that activation of the UPR leads to an increase in autophagy (nijholt et al., 2011), which is linked to the lysosome (Figure 15).
We have shown that activation of the UPR with 2-DG and dysfunction of the lysosome with Baf leads to a significant increase in tau dimers/oligomers. An increase between DG only treatment and 2-DG and Baf treatment would therefore be expected; however only a small increase was observed. There are two possible reasons for this observation. The first is that the system is saturated and a further increase in GF intensity is not possible. This is unlikely as we have measured higher increases in GF intensity with this assay. The other possibility is lysosome dysfunction through 2-DG treatment; which has already been measured for Tm and Tg treatment (Lee, Noh et al. 2012). Activation of the
UPR normally increases lysosomal activity however chronic activation of the UPR could deplete pro-lysosomal proteins and contribute to its dysfunction.
Dysfunction of the proteasome has been identified in tauopathies and is thought to be a secondary event due to its inhibition by tau NFT (Keck, Nitsch et al. 2003). In this study we screened 14 different known proteasome activators to identify whether they were able to reduce the level of tau in transfected HeLa cells. Surprisingly an early screen showed that the majority of these activators lead to an increase in tau, something which has already been observed with proteasome in activators (Delobel, Leroy et al. 2005). Compound 11 was able to significantly reduce the total level of the full length tau protein after 24 hours (~50%) but not after 48 hours (data not shown) which may indicate the half-life of the compound. This has therapeutic potential as it could reduce the level of tau, in the cytoplasm, and hence help to stop the formation of tau oligomers. This is the first study to screen a number of different proteasome activators to look at their affect on total tau; a line of research which has been hindered by the lack of available proteasome activators. However there has been a study which observed a decrease in toxicity, in a cell model for Huntington’s disease, using proteasome activators (Seo, Sonntag et al. 2007). Compound 11 has therapeutic potential but this should only aim to restore normal proteasome function as deficits have been observed in mice lacking tau (Ke, Suchowerska et al. 2012) and the normal function of other proteins may be affected.
The main limitation of this study is that HeLa cells were used. Neurons would more closely resemble tauopathy, but the UPR is ubiquitous between cell types. It is unknown how the BiFC assay affects normal tau function and also whether this resembles tau oligomerisation in disease. There is also a high basal GF with this assay, which is likely due to the formation of tau dimers/oligomers due to the over expression of tau and its irreversible interaction. OA and some UPR activators were unsuccessful in causing oligomerisation. This could be due to the increase in cell death with treatments which decreased the statistical power or due to the cell type used. There are also a number of advantages with using the BIFC assay. Tau aggregation is a process which takes decades and therefore producing an accurate cell model over a couple of days is intrinsically difficult. For this reason there has been only limited success in producing a cell model for tau oligomerisation (Kolarova, Garcia-Sierra et al. 2012). The BiFC assay used in this study has a number advantages over similar approaches due to the relative ease of detecting differences in GF , in living cells, and the sensitivity of the assay allows for the visualisation of early tau species (Liu, Ni et al. 2004, Chun, Waldo et al. 2007; Chun, Waldo et al. 2011, Kolarova, Garcia-Sierra et al. 2012))
In this study we have shown that activation of the UPR leads to tau oligomerisation, in the cell model used. The UPR was activated by 2-DG which causes metabolic stress. This is significant as it shows chronic activation of the UPR through metabolic stress is sufficient for tau oligomerisation. Because metabolic stress is a significant epidemiological risk factor for AD and we have identified activation of the UPR in a number of tauopathies, this suggests that the UPR plays an important role in a large number of tauopathy cases ( Accardi, Caruso et al. 2012, Hoozemans, Veerhuis et al. 2005; Hoozemans, van Haastert et al. 2009; Nijholt, van Haastert et al. 2012). Hence the UPR represents a therapeutic target for which a number of drugs, which have already shown potential for treating neurodegeneration, have been identified (Hoozemans, Scheper 2012). Lysosome dysfunction leads to an increase in tau dimers/oligomers, in this cell model. As lysosome dysfunction is common in tauopathies this result further supports activation of the lysosome as a therapeutic target. This cell model can also be used as a platform to further understand how lysosome dysfunction causes tau neurotoxicity. Lastly we have identified a proteasome activator which leads to tau degradation. This compound has therapeutic potential as reducing tau would also reduce the formation of toxic tau oligomers. Figure 15 shows our model for the role of the UPR and the proteolytic system in tauopathies; based on this study and previous work.
4.1 Future research
The BIFC assay will be further optimised to reduce the GF observed in controls to improve the dynamic range. The BiFC assay will also be used in neurons using viral transfection. This will identify if some of the UPR activators and positive controls used did not increase tau dimers/oligomers due to the cell type used. The use of less toxic UPR activators would also be very useful as treatment times could be increased which would increase tau dimer/oligomer formation. Furthermore identifying the role of over expressing or knocking down different components of the UPR in tau phosphorylation as well as oligomerisation using the BIFC assay would be very interesting. This has already been done in vivo with a number of other neurodegenerative diseases, which has implicated the UPR (Sado, Yamasaki et al. 2009; Moreno, Radford et al. 2012; Vidal, Figueroa et al. 2012; Zuleta, Vidal et al. 2012). A native gel will also be used to visualise the tau species being produced. This will confirm whether larger tau oligomers/aggregates are forming which may stop the reconstitution of the venus protein, which could explain why an increase in GF intensity it not observed with all treatment types. But this is unlikely as Outeiro et al. (2011) have published data suggesting this is not the case. For Baf treatment SDS-PAGE analysis would identify whether the increase in tau dimers/oligomers observed is due to a decrease in their degradation or due to a causal mechanism.
Compound 11 (proteasome activator) will be tested using the BiFC assay to determine whether tau dimer/oligomer formation is reduced (when induced by 2-DG), further supporting it as a potential therapeutic. Total tau as a fracti on of contr ol
Figure 15. The UPR and proteolytic system in tauopathy. Under normal circumstances tau is degraded by the 26s proteasome. We hypothesise that activation of the UPR leads to tau phosphorylation through GSK-3β, which alters the structure of tau. This will lead to the formation of tau oligomers and then NFT which inhibit the proteasome. Under “normal”ER stress conditions activation of the UPR leads to activation of the lysosome and autophagy. However under chronic ER stress in tauopathies we hypothesise that this decreases the availability of pro-lysosomal proteins and hence contribute to its dysfunction. Lysosome dysfunction is commonly associated with tauopathies (26s proteasome picture from Marteijn, Jansen et al. 2006) which leads to an increase in tau oligomers and may also contribute to their formation.
Abstract
Tau is a microtubule associated protein which has been implicated in a number of diseases termed tauopathies, the most common being Alzheimer’s disease (AD). One of the main pathological hallmarks of tauopathies is tau aggregates which contain hyperphosphorylated, insoluble tau. Using brain material from AD and other tauopathies (without Aβ pathology) our group has shown that activation of the unfolded protein response (UPR) precedes tau aggregation. Cell models indicate that UPR activation leads to phosphorylation of tau. The UPR is primarily a protective mechanism however we hypothesise that chronic activation contributes to tau pathology. Interestingly changes in glucose metabolism, which is a high risk factor for AD, leads to UPR activation. This led to the question; does UPR activation lead to tau aggregation? In this study we used a bimolecular fluorescent complementation assay (BiFC) to visualise tau oligomers in a cell model. This showed that activation of the unfolded protein response (UPR) using 2-deoxyglucose, induces tau oligomerisation.
In another line of study we looked at the role of the proteolytic system. Inducing lysosome dysfunction using bafilomycin, in this cell model leads to an increase in tau oligomers, supporting it as a therapeutic target. We also screened 14 proteasome activators for their ability to reduce tau. We identified one potential candidate. In conclusion we have used a cell model for visualising tau oligomerisation, which is intrinsically difficult to achieve and we have shown both UPR activation and lysosome dysfunction significantly increase tau dimers/oligomers. We have also identified a proteasome activator which has therapeutic potential and can be tested in this system.
Key words
5 References
