α-synuclein induced synapse damage is enhanced by amyloid-β_1-42

The synapse degeneration in PD and DLB that is associated with oligomers of αSN [7‐9] was modelled in vitro. The addition of recombinant human αSN reduced the synaptophysin content of cortical neurons in a dose-dependent manner (Figure 1A). The synaptophysin content was reduced to 50% of control neurons (EC₅₀) following the addition of 500 nM αSN. This effect of αSN on synapses occurred at concentrations that did not kill neurons; for example, the addition of 10 μM αSN reduced the synaptophysin content of cortical neurons by greater than 80% without affecting their viability as measured by thiazolyl blue tetrazolium (100% cell survival ± 8 compared with 96% ± 6, n = 9, P = 0.22). Similarly, immunoblots showed that αSN reduced the amount of synaptophysin in neuronal extracts without affecting the amount of β-actin (Figure 1B). The addition of βSN, another member of the synuclein family of proteins [6], did not affect the synaptophysin content of cortical neurons. We found that the addition of αSN, but not βSN, also reduced the amount of synaptophysin in hippocampal neurons (Figure 1C), an observation consistent with a report that the loss of synapses in the hippocampus is characteristic of the PD patients that develop dementia [13].

The uptake of FM1-43, a fluorescent dye that is taken up into synaptic vesicles, was used as a measure of synaptic vesicle recycling and hence neurotransmission [21]. Here we report that the uptake of FM1-43 by cortical neurons was reduced following the addition of αSN, but not after the addition of βSN (Figure 2). This effect of αSN was observed at lower concentrations than that required to reduce the synaptophysin content of neurons and the concentration of αSN required to reduce synaptic vesicle recycling by 50% was approximately 30 nM.

Although transgenic mice studies showed that the expression of βSN reduced neurodegeneration in mice expressing human αSN [22, 23], the molecular mechanisms underlying the interactions between αSN and βSN are unknown. We found that pre-mixing αSN with an excess of βSN (1:10) reduced the αSN-induced loss of synaptophysin in cortical neurons, whereas pre-mixing αSN with human serum albumin (1:10) had no affect (Figure 3). As this result may have been caused by a direct effect of βSN on neurons, cortical neurons were pre-treated with 10 μM βSN, washed and then incubated with αSN. Pre-treatment of neurons with βSN did not affect the loss of synaptophysin induced by αSN (data not shown).

The amyloid hypothesis of Alzheimer's disease (AD) pathogenesis maintains that the primary event is the production of neurotoxic amyloid-β (Aβ) peptides following the proteolytic cleavage of the amyloid precursor protein into different sized fragments [24, 25]. These fragments include Aβ_1-42 which is widely regarded as a major pathogenic species in AD [26]. Since recent reports showed that αSN and Aβ_1-42 co-exist in heterologous oligomers [27, 28] the effect of Aβ_1-42 on αSN-induced loss of synaptophysin was examined by pre-mixing the two peptides. The addition of Aβ_1-42 in the ratio (1:50) increased αSN-induced synapse damage (Figure 4A). Thus, while the EC₅₀ of αSN alone was 500 nM, the EC₅₀ of Aβ_1-42:αSN (1:50) was 25 nM. These concentrations of Aβ_1-42 did not affect synapses when added on their own. In contrast, pre-mixing αSN with the control peptide Aβ_42-1 (1:50) did not affect αSN-induced loss of synaptophysin. Since the predominant Aβ species found within the brain is Aβ_1-40 [29] the effect of Aβ_1-40 on αSN was also tested. However, there was no significant difference in the synaptophysin content of cortical neurons incubated with αSN and neurons incubated with Aβ_1-40/αSN (1:50) (data not shown). These results may have been caused by a direct effect of Aβ_1-42 on the neurons. Our observation that pre-treatment of cortical neurons with 1 nM Aβ_1-42 did not affect αSN-induced loss of synaptophysin (Figure 4B) suggested that Aβ_1-42 did not sensitise neurons to the effects of αSN.

The ability of synthetic Aβ peptides to self associate results in a mixture of physical complexes ranging from small soluble oligomers to large fibrils. Since the dynamic nature of Aβ aggregation means that it is difficult to ascribe biological function to specific Aβ assemblies using synthetic peptides, the activity of naturally derived, stable Aβ oligomers was also examined. We found that pre-mixing αSN with 7PA2-conditioned medium (7PA2-CM), which contained naturally secreted stable Aβ oligomers [30, 31], increased the αSN-induced loss of synaptophysin in cortical neurons (Figure 4C). In contrast, pre-mixing αSN with CHO-CM had no effect.

The addition of Aβ oligomers also affected αSN-induced inhibition of synaptic vesicle recycling. Thus, pre-mixing αSN with 7PA2-CM enhanced αSN-induced inhibition of FM1-43 uptake into synapses. The concentration of αSN alone required to reduce synaptic vesicle recycling by 50% was 30 nM, while the concentration of αSN that had been mixed with 7PA2-CM to have a similar level of effect was 1 nM (Figure 5).

Next we examined whether non-toxic concentrations of αSN affected Aβ_1-42-induced loss of synaptophysin. Here we show that pre-mixing Aβ_1-42 with αSN increased the Aβ_1-42-induced loss of synaptophysin from neurons (Figure 6). Thus, while the EC₅₀ of Aβ_1-42 alone was 50 nM, the EC₅₀ of αSN:Aβ_1-42 (2:1) was 5 nM. In contrast, the addition of βSN (2:1) did not affect Aβ_1-42-induced loss of synaptophysin. Pre-treatment of cortical neurons with 10 nM αSN did not affect Aβ_1-42-induced loss of synaptophysin (data not shown). We also sought to determine if αSN affected another peptide that caused synapse damage. Synapse degeneration is a feature of human and experimental prion diseases [32, 33] which was modelled by the addition of the prion-derived peptide PrP82-146 to cortical neurons [15]. As shown in Figure 6B, there was no difference in the synaptophysin content of cortical neurons incubated with PrP82-146 and those incubated with a combination of αSN/PrP82-146 (2:1).

We explored the possibility that αSN increased the binding of Aβ_1-42 to synapses as an explanation of the effect of αSN on Aβ_1-42-induced loss of synaptophysin. Time course studies showed that Aβ_1-42 accumulated in synaptosomes isolated from cortical neurons after 1 hour. Therefore cortical neurons were incubated with 100 nM biotinylated Aβ_1-42, or 100 nM biotinylated-Aβ_1-42 that had been pre-mixed with 500 nM αSN for 1 hour. The amount of biotinylated Aβ_1-42 found in synaptosomes isolated from these neurons was not altered by pre-mixing with αSN (Figure 7A), indicating that αSN did not alter the binding, or trafficking of Aβ_1-42 to synapses. The effect of Aβ_1-42 on the accumulation of αSN in synapses was also examined. Cortical neurons were incubated with 200 nM αSN, or a combination of 4 nM Aβ_1-42 and 200 nM αSN (1:50), for 1 hour and synaptosomes prepared. Immunoblots showed that the amount of αSN found within synapses was not affected by presence of Aβ_1-42 (Figure 7B).

Cortical neurons were prepared from the brains of mouse embryos (day 15.5) as described [15]. Neurons were plated at 2 × 10⁵ cells/well in pre-coated 48 well plates (5 μg/ml poly-L-lysine) in Ham's F12 (PAA) containing 5% foetal calf serum (FCS) for 2 hours. Cultures were shaken (600 r.p.m for 5 minutes) and non-adherent cells removed by 3 washes in PBS. Neurons were grown in neurobasal medium (NBM) containing B27 components (PAA) for 10 days. Immunohistochemistry showed that the cells were greater than 97% neurofilament positive. Fewer than 3% of cells stained for glial fibrillary acidic protein (astrocytes) or for F4/80 (microglial cells). Hippocampal neurons were prepared from the brains of adult mice as described [54]. Hippocampi were dissected from brains and triturated in Ham's F12 containing 5% FCS, 0.35% glucose, 0.025% trypsin, and 0.1% type IV collagenase. After 30 minutes at 37°C, the cells were triturated and the cell suspension was passed through a 100 μM cell strainer. Cells were collected, washed twice and plated at 2 × 10⁵ cells/well in 48 well plates (pre-coated with poly-L-lysine). After 24 hours cultures were shaken (600 r.p.m for 5 minutes) to remove non-adherent cells, washed twice and the remaining neurons were cultured in NBM/B27 and 10 ng/ml glial-derived neurotrophic factor (Sigma) for 7 days. Neurons were incubated with peptides for 24 hours and the amount of synaptophysin in cell extracts measured.

Neurons were washed 3 times with PBS and homogenised in a buffer containing 150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 10 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate, 0.2% sodium dodecyl sulphate (SDS) and mixed protease inhibitors (4-(2-Aminoethyl) benzenesulfonyl fluoride hydrochloride (AEBSF), Aprotinin, Leupeptin, Bestain, Pepstatin A and E-46) (Sigma) at 10⁶ cells/ml. For immunoblots cells were homogenised in extraction buffer (as above) at 10⁷ cells/ml and digested with DNAse (Sigma) for 1 hour at 37°C. Cell debris was removed by low speed centrifugation (300 × g for 5 minutes).

The amount of synaptophysin in neuronal extracts was measured by ELISA [15]. Briefly, the capture mAb was anti-synaptophysin MAB368 (Millipore). Samples were added for 1 hour and bound synaptophysin was detected using rabbit polyclonal anti-synaptophysin (Abcam) followed by a biotinylated anti-rabbit IgG (Dako), extravidin-alkaline phosphatase and 1 mg/ml 4-nitrophenol phosphate. Absorbance was measured on a microplate reader at 405 nm and the synaptophysin content of samples was expressed as units where 100 units was defined as the amount of synaptophysin in untreated neurons.

The fluorescent styryl dye FM1-43 (Biotium) that is readily taken up into synaptic recycling vesicles [55] was used to determine synaptic activity as described [21]. Treated neurons were incubated with 1 μg/ml FM1-43 and 1 μM acetylcholine (ACh) for 10 minutes, washed 5 times in ice cold PBS and solubilised in methanol at 1 × 10⁶ neurons/ml. Soluble extracts were transferred into Sterlin 96 well black microplates and fluorescence was measured using excitation at 480 nm and emission at 625 nm. Background fluorescence was subtracted and samples were expressed as "% fluorescence" where 100% fluorescence was defined as the amount of fluorescence in untreated neurons incubated with FM1-43 and ACh.

Synaptosomes were prepared on a discontinuous Percoll gradient [56]. Briefly, 10⁶ cortical neurons were homogenized at 4°C in 1 ml of SED solution (0.32 M sucrose, 50 mM Tris-HCl, pH 7.2, 1 mM EDTA, and 1 mM dithiothreitol, and centrifuged at 1000 × g for 10 minutes. The supernatant was transferred to a gradient of 3, 7, 15, and 23% Percoll in SED solution and centrifuged at 16,000 × g for 30 minutes at 4°C. The synaptosomes were collected from the interface of the 15% and 23% Percoll steps. The fraction was washed twice (16,000 × g for 30 minutes at 4°C) and suspended in extraction buffer containing 150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 10 mM EDTA, 0.2% sodium dodecyl sulphate and mixed protease inhibitors.

The amounts of biotinylated Aβ_1-42 in extracts were determined by ELISA. Nunc Maxisorb Immunoplates were coated with 1 μg/ml protein A (Innova) followed by 0.1 μg/ml mAb reactive to amino acids 1 to 16 of β-amyloid (clone 6E10 - Signet) and blocked with 5% milk powder. Samples were boiled in 0.2% SDS, cooled and incubated for 1 hour. Biotinylated Aβ_1-42 was detected with extravidin-alkaline phosphatase and 1 mg/ml 4-nitrophenyl phosphate (Sigma). Absorbance was measured on a microplate reader at 405 nm and results were calculated by reference to a standard curve generated form serial dilutions of biotinylated Aβ_1-42.

Samples were mixed with an equal volume of Laemmli buffer, boiled, and subjected to electrophoresis on a 15% polyacrylamide gel (Invitrogen). Proteins were transferred onto a Hybond-P PVDF membrane (Amersham Biotech) by semi-dry blotting. Membranes were blocked using 10% milk powder; synaptophysin was detected using a mouse monoclonal antibody (mAb) anti-synaptophysin SY38 (Abcam), β-actin was detected by incubation with a mouse mAb (clone AC-74, Sigma) and human αSN was detected by incubation with mAb 211 raised against amino acids 121 to 125 of human αSN (Santa Cruz Biotech). These were visualised using a combination of biotinylated secondary antibodies (Dako), extravidin-peroxidase and an enhanced chemiluminescence kit.

Recombinant human αSN and βSN were purchased from Sigma. Synthetic peptides containing the amino acids 1 to 42 (Aβ_1-42) or 1 to 40 (Aβ_1-40) of the Aβ protein, biotinylated-Aβ_1-42 and a control peptide consisting of amino acids 1 to 42 in reverse order (Aβ_42-1) were obtained from Bachem. Peptides containing amino acids 82 to 146 of the human PrP protein (PrP82-146) and was a gift from Professor Salmona (Mario Negri, Milan). Aβ peptides were first dissolved in hexafluoroisopropanol, lyophilised and subsequently solubilised and stored at 1 mM in DMSO. Stock solutions of peptides were stored at 1 mM, thawed on the day of use and diluted/mixed in NBM for 1 hour at 37°C. Dilutions/mixtures were subjected to vigorous shaking (Disruptor Genie, full power for 10 minutes) before they were added to neurons. Chinese hamster ovary (CHO) cells stably transfected with a cDNA encoding APP₇₅₁ containing the Val717Phe familial AD mutation (referred to as 7PA2 cells) were cultured in DMEM with 10% FCS [30, 31]. Conditioned medium from these cells contains stable Aβ oligomers (7PA2-CM). Conditioned medium from non-transfected CHO cells (CHO-CM) was used as controls. These were mixed with peptides and subjected to vigorous shaking (as above), diluted in NBM and added to neurons.

Differences between treatment groups were determined by 2 sample, paired T-tests. For all statistical tests significance was set at the 1% level.