Background
Microglia are a dynamic immune cell population of the central nervous system (CNS) [
1‐
3]. They are involved in chemotaxis, phagocytosis, and proinflammatory cytokine secretion [
4,
5] as components of their surveillance function. A number of chronic neurodegenerative diseases, including Parkinson's disease (PD), Alzheimer's disease, and multiple sclerosis display an apparently aberrant microglial behavior that is hypothesized to contribute to disease progression [
6‐
8]. Specifically, microglia appear to have a chronically activated phenotype exemplified by increased levels of various proinflammatory markers as well as elevated cytokine secretion.
It is interesting to note that PD brains have been characterized by progressive loss of dopaminergic neurons in the substantia nigra par compacta (SNpc) [
9,
10], a region with the reportedly highest density of brain microglia [
11]. It is, therefore, not surprising that increased numbers of reactive microglia in the substantia nigra are characteristic of disease and reactive microglia numbers expand to other brain regions during progressive neuron loss and disease [
12,
13].
To explore the possibility that microglial activation plays a causative role in the proinflammatory and neurodegenerative changes observed in PD, we elected to model a familial form of disease which results from over-expression of wild type or mutant α-synuclein [
14‐
16]. α-Synuclein is a 140 amino acid protein that is highly expressed in the central nervous system immuno-localizing to presynaptic terminals of neurons [
17‐
19] as well as glia and macrophage [
20‐
24]. α-Synuclein reportedly functions in regulating synaptic vesicle pools [
18], interacts with a variety of proteins [
25‐
27], and regulates lipid metabolism [
28,
29]. We have also demonstrated that α-synuclein expression regulates the behavior of microglia [
30]. A reactive microglial phenotype was increased in α-synuclein knock-out compared to wild type microglia [
30]. However, whether over-expression of wild type or mutant forms of α-synuclein may also regulate microglial phenotype remains unclear.
In order to characterize the behavior of microglia that over-express wild type or mutant α-synuclein, the mouse microglial cell line, BV2, was transiently transfected to express either human wild type (WT), A30P, or A53T mutant α-synuclein to assess the impact of intracellular over-expression on microglial behavior, rather than phenotype changes due to stimulation with extracellular α-synuclein. This study offers insight into varied mechanisms in which α-synuclein may contribute to phenotype changes in microglia during disease.
Methods
Materials
The anti-α synuclein antibody was obtained from Covance (Emerryville, CA). The anti-Cox-2, anti-LAMP-1, anti-actin and anti-GAPDH antibodies were purchased from Santa Cruz Biotechnology. Anti-PLD1 and PLD2 antibodies were purchased from Abcam (Cambridge, MA). The anti-Cox-1 antibody was purchased from Cayman (Ann Arbor, MI). The anti-MAP2 antibody was from Sigma (St Louis, MO). Anti-mouse, anti-rabbit and anti-goat horseradish peroxidase conjugated secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Lipopolysaccharide (LPS) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). FITC-labeled Escherichia coli (K-12 strain) Bioparticles were purchased from Molecular Probes (Eugene, OR). The LDH assay kit was obtained from Promega (Madison, WI).
Microglial Culture
The BV2 immortalized microglial cell line [
31] was obtained from Dr. Gary E. Landreth (Cleveland, OH). BV2 cells were grown in Dulbecco's modified Eagle's medium:Nutrient Mixture F-12 Ham (DMEM/F-12) (Gibco RBL, Rockville, MD) supplemented with 5% horse serum (Equitech-Bio, Inc., Kerrville, TX) and 10% fetal bovine serum (U.S. Biotechnologies Inc., Parkerford, PA) and 1.5 μg/mL penicillin/streptomycin/neomycin in a humidified atmosphere of 5% CO
2 and 95% air at 37°C.
Transient Transfection
BV2 cells were transiently transfected with constructs (parent construct pcDNA3.1) containing cDNAs coding for human WT, A30P, or A53T α-synuclein (1 × 106 cells, 2 μg DNA per transfection) using an Amaxa Mouse Macrophage Nucleofection Kit (Lonza Group Ltd, Switzerland) according to the manufacturer's protocol. Constructs were generously provided by Dr. Nelson Cole (NIH). Transfected cells were plated at 1 × 106 cells/condition in serum containing DMEM/F12 and harvested after 48 hours post-transfection.
Neuron Culture
Primary cortical neuron cultures were generated as previously described from cortices of embryonic day 16 C57BL/6 mice [
32]. Meninges-free cortices were isolated, trypsinized and plated onto poly-L-lysine-coated (0.05 mg/mL) tissue culture wells (260 cell/mm
2) for 7 days. The neuronal growth media was Neurobasal media supplemented with B27 and glutamine (Life Technologies, Rockville, MD, USA), which consistently provide neuronal cultures that are at least 95% pure. Culture purity was routinely evaluated by cell counting after immunostaining, to identify the neuronal cytoskeletal protein, microtubule-associated protein 2 (MAP2).
Western Blot
To perform Western blot analyses, BV2 cells were untreated, mock transfected or transfected to express WT, A30P, or A53T α-synuclein for 48 hours. At 48 hours post-transfection cells were lysed with RIPA buffer, sonicated, and centrifuged at 14,000 RPM, 4°C for 10 minutes. Protein concentrations were quantitated using the method of Bradford [
33]. Proteins were resolved by 10% or 15% SDS-PAGE and then transferred to PVDF membrane and Western blotted using anti-α-synuclein, anti-cPLA
2, anti-Cox-1, anti-Cox-2, anti-PLD1, anti-PLD2, anti-LAMP-1, anti-actin or anti-GAPDH antibodies, followed by incubation with horseradish peroxidase-conjugated secondary antibodies. Antibody binding was visualized using enhanced chemiluminescence (Pierce, Rockford, IL). Experiments were repeated 5 independent times. To quantify protein levels, optical density (O.D.) of protein bands were normalized against their respective loading control (GAPDH or actin) using Adobe Photoshop software (Adobe Systems, San Jose, CA). Ratios were averaged for all five experiments (± SD) for statistical analysis.
Enzyme Linked Immunosorbent Assay (ELISA)
The concentrations of secreted TNF-α and IL-6 from BV2 cultures were determined using commercially available mouse TNF-α and IL-6 colorimetric sandwich ELISA reagents purchased from R & D Systems (Minneapolis, MN). Briefly, cells were transfected and then stimulated with or without 25 ng/mL LPS (Sigma) at 48 hours post-transfection. Media was transferred to an ELISA plate and the levels of TNF-α and IL-6 were detected according to the manufacturer protocol. Experiments were performed with 8 replicates per condition and repeated three times to identify mean values (± SD).
Griess Assay
The levels of nitrite secreted from BV2 cells were detected using Griess reagent obtained from Alexis Biochemicals (San Diego, CA). Briefly, after 48 hours post-transfection, media was transferred to 96 well plates and incubated with Griess reagent for 10 minute at room temperature. The nitrite levels were read via microplate reader at 546 nm. Experiments were performed with 8 replicates per condition and repeated three times to identify mean values (± SD).
Phagocytosis Assay
Phagocytosis was quantified by measuring the uptake of a FITC-labeled bioparticle. Briefly, transfected BV2 cells, in 96 well plates, were incubated with or without FITC-labeled bioparticle (0.25 mg/mL) for 3 hours. To quench the signal from extracellular or outer plasma membrane associated bioparticle, medium was removed and the cells were rinsed with 0.25 mg/mL Trypan blue in phosphate buffer saline (PBS). Intracellular fluorescence was read via fluorescent plate reader (Bio-Tek, Winooski, Vermont) at 480 nm excitation and 520 nm emission. Experiments were performed with 8 replicates per condition and repeated a minimum of three times to determine mean values (±SD).
Lactate Dehydrogenase (LDH) Assay
LDH release was measured a using CytoTox 96 non-radioactive cytotoxicity assay kit according to the manufacturer protocol (Promega). Optical densities were measured by a microplate reader at 490 nm. Each condition was performed with a replicate of 8 and mean values (± SD) from three independent experiments were determined.
To assess the microglia-mediated neurotoxicity, neurons were co-incubated either alone or with mock transfected, WT, A30P, or A53T transfected BV2 cells for 72 hours. Neurons were plated onto 24 well plates (40,000 cells/well) and at 7 days in vitro were co-cultured with BV2 cells (4,000 cells/insert) that were plated onto cell culture inserts (0.4 μm Millicell, Millipore) in Neurobasal medium with or without 25 ng/ml LPS for 72 hours. After the 72 hour incubation, neurons were fixed in 4% paraformaldehyde and immunostained with antibody recognizing the neuronal cytoskeletal protein, microtubule-associated protein 2 (MAP2). A counting grid placed on the bottom of the wells was used to determine the number of viable neurons. Neurons from 4 independent fields/well from 8 wells per condition were counted. Neurons were counted as viable if they were MAP2 positive, had a visible nuclei and immunostained processes which were at least two times the length of the cell body. Mean values (± SD) from three independent experiments were determined.
Statistical Analysis
Mean values (± SD) for each experiment were determined and values statistically different from controls were calculated using one-way ANOVA. The Newman-Keuls multiple comparisons post-test was used to determine p-values. GraphPad Prism 4 software was used for analysis (GraphPad, San Diego, CA).
Discussion
This study demonstrated that over-expression of α-synuclein modulates the phenotype of a commonly used microglial model, BV2 cells. It is important to point out that these data are derived from a microglial cell line and the possibility exists that a more accurate prediction of microglial behavior in response to α-synuclein over-expression would emerge through the use of primary cells. For instance, microglia grown from transgenic mouse lines over-expressing mutant or wild type protein could be used in future work. Nevertheless, the current data set demonstrated a clear change in cellular behavior of microglial BV2 cells that involved an increase in Cox-2 protein levels in cells over-expressing human WT, mutant A30P and A53T α-synuclein. In addition, over-expression of A30P and A53T mutants as well as human WT α-synuclein decreased phagocytic ability of the BV2 cells while increasing their secretion of TNF-α, IL-6, and nitric oxide. However, in spite of these robust changes in behavior, the α-synuclein over-expressing BV2 cells did not demonstrate any increase in neurotoxic capacity.
Although extracellular α-synuclein in PD may be acting as one of the sources for the induction of microgliosis, our efforts were to identify a fundamental role for intracellular α-synuclein in regulating microglial phenotype in particularly the familial form of disease. Therefore, a distinction of our work from several prior reports is that we have examined effects of α-synuclein expression on microglial phenotype rather than effects of adding α-synuclein to microglia as a ligand. It is assumed that microgliosis occurs during disease in part due to neuronal secretion of α-synuclein which directly stimulates microglia in a fashion requiring CD36 [
16,
67]. Several studies demonstrate that extracellularly applied α-synuclein directly stimulates phagocytic cells such as microglia, macrophage, and monocytes to acquire a reactive phenotype characterized by a number of changes including increased production of matrix metalloproteases [
68,
69], increased Cox-2 levels [
19], increased cytokine secretion [
18], increased neurotoxin secretion [
18,
70], and increased cellular migration [
69]. Our findings indicate that expression of α-synuclein also drives microglia into a form of activation characterized by elevated proinflammatory cytokine secretion and Cox-2 levels accompanied by impaired phagocytosis that appears unique from activation characterized by extracellular α-synuclein stimulation. These findings may help to elucidate the biology of early onset disease and indicate that microgliosis occurs as not only a reaction to neuronal death and α-synuclein secretion but may also cause neuronal dysfunction through impaired homeostasis and a proinflammatory phenotype.
Although it is our expectation that the phenotype changes observed in the BV2 cells was due to effects of α-synuclein expression we cannot exclude the possibility that a portion of the translated protein is being exocytosed into the media and acting as an extracellular ligand as others have reported [
71]. It is possible that BV2 cells over-expressing α-synuclein could be secreting the protein to provide a pool for extracellular, autocrine stimulation. A prior report demonstrated that extracellular stimulation with α-synuclein compared to over-expression of α-synuclein in BV2 cells produced similar changes in cellular migration and CD44 expression [
69]. This supports the idea that some component of the phenotype change we observed may be due to autocrine, feed-forward stimulation of exocytosed α-synuclein combined with the consequences of over-expression on cellular behavior. More importantly, this suggests again that microgliosis in disease is not only a consequence of α-synuclein expression by microglia but is also modified by secreted α-synuclein that could be coming from neurons but also other cells in the brain such as microglia and astrocytes [
71,
72]. Dissecting the differences between the two types of α-synuclein-mediated activation will offer a clearer target for attenuating the microglial contributions to disease.
Histologic data from human PD brains has demonstrated increased Cox-2 immunoreactivity within both microglia and neurons of the substantia nigra suggesting that Cox-2-dependent prostaglandin production contributes to inflammatory gliosis and neuron death [
73,
74]. A role for Cox-2 dependent inflammation and cell death in disease is supported by the MPTP toxin model of PD in which Cox-2 activity is required for the observed neuronal death [
75‐
77]. Although a myriad of mechanisms are feasible, one intriguing idea is that the increased Cox-2 activity leads to dopamine oxidation and subsequent α-synuclein accumulation as Lewy bodies [
78]. We now extend this data to demonstrate that α-synuclein over-expression of particularly the missense mutant forms increased microglial Cox-2 expression.
The consequences of increased Cox-2 activity in microglia are several. For instance, Zhang et al. (2005) demonstrated that microglial stimulation with extracellular aggregated α-synuclein enhances microglial production of PGE
2 required for subsequent neurotoxin secretion [
70]. On the other hand, PGE
2 stimulation inhibits TNF-α secretion from BV2 and microglial cells [
58‐
60] and impairs phagocytic ability [
79]. Although the specific prostaglandin formation and function downstream of increased microglial Cox-2 expression or activity is far from resolved our data correlates well with an emerging theme that arachidonic acid metabolism is disrupted during disease due to a fundamental role of α-synuclein in regulating lipid metabolism [
39,
42]. In addition, other studies have demonstrated that a disruption in particular prostaglandin levels correlates with disease, α-synuclein expression, and in some cases induces disease phenotype [
39,
80‐
83]. Therefore, alterations in arachidonic acid metabolism appear central to both sporadic and familial disease across cell several types including microglia and neurons and defining the specific production patterns and consequences of individual prostaglandins will be critical in defining their role during PD.
Another interesting phenotype change produced by over-expression of α-synuclein was the attenuated ability of microglia to phagocytose the
E. coli bioparticles. This correlated with a significant decrease in LAMP-1 levels suggesting lysosomal dysfunction, at least, is a component of the uptake problem. It is clear that α-synuclein has a role in modulating vesicular trafficking in other cells types [
84,
85] so it is not unreasonable that a similar regulatory role exists in phagocytic cells such as microglia. For example,
in vitro studies demonstrate that microglia are capable of taking up α-synuclein, in particular its monomeric form, in what appears to be a classic-clathrin dependent mechanism [
5,
86]. Indeed the monomeric protein facilitates overall microglial phagocytic ability, while the aggregate form attenuates phagocytosis [
5]. It is possible that α-synuclein expression by microglia actually attenuates the ability of these cells to take up aggregate α-synuclein thus contributing to disease.
Another interesting possibility is that the decrease in phagocytic ability by α-synuclein over-expressing cells is due to an increase in Cox-2 mediated prostaglandin formation. Extracellular α-synuclein aggregates stimulate microglia
in vitro and attenuate their ability to phagocytose the aggregates in a fashion requiring PGE
2 stimulation of its EP2 receptors [
79]. In addition, PGE
2 stimulation of its EP2 receptor downregulates microglial ability to take up another aggregate protein, beta amyloid [
53,
54]. This collectively supports the idea that specific prostaglandin stimulations modulate microglial behavior. Therefore, over-expression of α-synuclein may impair the general homeostatic role of microglia as brain phagocytes and while certainly of relevance to PD this implicates a broader role for this protein in how microglia function in the brain.
It is somewhat surprising that the α-synuclein over-expressing cells did not demonstrate increased neurotoxic capacity in our co-culture paradigm with and without LPS stimulation. One possibility is that altered culture conditions including different cellular ratios or incubation times as well as the use of primary microglia instead of BV2 cells may produce different results in neuronal or synaptic viability. However, the fact that cells over-expressing α-synuclein demonstrated significant changes in secretory phenotype without a correlating change in neurotoxic capacity in the conditions tested is still of importance. For example, there is certainly
in vivo evidence that microgliosis can occur as a consequence of α-synuclein expression that involves proinflammatory change without robust neuron death. Specifically, over-expression of α-synuclein results in an early increase in microgliosis prior to neuron death in some rodent models of disease [
16,
67]. Another study has shown using an adeno-associated virus model to over-express α-synuclein in mice that a robust increase in microgliosis as well as T and B cell infiltration occurs in the absence of any robust neurodegeneration [
87]. Others report that the reactive microglial phenotype varies in response to neuronal α-synuclein expression depending upon whether or not neurons are dying [
88]. Collectively, it appears that microgliosis
in vitro and
in vivo may be heterogeneous in response to α-synuclein expression.
For instance microgliosis during disease may be a consequence of extracellular α-synuclein stimulation as a distinct and additional mechanism of activation than that induced by over-expression of α-synuclein. This suggests that a heterogeneous range of microgliosis phenotypes exist during disease and across brain regions that have yet to be fully described in which each type of activation may contribute differently to disease. Extracellular stimulation of microglia with exocytosed α-synuclein may be responsible for a form of gliosis while direct effects on microglial phenotype due to expression of α-synuclein may produce a similar yet unique phenotype. Future efforts examining primary microglia rather than BV2 cells over-expressing α-synuclein in neuron microglia co-culture using different cellular ratios, neuronal populations, or time points may provide a clearer picture of changes in neuron viability or, more importantly, synaptic integrity. It will be important to determine whether a temporal effect of either form of gliotic stimulation, α-synuclein over-expression or extracellular stimulation, occurs during disease and whether or not they provide combined or singular insults to the neuronal death that occurs.
Authors' contributions
L.R. was responsible for conducting all experiments, interpreting data, and writing the initial version of the manuscript.
E.J.M. was involved in data interpretation and revising the final version of the manuscript.
C.K.C. was involved in overall experiment design, data interpretation, and revising the final version of the manuscript.
All authors have read and approved of the final version of the manuscript.
Competing interests
The authors declare that they have no competing interests.