Background
Lewy bodies are intracellular deposits containing the ubiquitous CNS protein α-synuclein (SNCA), and are the pathological hallmark of Parkinson's disease (PD) [
1]. However, speculation continues as to the exact role of the protein under both healthy and pathological conditions. Brundin
et al. have proposed that extracellular SNCA (eSNCA) may be responsible for propagating PD pathology in grafted tissue [
2]. The mechanism for the physiological release is debated, but
in vitro studies have demonstrated that SNCA is secreted by living neurons and enters the surrounding medium [
3]. It is also possible that dying neurons release SNCA into the extracellular space. eSNCA is present in measurable quantities in cerebrospinal fluid and plasma of individuals with PD. Various groups have suggested that eSNCA over-stimulates the immune system resulting in a neurotoxic central immune phenotype [
4,
5]. These studies frequently use
in vitro techniques or over-expression models where the use of specific vectors may interfere with the immune process [
6]. Here we use direct
in vivo application of SNCA protein to study the inflammatory response.
The inflammatory response described in PD is thought to initially result from the activation of microglia [
5]. Traditionally, this has been seen to induce a cascade of proinflammatory cytokines that results in feed-forward immune stimulation and a hyperactive inflammatory response. While proinflammatory cytokines have been shown to be present in both post-mortem brains and the cerebrospinal fluid of PD patients [
7‐
9], it is possible that these inflammatory responses may,
in vivo, be relatively brief and disguise the true role of microglia in PD. It is possible, in terms of eSNCA, that they play a scavenger-like role, merely clearing debris rather than establishing an inflammatory response on the scale of that seen with more traditional mediators of inflammation.
In order to have a good basis for comparison, we employed an intranigral lipopolysaccharide (LPS) injection as a positive control. Recent PD research has used intranigral injections of LPS as a model of disease [
10,
11]. What is important about these studies is that they produced a model of dopaminergic cell death caused by a local inflammatory response. However, the inflammation in PD is unlikely to be triggered by the same pathways activated by LPS. To date, little comparison of the
in vivo inflammatory effects of SNCA and LPS has been made.
The aim of the present study was to determine the immune profile resulting from intranigral eSNCA and further to establish whether this profile differs significantly from that induced by LPS administration. Here, we show that SNCA does generate a significant inflammatory response in vitro and in vivo when compared to control protein, but the response is far less marked than that seen with LPS. However, we also show that the inflammatory response to eSNCA is greatly enhanced if an animal is also given a systemic injection of LPS. Thus, while the inflammatory profile resulting from stimulation with eSNCA or LPS are considerably different, systemic activation of the immune system can produce a local inflammatory response to eSNCA that is comparable to LPS.
Methods
Materials
SNCA peptide (rPeptide, Georgia, USA) was maintained as a stock solution of 6 μg/μl in phosphate-buffered saline (PBS; Invitrogen). Amyloid-β peptide (California Peptide Inc., California, USA) was maintained as a stock solution of 6 μg/μl in 0.1% dimethylsulfoxide/PBS. Bovine serum albumin (BSA; Sigma-Aldrich, Poole, UK) was maintained at a stock solution of 6 μg/μl in PBS. LPS (
E. coli 026:B6, Sigma-Aldrich) was maintained as a stock solution of 10 μg/μl in PBS. Peptide concentrations were chosen based on
in vitro dose response data (not included). LPS doses, both central and peripheral, were based on those currently used in the literature to produce a robust inflammatory response [
11,
12].
Cell culture
BV2 cells (a kind gift from Dr. David Brough, University of Manchester) were maintained in DMEM (GIBCO, Invitrogen, Paisley, UK) with 10% heat-inactivated FCS (GIBCO). Cells were treated with LPS, SNCA or amyloid-β in 12-well plates (1.5 × 105 cells/well) and supernatant samples were removed at time points up to 48 hours after treatment. Supernatants were analysed for TNFα release by ELISA (R&D Systems, Abingdon, UK) and plates were read using a BioRad Model 680 Microplate Reader (BioRad, Hemel Hempstead, UK). For microscopy, cells were grown on sterile coverslips and fixed in an ice-cold 3:1 acetone:methanol solution prior to mounting with DAPI mounting medium (Vector Laboratories).
Animals
Adult female ABH-Biozzi mice (6 months) were obtained from Charles River and housed under a standard 12-hour light/dark cycle. Animals were provided with food and water ad libitum and all procedures were carried out in accordance with the UK Animals (Scientific Procedures) Act, 1986. Animals were anaesthetized in a 2% isoflurane/oxygen mix (2 L/min) and placed in a stereotactic frame (Stoetling Co., USA) under maintenance anaesthesia (1.5%).
Intranigral injection of peptide
The skull was exposed and a hole drilled above the position of the substantia nigra pars compacta (SNpc) which lies -2.9 mm anterior, -1.3 mm lateral and -4.1 mm ventral from Bregma [
13]. Injections of 0.5 μl of stock solutions were made using a graduated glass capillary tube (Drummond Scientific Company, Broomall, PA, USA) over 5 minutes (0.1 μl/min) followed by 2 minutes of rest, to allow diffusion of the injected material, prior to removal of the needle.
LPS challenge
A subset of animals were challenged with LPS at 0.5 mg/kg i.p. 18 hours after receiving intranigral injections.
RNA extraction and cDNA preparation
mRNA was extracted using the RNeasy Mini Kit (Qiagen, Crawley, UK) according to the manufacturer's instructions. Briefly, 10-20 mg of frozen tissue was submerged in 300 μl lysis buffer containing 0.001% β-mercaptoethanol. Tissue samples were homogenised using a motor-driven disposable plastic pestle and the resulting suspension transferred to a Qiashredder Mini Spin® column which was then centrifuged at maximum speed in a microcentrifuge. The resulting lysate was mixed 1:1 with 70% ethanol and centrifuged through an RNeasy Mini Spin® column. The column was washed and treated with DNAse 1 for 15 minutes. The column was washed again to remove any final contaminants and RNA was eluted using RNase-free water. RNA samples were then diluted as necessary in order to input 400 ng total RNA into a 10 μl-reverse-transcription reaction. cDNA was synthesised using a Taqman® Reverse Transcription Reagent Kit (Applied Biosystems, Warrington, UK) as per the manufacturer's instructions.
Quantitative PCR
RT-PCR assays were performed as previously described [
14]. Samples were run against standard curves generated from serially-diluted cDNA from LPS-challenged mouse liver. Primer and probe sets for mouse NFkB, IL-1β, TNFα, TGFβ, COX2 and IL-6 were designed using the Roche universal probe library assay design centre. Samples were analyzed using a Roche Light Cycler 480
® (Roche Diagnostics, Welwyn Garden City, UK) and all reagents were used according to manufacturer's instructions. Briefly, gene-specific primers were designed and combined with a FAM/TAMRA labelled hybridization probe. PCR was run according to standard conditions [
14]. Analysis was performed using the standard curve to determine reaction efficiency followed by a comparative-cycle-threshold method. Results were expressed as relative expression corrected to the house-keeping gene glyceraldehyde phosphate dehydrogenase (GAPDH).
Nuclear and cytosolic p65 protein analysis
Fresh tissue was extracted from the injection site and run through the ProteoExtract kit (Merck, Nottingham, UK). Briefly, tissue was mixed with 250 μl Extraction Buffer 1 (including protease inhibitors) and incubated at 4°C for 10 minutes under agitation. Insoluble material was pelleted at 1000 G at 4°C for 10 min and the resulting supernatant, the cytosolic subproteome, was removed and stored. The pellet was mixed with 250 μl Extraction Buffer 2 and incubated for 30 min at 4°C under agitation. The insoluble material was pelleted at 6000 G at 4°C for 10 min. The supernatant, the membrane/organelle subproteome, was removed and the pellet mixed with 125 μl Extraction Buffer 3 (including 1.5 μl Benzonase). Following 10 min of incubation at 4°C the insoluble material was pelleted at 7000 G at 4°C for 10 min and the supernatant, the nuclear fraction, was removed. The final fraction, the cytoskeletal subproteome, was discarded. Western blots were performed on the nuclear and cytosolic subproteomal fractions. Subcellular fractions were analyzed by 1DE western blot probing for p65 (AbCam, UK), using actin (cytosolic) and HCDA1 (nuclear) as housekeeping proteins. Quantification was performed using ImageJ software using BSA injected animals as controls.
Tissue preparation
Animals were surgically anaesthetised with 0.1 ml pentobarbitone and transcardially perfused with heparinised saline (0.9%) followed by a periodate lysine paraformaldehyde solution (PLP: 2% paraformaldehyde, lysine, periodate and 0.05% glutaraldehyde). Brains were removed, post-fixed in PLP for 4 hours and further fixed in 30% sucrose for > 12 hours. 10 μm-sections were cut on a cryostat (Leica, Bucks, UK) and mounted on gelatine-coated slides.
Immunohistochemistry
An avidin-biotin-peroxidase method was employed for staining the tissue sections [
15]. Antigens were detected using antibodies against Iba-1 (Abcam, Cambridge, UK) to detect activated microglia and ICAM-1 (Abcam) Binding was detected using a biotinylated secondary antibody and an ABC standard kit (Vector Laboratories). Visualization was performed using a 0.05% diaminobenzene hydrochloride solution (DAB; Sigma). ICAM-1 and Iba1 analysis was performed using a light microscope (Nikon Labophot-2, Surrey, UK) fitted with an eyepiece graticule of known area. Vessels were counted in areas of highest density around the site of injection and expressed as number of vessels per mm
2. TUNEL labelling was performed using a NeuroTacs kit (R&D Systems, Abingdon, UK) as per the manufacturer's instructions and developed using a light microscopy-based method (DAB).
Discussion
The purpose of this study was to investigate the inflammatory properties of SNCA in vitro and in vivo. Our data reveals that non-aggregated wild-type (wt) synthetic human SNCA produces an inflammatory response in vitro, as demonstrated by the translocation of the p65 subunit of NFĸB and the release of TNFα into the surrounding medium, as well as an in vivo response. However, to put the qPCR data from SNCA-injected animals into context, a second group of mice received a single intranigral injection of LPS and were allowed to survive for 24 hours prior to tissue processing. The expression levels of proinflammatory mRNA were consistently increased for LPS-injected mice compared to SNCA-injected mice, suggesting that the LPS challenges used to model PD are unlikely to be pathophysiologically relevant. However, when the animals were injected intracerebrally with SNCA and subsequently challenged with systemic LPS the level of production of IL-1β in the SNpc was comparable to that induced by the direct injection of LPS into the brain. The injection of albumin into the SNpc with a peripheral LPS challenge did not provoke the production of IL-1β. Thus, the inflammatory response to eSNCA is context sensitive and suggests the contribution of inflammation induced by eSNCA to the progression of PD is greatly augmented by peripheral immune activation. The significance of these data is discussed below.
NFĸB exists as a heterodimer of p65 and p50, which, in the resting state, has a cytoplasmic distribution. The activation of macrophages by proinflammatory cytokines and by scavenger receptors causes the NFĸB heterodimer to dissociate from its IkB inhibitory accessory subunit and translocate to the nucleus [
16]. Our initial findings show that both wt SNCA and wt amyloid-β induce NFĸB-p65-subunit translocation to the nucleus. Others have isolated cytosolic and nuclear fractions of murine microglia and demonstrated the translocation of p65 after one hour of treatment with nitrated SNCA, but no comparison with amyloid-β was performed [
17]. While SNCA and amyloid-β both caused translocation of p65, downstream TNF production in response to these challenges was quite different. SNCA gave rise to a rapid and sustained TNF response in BV2 microglial cells, which exceeded the response generated by LPS. Amyloid-β-induced TNF production was trivial by comparison. These results suggest that LPS, SNCA, and amyloid-β activate BV2 cells in a distinct manner. It is also possible that the amyloid-β is cleared more rapidly than SNCA by the BV2 cells. Indeed, results indicate that amyloid-β is rapidly produced and cleared from the CNS in humans and that about 8% of extracellular amyloid-β is turned over each hour [
18]. Thus microglial cells are likely to be constantly exposed to amyloid-β, and eSNCA exposure is likely to be more rare. The exposed residues of peptides such as SNCA and amyloid-β may also alter the inflammatory response. Work with the shorter 25-35-residue fragment of amyloid-β is often undertaken because it produces exaggerated toxicological properties when compared to the full-length peptide [
19‐
21]. Indeed, the treatment of astrocytes with full-length amyloid-β has been reported to decrease TNF production when they are co-cultured with monocytes [
22]. Reynolds, working on nitrated SNCA, also reported TNF release within the same range as we found here [
23]. However, Zhang and colleagues failed to show TNF production by primary mixed neuron-glial cultures with full-length amyloid-β [
5]. This might be attributable to differences in the aggregation process. However, together these results highlight the difficulty associated with the interpretation of
in vitro experiments.
In vivo, modest increases in the mRNA of the transcription factor NFĸB were noted in both SNCA- and LPS-treated animals. This was largely expected of a constitutively active, translocating protein. However, it is known that NFĸB proteins in monocytes alter significantly following cellular maturation in culture and we were interested to note whether any of the responses to the
in vivo challenges would be underpinned by an increase in the overall level of NFĸB signal transduction machinery [
24]. The level of NFĸB mRNA was not significantly increased following the peripheral administration of LPS. However, the mechanistic principle of NFĸB activation is nuclear translocation rather than an increase in expression and
in vivo experiments have shown that p65 translocation to the nucleus is increased in SNCA-treated, and LPS challenged animals.
A 1.25-fold increase in the level of TNF mRNA was observed after SNCA treatment, which is less than that observed in PD brains [
17], but similar to that observed in SNCA over-expressing mice [
25]. This is likely to represent differences in the length of disease progression and may also be a consequence of the altered reactivity of the aged brain [
26] and the presence of environmental challenges or co-morbidity [
27]. IL-1β mRNA was increased 5-fold, an increase similar to that measured in SNCA over-expressing mice [
25], but considerably smaller than the 1000-fold increase generated by LPS in the SNpc. High levels of IL-1β are a common feature of peripheral inflammatory disease where cell death is a prominent feature [
28], but levels of IL-1β are low in the degenerating brain [
29]. The principle source of IL-1 in the naïve brain is the microglial cell where it contributes, among others, to the regulation of sleep cycles, appetite and temperature control. IL-6 is also implicated in a number of normal physiological processes within the CNS, including temperature regulation. It is perhaps, therefore, no surprise that IL-1β and IL-6 expression in response to SNCA is limited given the important role of these cytokines in homeostatic processes in the normal brain. However, it is clear that it is possible to overcome any resistance to cytokine production in circumstances where microglial cells have become 'primed' and are then faced with environmental challenges. Here, host survival mechanisms may take precedence over the potential damaging consequences of increased cytokine production in the brain. TNF levels were significantly increased in the ipsilateral and the contralateral hemispheres following the systemic LPS challenge. IL-1β and IL-6 were also increased in the contralateral hemisphere as well as the ipsilateral, but only in those animals that had been injected with SNCA. We have previously noted that distinct cytokine induction pathways do exist in the CNS and that the cytokine network that is often described in the periphery cannot be extrapolated to the brain [
30]. The presence of 'mirror' lesions in the contralateral hemisphere of MS patients has been recognised for a long time [
31], and, while it is often stated that MS lesions localize in a random fashion, more careful analysis reveals a distinct symmetry when small plaques and all demyelinating activity are taken into account. The signal for contralateral cytokine induction is unknown and it is unclear whether such cross-talk might contribute to PD pathology.
SNCA injection resulted in an increase in TGFβ and COX-2 mRNA but, unlike the proinflammatory cytokines, the levels were not increased by the same order of magnitude after central LPS injection. The values obtained are similar to those reported by others after hippocampal LPS injection [
32,
33]. TGFβ has been shown to promote neuronal survival in neonatal animals [
34] and is often described as an anti-inflammatory cytokine. The pattern observed in this study is similar that observed in murine prion disease [
29,
32,
33]. TGFβ and COX-2 are associated with a non-inflammatory microglial response to neurodegeneration that presumably reflects a phagocytic phenotype, similar to an M2 macrophage, rather than a classically activated M1-like macrophage. The induction of IL-1β and TNF in the SNCA-injected brain, albeit small, suggests that a mixture of phagocytic and inflammatory microglia is generated [
35]. Following the central injection of LPS the inflammatory cytokine response dwarfed the observed elevation in TGFβ and COX, and reveals the potential for cytokine production in the brain. It is interesting to note that TGFβ was induced by the combination of peripheral LPS with SNCA to a similar extent as central LPS, but COX-2 was unchanged by LPS with SNCA, which reflects dissociation between the regulatory pathways of these two 'anti-inflammatory' mediators. It should also be noted that SNCA injection did not result in significant neuronal cell death, unlike central LPS injection.
Administration of SNCA provoked a significant inflammatory profile in BV2 cells
in vitro, which was mimicked
in vivo with the expression of mRNA for the major proinflammatory cytokines. However, compared the central administration of LPS to the SNpc, cytokine production was only a small fraction of the potential production. Recently LPS has been used as a model of inflammatory PD and while it shows degeneration of tyrosine hydroxylase-positive neurons within the SN, this response is clearly distinct from the microglial response to eSNCA. Initially, such results suggested that the injection of LPS into the SNpc is not a relevant model of PD. However, the peripheral injection of LPS did induce a marked increase in inflammatory activity in the SNCA-injected SNpc. The LPS injections into the SNpc may, therefore, be modeling the consequences of a systemic infection on the degenerating SN. Recent work by Gao and colleagues has substantiated this finding by using a two-hit approach, systemic inflammation in a transgenic model of PD [
36]. Transgenic models remain important for studying the long term effects of gene-environment interactions, however, as the authors describe, there are often discrepancies between the progression of the disease in humans and that which occurs in the animal models [
37,
38]. We sought to study the immediate interactions between SNCA and the immune system and our findings are similar to recent studies in which systemic LPS injections in models of either brain injury or chronic neurodegenerative disease exacerbate brain damage and modify the local inflammatory response [
39‐
42]. For example, a link between MS relapse and infection has been proposed [
43]. A causative link between an infective agent and PD has also been sought for some time, but the evidence has been unconvincing. More recently, proponents of the infective hypothesis for idiopathic PD have suggested that certain specific pathogens, such as
H. Pylori, will be contributory rather than causative, which seems more likely [
44]. However, following the studies presented here, it is now our hypothesis that any activator of the innate immune system, including the presence of a pathogen in the periphery, will have the capacity to alter the pathogenesis of PD by generating a transient proinflammatory, destructive microenvironment that is likely to accelerate local neurodegeneration.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
YC and LA-E conducted the experiments and drafted the manuscript. NRS and MJAW assisted with revision of the manuscript. DCA provided intellectual input into experimental design and drafted the manuscript. All authors have read and approved the final manuscript.