J. Neurochem. (2008) 104, 1504–1525.
Microglial neuroinflammatory processes play a primary role in dopaminergic neurodegeneration for Parkinson’s disease (PD). This can occur, in part, by modulation of glial function following activation by soluble or insoluble modified alpha-synuclein (α-syn), a chief component of Lewy bodies that is released from affected dopaminergic neurons. α-Syn is nitrated during oxidative stress responses and in its aggregated form, induces inflammatory microglial functions. Elucidation of these microglial function changes in PD could lead to new insights into disease mechanisms. To this end, PD-associated inflammation was modeled by stimulation of microglia with aggregated and nitrated α-syn. These activated microglia were ameboid in morphology and elicited dopaminergic neurotoxicity. A profile of nitrated, aggregated α-syn-stimulated microglia was generated using combinations of genomic (microarrays) and proteomic (liquid chromatography-tandem mass spectrometry, differential gel electrophoresis, and protein array) assays. Genomic studies revealed a substantive role for nuclear factor-kappa B transcriptional activation. Qualitative changes in the microglial proteome showed robust increases in inflammatory, redox, enzyme, and cytoskeletal proteins supporting the genomic tests. Autopsy brain tissue acquired from substantia nigra and basal ganglia of PD patients demonstrated that parallel nuclear factor-kappa B-related inflammatory processes were, in part, active during human disease. Taken together, the transcriptome and proteome of nitrated α-syn activated microglia, shown herein, provide new potential insights into disease mechanisms.
atomic force microscopy
differential gel electrophoresis
Dulbecco’s modified Eagle’s medium
fetal bovine serum
liquid chromatography-tandem mass spectrometry
nuclear factor-kappa B
nitrated alpha synuclein
reactive oxygen species
substantia nigra pars compacta
Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by resting tremor, rigidity, bradykinesia, and gait disturbances (Fahn et al. 1998; Mayeux 2003; Fahn and Sulzer 2004). Presently, 1.5 million Americans are afflicted. Disease incidence rises with increasing age, with 120/100 000 contracting PD over the age of 70 (Dauer and Przedborski 2003). Pathologically, PD is characterized by the progressive loss of dopaminergic neuronal cell bodies in the substantia nigra pars compacta (SNpc) and their termini in the dorsal striatum (Hornykiewicz and Kish 1987). These pathological findings commonly parallel microglial activation observed in association with deposits of aggregated alpha synuclein (α-syn) in intracellular inclusions, known as Lewy bodies (LB) (Spillantini et al. 1997; Croisier et al. 2005). Although host genetics and environmental factors affect the onset and progression of PD (Tanner 1992) significant clinical, epidemiologic, and experimental data also support a role for microglial inflammation in disease pathogenesis (Forno et al. 1992; Banati and Blunt 1998;McGeer and McGeer 1998; Mirza et al. 2000;Cicchetti et al. 2002; Block and Hong 2005; Hong 2005; Wang et al. 2005).
The mechanisms underlying microglial activation in PD and how it affects neuronal survival is incompletely understood. One line of investigation is that neuronal death itself drives microglial immune responses (Giasson et al. 2000; Przedborski et al. 2001; Mandel et al. 2005). Alternatively, we, as well as others, have proposed that activation occurs as a consequence of release of aggregated proteins from the cytosol or within LB to the extracellular space. In this way, the death of dopaminergic neurons leads to release of modified protein aggregates that activate microglia inciting a lethal cascade of neuroinflammation and neuronal demise (Zhang et al. 2005; Wersinger and Sidhu 2006). Several lines of experimental evidence support this contention (Spillantini et al. 1997; Goedert 1999; Giasson et al. 2000; Kakimura et al. 2001; Croisier et al. 2005; Lee et al. 2005). First, aberrant expression of α-syn and PD pathogenesis are linked. This is derived from the discovery that mutations and multiple copies of the gene encoding α-syn (SNCA and PARK1) are linked to familial early onset PD (Kruger et al. 1998; Spira et al. 2001; Zarranz et al. 2004; (Singleton et al. 2003; Chartier-Harlin et al. 2004). Second, oxidation and nitration of α-syn leads to formation of aggregates and filaments that comprise LB (Giasson et al. 2000; Souza et al. 2000). Third, portions of α-syn are secreted rendering it more vulnerable to aggregation (Lee et al. 2005) and oxidative modification (Kakimura et al. 2001). Fourth, α-syn itself can activate microglia, inducing reactive oxygen species (ROS) (Thomas et al. 2007) and subsequent neurotoxicity (Zhang et al. 2005). Fifth, microglial products including cytokines, chemokines, excitotoxins, and proteins of the classical complement cascade affect a broad range of neurological diseases (McGeer and McGeer 1998; Bal-Price and Brown 2001; Liu and Hong 2003; Block and Hong 2005). Sixth, endogenous activators of microglia show a neuroinflammatory fingerprint reflective of what can occur during PD (Zhou et al. 2005; McLaughlin et al. 2006). Lastly, attenuation of microglial activation can protect up to 90% of dopaminergic neurons in PD animal models (Du et al. 2001; Teismann and Ferger 2001; Wu et al. 2002; Teismann et al. 2003; Kurkowska-Jastrzebska et al. 2004; Choi et al. 2005; Vijitruth et al. 2006).
Based on these observations, we investigated changes in the microglial transcriptome and proteome as a consequence of the cells’ engagement with nitrated and aggregated α-syn (N-α-syn). N-α-syn stimulation of microglia induced morphological cell transformation and neurotoxic secretions. A N-α-syn-activated ‘microglial signature’ was determined by gene microarrays, 2D differential in-gel electrophoresis (DIGE), and by cytokine profiling. N-α-syn induced a microglia inflammatory phenotype characterized by the expression of neurotoxic and neuroregulatory factors. Most importantly, the inflammatory signature seen in laboratory assays were, in part, mirrored in parallel tests performed on postmortem brain tissues from PD patients. These observations, taken together, serve to support both a ‘putative’ role for N-α-syn-activated microglia in disease.
Materials and methods
Parkinson’s disease brain tissues
Autopsy materials from the substantia nigra (SN) and basal ganglia (BG; caudate nucleus and putamen) of 10 patients who died with signs and symptoms of PD, three with Alzheimer’s disease (AD), and 10 age-matched controls were secured from the National Research Brain Bank Tissue Consortium. The 10 controls ranged in age from 62 to 91 and died of diseases unrelated to neurological impairments. This included atherosclerotic and metabolic diseases, infections, and cancer (Table 1).
|1||PD||Female||78||25||Hypopigmentation and LB formation in SN and locus ceruleus|
|2||PSP||Male||83||N/A||Cerebral atrophy of frontal and superior temporal|
|6||LBD||Male||82||8||Neocortical LB and hippocampal sclerosis|
|7||PSP||Male||82||24||Degeneration of subcortical nuclei with neurofibrillary tangles|
|8||PD||Male||67||17||Hypopigmentation, neuronal loss, gliosis, LB in SN, and locus ceruleus|
|9||PD||Female||85||3||Hypopigmentation, gliosis, and neuronal loss in SN, globus pallidus, and caudate-putamen|
|10||PD||Male||86||5||Hypopigmentation, LB, neuronal loss in the SN and locus ceruleus|
|11||AD||Male||73||8||Neuritic and diffuse plaques and neurofibrillary tangles in neocortex and hippocampus. Neuronal loss in Locus ceruleus.|
|12||AD||Female||84||N/A||Remote infarct in the occipital lobe and cerebellum|
|13||AD||Male||76||N/A||Senile plaque-predominant AD|
An antibody to N-α/β-syn (clone nSyn12, mouse ascites; Upstate, Charlottesville, VA, USA) that recognizes nitrated human N-α-syn (14.5 kDa) and N-β-syn (17 kDa) was used for immunoprecipitation. Samples of SN from control, AD, and PD autopsy brain tissues (Table 1) were homogenized in ice-cold radioimmunoprecipitation (RIPA) buffer, pH 7.4 and centrifuged at 10 000 g for 10 min at 4°C to remove cellular debris. Protein A/G PLUS-agarose beads (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) were added to 1 mg total cellular protein and incubated for 30 min at 4°C. Beads were pelleted by centrifugation at 1000 g for 5 min at 4°C. The supernatants were incubated overnight at 4°C with 2 μg of primary antibody, then with Protein A/G PLUS-agarose beads for 6 h on a rotating device at 4°C. Immunoprecipitates (IP) were collected after centrifugation at 1000 g for 5 min at 4°C, washed with phosphate-buffered saline (PBS), and resuspended in 20 μL of 1x electrophoresis sample buffer.
Nitrated-α-Syn IP (20 μL) were fractionated by 16% Tricine sodium dodecyl sulfate–polyacrylamide gel electrophoresis (PAGE) (Jule Inc., Milford, CT, USA and BIORAD Laboratories Inc., Los Angeles, CA, USA) at constant voltage for 1.5 h. The gels were fixed and stained with Coomasie Blue to visualize protein bands. Bands corresponding to the molecular weights encompassing 14.5 kDa (α-syn) were excised, digested by trypsin, column purified, and sequenced by liquid chromatography-tandem mass spectrometry (LC-MS/MS) for protein validation. Sequenced peptides were distinguished by peptide matches to the human α-syn sequence (NCBI Accession: AAI08276).
Purification, nitration, and aggregation of recombinant mouse α-syn
Purification, nitration, and aggregation of recombinant mouse α-syn were performed as previously described (Thomas et al. 2007). Five individual lots of α-syn were tested for endotoxin by Limulus amebocyte lysate tests and all were below the limit of detection for endotoxin (< 0.05 endotoxin units). Amino acid analysis to determine protein concentration using HPLC was performed by the University of Nebraska Medical Center Protein Structure Core Facility. Proteins were separated by PAGE using 4–12% NuPAGE gels (Invitrogen, Carlsbad, CA, USA). After electrophoresis, the gels were transferred onto polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA) and probed with primary mouse IgG1 anti-α-syn (1 : 500; Transduction Laboratories/BD Biosciences, Franklin Lakes, NJ, USA) or primary rabbit IgG anti-nitrotyrosine (1 : 2000; Upstate). Signal was detected with horseradish peroxidase-conjugated anti-mouse IgG or anti-rabbit IgG (both from Zymed Laboratories, South San Francisco, CA, USA) using chemiluminescence systems (SuperSignal® West Pico Chemiluminescent Substrate; Pierce Biotechnology Inc., Rockford, IL, USA). For visualization of the protein by atomic force microscopy (AFM), samples were deposited on mica, glued to a glass slide, and dried under argon gas flow. The image was taken in air, height, amplitude, and phase modes using a Molecular Force Probe 3D controller (Asylum Research Inc., Santa Barbara, CA, USA).
Isolation, cultivation, and N-α-syn activation of murine microglia
Microglia from C57BL/6 mice neonates (1- to-2-days old) were prepared according to well described techniques (Dobrenis 1998). All animal procedures were in accordance with National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center. Brains were removed and placed in Hanks’ Balanced Salt Solution at 4°C. The mixed glial cells were cultured for 7 days in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal bovine serum (FBS), 10 μg/mL gentamicin, and 2 μg/mL macrophage colony stimulating factor (a generous gift of Wyeth Inc., Cambridge, MA, USA). To obtain homogenous microglial cell populations, culture flasks were gently shaken and non-adherent microglia were transferred to new flasks. The flasks were incubated for 30 min to allow the microglia to adhere, and loose cells removed by washing with DMEM. Microglia were plated at 2 × 106 cells per well in six-well plates in DMEM containing 10% FBS, 10 μg/mL gentamicin, and 2 μg/mL macrophage colony stimulating factor. One week later, cells were stimulated with 100 nmol/L of aggregated N-α-syn/well or no stimulation for 4 h. Media were replaced with serum-free DMEM without phenol red or other additives (Invitrogen) and incubated for 24 h in a 37°C, 5% CO2 incubator.
Inflammatory genomic and PCR assays
RNA from N-α-syn stimulated primary murine microglial cells and unstimulated control was extracted with TRIzol (Invitrogen), column purified (Qiagen, Valencia, CA, USA), precipitated with ammonium acetate, amplified and labeled using the T7-based TrueLabeling-AMP 2.0 kit (Superarray, Frederick, MD, USA). The resultant cRNA was hybridized to an oligo-based microarray for mouse general pathway (Superarray #OMM-014) and nuclear factor-kappa B (NF-κB)-related genes (Superarray #OMM-025). The arrays were washed, incubated sequentially with streptavidin-bound alkaline phosphatase and chemiluminescent substrate before exposure to X-ray film. Subsequent analysis of the microarrays was performed using the GEArray expression analysis suite (Superarray).
Total RNA obtained from analysis of the microglial transcriptome was reverse transcribed with random hexamers and SSII reverse transcriptase (Applied Biosystems, Foster City, CA, USA). Murine-specific primer pairs were: Ccl2: CCCCAAGAAGGAATGGGTCC and GGTTGTGGAAAAGGTAGTGG; Il1β: GTTCCTTTGTGGCACTTGGT and CTATGCTGCCTGCTCTTACTGACT; Il10: CAGTTATTGTCTTCCCGGCTGTA and CTATGCTGCCTGCTCTTACTGACT; Ifng: TTTGAGGTCAACAACAACCCACA and CGCAATCACAGTCTTGGCTA; and Nos2: 5′-GGCAGCCTGTGAGACCTTTG-3′ and 5′-GAAGCGTTTCGGGATCTGAA-3′. TaqMan gene expression assays specific for murine Tnf, Tnfrsf1a, Stat1, Rela, Bdnf, and Gdnf were purchased from Applied Biosystems, and normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) expression. Tissue samples obtained from PD and control patients were snap frozen on dry ice and stored at −80°C. RNA was prepared from the samples using TRIzol reagent (Invitrogen) and purified with the RNeasy Mini Kit (Qiagen), prior to cDNA synthesis. Human specific primers for TNF, TNFRSF1A, STAT1, NFKB1, RELA, BDNF, and GDNF were analyzed using TaqMan gene expression assays. Gene expression was normalized to the housekeeping gene Gapdh. Real-time quantitative PCR was performed with cDNA using an ABI PRISM 7000 sequence detector (Applied Biosystems). Reverse SYBR Green I detection system was used, and the reactions generated a melting temperature dissociation curve enabling quantitation of the PCR products.
Microglia were plated at a density of 2 × 106 cells/well in a six-well plate and stimulated with 100 nmol/L aggregated N-α-syn, and 100 ng/mL lipopolysaccharide (LPS, Escherichia coli; Sigma-Aldrich, St. Louis, MO, USA) in serum-free media. Fifty microliter aliquots were collected at 8, 24, and 72 h of incubation in triplicate and frozen at −80°C. For assay, the samples were analyzed using the BD Cytometric Bead Array Mouse Inflammation Kit (BD Biosciences, San Jose, CA, USA) according to the manufacturer’s protocol. Samples of culture supernatants from microglia were diluted 1 : 3 and 1 : 10 in assay diluent and analyzed for cytokine concentration with a FACSCalibur flow cytometer (BD Biosciences). Concentrations of cytokines were determined from a standard curve created with serial dilutions of the cytokine standards provided by the manufacturer.
MES23.5 cells, kindly provided by Dr Stanley Appel, were cultured in 75-cm2 flasks in DMEM/F12 with 15 mmol/L HEPES (Invitrogen) containing N2 supplement (Invitrogen), 100 U/mL of penicillin, 100 μg/mL streptomycin, and 5% FBS. Cells were grown to 80% confluence then co-cultured at 1 : 1 ratio with previously plated microglial cells. To assess cell viability microglia cells were plated at a density of 5 × 104 cells on sterile glass coverslips, and co-cultures were prepared with a 1 : 1 ratio microglia: MES23.5 cells. After 24–48 h, cells were stimulated with aggregated 100 nmol/L N-α-syn or 100 nmol/L α-syn for 4, 8, 24, and 72 h. CD11b+ microglial cells were distinguished from MES23.5 cells by APC-conjugated CD11b (1 : 200; Invitrogen) immunocytochemistry. For tyrosine hydroxylase (TH) cytostaining, cells were fixed in 4%p-formaldeyde, permeablized, and blocked in 2% normal goat serum with 0.25% Triton X-100 in PBS for 30 min, and probed with rabbit polyclonal anti-TH (1 : 1000; EMD Biosciences Inc., San Diego, CA, USA), followed by FITC goat anti-rabbit IgG. For western blot analysis, 10 μg of protein sample from cell lysates of each treatment group was loaded onto a 12% NuPAGE Bis–Tris gel (Invitrogen). Following transfer onto a PVDF membrane, the membrane was blocked and then probed overnight with anti-TH (1 : 1000). Signal was detected with horseradish peroxidase-conjugated anti-rabbit IgG (1 : 10 000; Zymed Laboratories) using chemiluminescence system (SuperSignal® West Pico Chemiluminescent substrate; Pierce Biotechnology Inc.). Densitometric analysis was performed using ImageJ software and normalized to β-actin (1 : 1000; Abcam, Cambridge, MA, USA). Assays for viable and dead mammalian cells (Live/Dead Viability/Cytotoxicity; Invitrogen) were performed according to manufacturer’s protocol. The protocol was revised so that the concentration of each dye was 1 μmol/L to avoid high background. Live cells were distinguished by the uptake of calcein acetoxymethyl ester to acquire a green fluorescence [excitation/emission (ex/em) 495/515 nm], while dead cells acquired a red fluorescence (ex/em 495/635 nm) because of the uptake of ethidium homodimer-1. Cell enumerations were performed using fluorescence microscopy (200× magnification) and a M5 microplate fluorometer (Molecular Devices, Sunnyvale, CA, USA) (Limit 1 ex/em 490/522 nm and Limit 2 ex/em 530/645 nm). The number of viable MES23.5 cells in each treatment group was normalized as the percentage of surviving cells in unstimulated culture controls.
Protein purification, 2D DIGE, and DeCyder analyses
Cell lysates of microglia were prepared with 5 mmol/L Tris–HCl, pH 8.0, 1% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propane sulfonate and a cocktail of protease inhibitors (Sigma-Aldrich). Protein content was quantitated using a DC Protein Assay (BioRad, Hercules, CA, USA). Factors known to interfere with isoelectric focusing (first dimension separation in 2D sodium dodecyl sulfate–PAGE) such as salts and detergents were removed from cell lysates using the 2D Cleanup kit (GE Healthcare, Piscataway, NJ, USA) according to manufacturer’s protocol. Protein concentration was determined using 2D Quant (GE Healthcare). Samples of control and stimulated cell lysates (25 μg of each lysate) were labeled with 400 pmol of CyDye 2. A 50 μg protein sample of control cell lysate was labeled with 400 pmol of CyDye 3; and a 50 μg protein sample of stimulated cell lysate was labeled with 400 pmol of CyDye 5. Labeling was performed following the manufacturer’s protocols. The samples were pooled, resuspended in rehydration buffer to a total volume of 450 μL, then loaded onto an immobilized pH gradient strip, and left for 18 h for rehydration. In the first dimension, samples were run in IPGphor and in Ettan DALTsix electrophoresis apparatus (GE Healthcare) for the second dimension. CyDye 3- and CyDye 5-derivatized proteins were detected in gels using a Typhoon 9400 Variable Mode Imager with ex–em filters at 540/590 nm for CyDye 3 dyes and 620/680 nm for CyDye 5 dyes (GE Healthcare). Analysis of CyDye 3-CyDye 5 image pairs, adjustment to CyDye 2 control images, and detection of protein spots with relative spot volumes were performed using DeCyder software (GE Healthcare) to locate and analyze multiplexed samples within the gel. Selected protein spots of interest were excised from the 2D gel using an Ettan Spot Picker. The proteins from gel pieces were digested with trypsin, as described below, and resultant peptides were analyzed using LC-MS/MS system (ThermoElectron Inc., Waltham, MA, USA). Protein identification was completed using BioWorks 3.1 software.
In gel tryptic digestion and protein identification by LC-MS/MS
Specific protein spots were excised from the gels by an automated Ettan spot picker. Following column purification (ZipTip CU-18; Millipore) with 50% acetonitrile (ACN), 50 mmol/L NH4HCO3/50% ACN, and 10 mmol/L NH4HCO3/50% ACN, gel pieces were dried and incubated with trypsin (100 ng/μL) (Promega, Sunnyvale, CA, USA) for 12–18 h. Samples were extracted by 0.1% trifluoroacetic acid/60% ACN, pooled, and dried.
Dried peptide samples were reconstituted in 0.1% formic acid/HPLC-grade water, detected on a ProteomeX LCQTM DECA XP Plus mass spectrometer (ThermoElectron Inc.), and identified using BioWorks 3.1SR software. Proteins identified by peptides having a Unified Score (BioWorks 3.1SR, ThermoElectron Inc.) greater than 3000 were marked for further analysis.
Cell lysates were prepared from SN of PD and control patients by homogenization in PBS. Cells were collected following centrifugation at 500 g for 5 min. Cytosol and nuclear fractions were prepared using the Nuclear/Cytosol Fractionation Kit (BioVision, Mountain View, CA, USA) according to manufacturer’s protocol.
Western blot assays
Protein was prepared from cell lysates in RIPA buffer supplemented with protease inhibitors (Pierce Biotechnology Inc.). Protease inhibitor cocktail was added to each conditioned media sample fraction prior to processing. Following centrifugation at 10 000 g for 10 min, the supernatants were removed and allowed to dialyze against water overnight. Tissue samples obtained from PD and control patients were snap frozen on dry ice and stored at −80°C. Protein lysates were prepared from individual samples through homogenization in RIPA buffer supplemented with protease inhibitors (Pierce Biotechnology Inc.). Protein quantification was performed using the bicinchoninic acid kit (Pierce Biotechnology Inc.). Protein concentration of each sample was determined using a calibration curve generated from purified bovine serum albumin. A total of 20 μg of each sample was loaded onto 4–12% Bis–Tris NuPAGE gels (Invitrogen) and transferred onto PVDF membranes (BioRad). Primary antibodies to calmodulin (1 : 1000) and 14-3-3σ (1 : 200) (Millipore), biliverdin reductase (1 : 5000), thioredoxin (1 : 2000), β-actin (1 : 5000), and α-tubulin (1 : 5000) purchased from Abcam, l-plastin (1 : 1000), α-enolase (1 : 1000), glutathione-S-transferase (1 : 1000), and NF-κB p65 and p50 (1 : 200) purchased from Santa Cruz Biotechnology Inc. were used for analyses. Blots were probed with the respective horseradish peroxidase-conjugated secondary antibodies (1 : 5000; Invitrogen) and detected using SuperSignal West Pico Chemiluminescent substrate (Pierce Biotechnology Inc.). The intensity of protein bands was quantified using ImageJ and normalized to Gapdh (1 : 5000; Santa Cruz Biotechnology Inc.) level in the same sample.
All values are expressed as mean ± SEM. Differences among means were analyzed by Student’s t-test or by one-way anova followed by Bonferroni post hoc testing for pair-wise comparison.
Aggregated N-α-syn and microglial activation: laboratory and pathological studies
To investigate the mechanisms by which N-α-syn-mediated microglial activation affects dopaminergic neurodegeneration, we created a cellular model that would reflect the salient features of neuroinflammation as it could occur in PD. To this end, we first determined if N-α-syn was present in regions of brain where microglial activation is known to be present in PD. Whole cell lysates consisted of several protein bands following gel electrophoresis (data not shown) and Coomassie staining (data not shown). IP assays performed from SN tissues of PD patients using a primary antibody against nitrated α/β-synulcein showed a greater than twofold increase in intensity of the protein band corresponding to 14–14.5 kDa (p < 0.001) than that present in control patients (Fig. 1a) or in patients diagnosed with AD (data not shown) along with higher molecular weight species greater than 17 kDa (data not shown). Peptide sequence analyses by LC-MS/MS revealed that the protein band encompassing the 14–14.5 kDa of the anti-N-α-syn IP contained peptides with sequence homology to human α-syn in SN samples recovered from PD brains (Fig. 1a, highlighted sequences). Interestingly, such sequence homologies to α-syn were not identified from 14 to 14.5 kDa proteins in either control or AD samples. Thus, we next sought to develop an in vitro model to reflect conditions present in an affected human host, but using the murine analog. Here, recombinant mouse α-syn was purified, nitrated, and aggregated for use as a microglial stimulant. Western blot assays showed cross-linking of N-α-syn monomers (Souza et al. 2000) and higher molecular weight aggregates, thus verifying the nitration and aggregation of α-syn (Fig. 1b). The aggregated N-α-syn contained a substantially reduced monomeric band (corresponding to a band at ∼14 kDa) but higher molecular weight banding aggregates. Analysis of protein aggregation was also assessed by AFM. Samples of N-α-syn contained low numbers of globular aggregates (2–6 nm in height) prior to aggregation. However, following aggregation, N-α-syn was present predominately as oligomers (2–6 nm in height). In addition, there were few protofibrils (1.5–2.5 nm in height), filaments, and fibrils (∼5–8 nm in height) present (Fig. 1c). Non-nitrated α-syn was present in similar configurations (data not shown).
We next evaluated the stimulatory effects of N-α-syn on microglia. The dose of 100 nmol/L (14.5 ng protein/mL) was selected based on previous extensive works performed in our laboratories demonstrating that, following a dose-response of N-α-syn, 100 nmol/L (50% over control) is required to induce substantive ROS from activated microglial cells (Zhang et al. 2005; Thomas et al. 2007) as well as cytotoxicity. ROS production was slightly decreased in comparison with either 50 or 500 nmol/L of N-α-syn. While native α-syn is ubiquitously expressed, the physiological concentration of N-α-syn in disease has not been elucidated. However, based on concentrations of modified α-syn in affected PD brain tissues, 100 nmol/L concentration is at physiologically relevant levels (Halliday et al. 2005) and is below that detectable by immunohistochemistry in neuronal inclusions within the SN of PD brains (≥ 100 ng). Phenotypic transformation into an ameboid morphology commonly follows microglial activation with different pro-inflammatory stimuli (Giulian et al. 1995; Vilhardt 2005). Thus, we examined if changes in microglial morphology would be elicited following N-α-syn activation. Resting microglia were both round and ellipsoid shaped with retracted processes that were characteristic of a relatively quiescent phenotype (Fig. 1d). In contrast, N-α-syn activated microglia assumed a more ameboid appearance with extensive processes, characteristic, in part, of an activated phenotype. N-α-syn-stimulated microglia co-cultured with MES23.5 cells acquired a rod-like appearance and further extension of processes.
We recently demonstrated that 100 nmol/L of aggregated N-α-syn could activate microglia to produce copious amounts of ROS (Thomas et al. 2007). In contrast, unaggregated N-α-syn or minced neuronal membrane fractions failed to induce significant amounts of ROS above control levels. This suggested that the microglial response to N-α-syn was specific and could not be elicited in response to unaggregated protein or by phagocytosis under the same conditions. Therefore, we assessed the extent of the neuroinflammatory phenotype induced by N-α-syn stimulation of microglia. Quantification of common cytokines and chemokines that are secreted in response to inflammatory stimuli was performed by cytometric bead array. LPS-activated microglia served as a positive control. Stimulation with N-α-syn enhanced the secretion of TNF-α, IL-6, MCP-1 (Fig. 1e), and IFN-γ (data not shown) compared with basal levels observed in unstimulated microglia. These results are consistent with the induction of an inflammatory microglial phenotype following N-α-syn stimulation. The parallels between N-α-syn and LPS-induced cellular effects support a commonality for innate immune responses in disease and suggest that these pro-inflammatory processes may be common among mononuclear phagocytes that recognize disparate activators.
N-α-syn-stimulated microglia are neurotoxic to MES23.5 dopaminergic cells
To determine the effect of N-α-syn-activated microglia on neuronal survival, the dopaminergic MES23.5 cell line was used as an indicator for cytotoxicity measurements by co-culture with stimulated and unstimulated microglia. MES23.5 cell death was determined by measuring immunoreactivity for the rate-limiting enzyme in dopamine synthesis, TH, expressed by MES23.5 cells, and the Live/Dead cell assay. During stimulation with 100 nmol/L N-α-syn, the number of TH+ cells declined in the stimulated cultures, resulting in a significant diminution in TH-immunoreactive cells (8 h: 74.6% of control; 24 h: 53.4% of control, p < 0.01; 72 h: 48.5% of control, n = 6, p < 0.01). Western blot analysis confirmed this observation, as TH expression decreased in a time-dependent manner over the course of N-α-syn stimulation (TH+/β-actin ratio at 8 h: 94.6% of control; 24 h: 86.2% of control; 72 h: 64.9% of control, p < 0.01). Analysis of cell viability with the Live/Dead cell assay demonstrated that stimulation of microglia with 100 nmol/L of N-α-syn followed by MES23.5 co-culture resulted in remarkable reduction of viable cells with concomitant increase in dead MES23.5 cells; whereas, fewer dead cells were observed in co-cultures with microglia stimulated with α-syn (non-nitrated) after 24 h (Fig. 2a). Percentage of MES23.5 cell survival was less in co-cultures with microglia stimulated with α-syn (83%) and N-α-syn (58%) compared with unstimulated controls (95%) at 72 h (Fig. 2b). The more sensitive fluorometric analysis revealed as early as 24 h after stimulation a similar pattern of progressive decline in viable cells in the presence of α-syn and N-α-syn stimulated microglia to 76% and 65% of controls at 24 h of stimulation, respectively (Fig. 2c). Moreover, N-α-syn-mediated cytotoxicity was restricted to MES23.5 cells, as stimulation of microglia in the absence of MES23.5 cells neither affected microglial survival (Fig. 2d) nor yielded a significant difference in the number of dead CD11b+ cells between control and stimulated cultures (data not shown). In addition, cytotoxicity of MES23.5 cells was not elicited with N-α-syn in the absence of microglia (Fig. 2d). Furthermore aggregation of N-α-syn was necessary for inducing microglia cytotoxicity (Fig. 2d). Importantly, a decrease in the cell survival was observed when microglia were stimulated with either aggregated α-syn (93%) or N-α-syn (86%) for 24 h, and co-cultured with MES23.5 cells in transwell inserts, but not unaggregated protein. MES23.5 cultures incubated with supernatants obtained from microglia stimulated with either α-syn or N-α-syn resulted in decreased cell survival (89% and 84%, respectively) compared with supernatants from unstimulated microglia (Fig. 2e).
NF-κB gene expression and nuclear translocation in PD
NF-κB pathway activation is critical for the initiation of inflammatory events including the production of inflammatory cytokines and chemokines linked to inflammation and microglial activation. We hypothesized that acquisition of such an inflammatory phenotype begins with induction of gene products that ultimately leads to neurotoxic factor production, cell migration, and apoptosis. To determine the extent to which this pathway was operative in PD, the SN and BG of PD brains (clinical and neuropathological profiles shown Table 1) and controls (those without neurological disease) were analyzed for NF-κB-related genes as well as neurotrophin expression (Fig. 3). Increases, albeit modest, were seen in NFKB1 expression from samples of SN from PD patients compared with controls; whereas, no significant difference was observed for RELA expression (data not shown). However, an eightfold increase in TNF expression was observed in the SN and BG together with a twofold increase in expression of its receptor TNFRSF1A. STAT1 was minimally decreased in PD brains. Similarly, analysis of AD brain tissues as a control for neuroinflammatory pathology also revealed a moderate induction of NF-κB transcription factors NFKB1 and RELA, while TNF expression was increased 40- and 10-fold in the SN and BG along with modest elevations of STAT1 in AD brain tissues compared with controls (data not shown). Based on these findings, we reasoned that a compensatory trophic mechanism could be operative in PD. Indeed, BDNF was shown to be increased greater than sixfold in the SN and twofold in the BG in PD. Consistent with recent observations by others (Backman et al. 2006), GDNF was increased greater than 10-fold in the BG but no significant changes were observed in the SN.
A recent investigation by immunofluorescence analysis of midbrain sections revealed a marked increase in expression of NF-κB p65 in the SN of PD patients compared with controls, which co-localized to CD11b+ microglia in addition to affected neurons (Ghosh et al. 2007). In the current study cytosolic and nuclear fractions were prepared from the lysates of SN of PD and control brain tissues, and lysates analyzed for NF-κB protein subunits p50 and p65. Increased expression of NF-κB subunits in both the cytosolic fractions and nuclear fractions were observed in PD brain tissues (Fig. 4). Phosphorylation of serine 536 (pS536) critical for RelA/p65 transcriptional activity was also increased in PD brain tissues.
N-α-syn-activated microglia and the PD transcriptome are linked through NF-κB
The increased expression of NF-κB transcription identified in the SN of PD brains and the microglial response to N-α-syn stimulation that were consistent with inflammatory responses suggested that one major signaling pathway induced by N-α-syn involves NF-κB activation. Use of a general microarray confirmed that NF-κB expression was increased by stimulation with N-α-syn (Fig. 5a). Using NF-κB-focused microarrays (Fig. 5b, Table 2), we showed increased expression of genes encoding pro-inflammatory cytokines, including Tnf, Ccl2, Il6, and Il1β. Also induced were those genes encoding the NF-κB transcription factor subunits, Nfkb1, Nfkb2, and Rela. In addition, N-α-syn induced genes involved in other pathways, particularly those of the mitogen-activated pathway, as indicated by the induction of the immediate early genes, Fos and Raf1. At 4 h post-stimulation, expression of most NF-κB-related genes peaked. The majority of genes induced at 1 h remained elevated, with the addition of the apoptosis-regulatory genes Card10 and Casp8. The NF-κB inhibitor, Nfkbia, was also induced (data not shown) but may become apparent only after removal or clearance of the stimulus, as Ikbkb expression was also induced at this time. Removal of N-α-syn from microglial cells after 4 h of stimulation reduced most NF-κB genes to pre-stimulatory levels. At 8 and 16 h following removal of N-α-syn from culture, several apoptosis-regulatory genes (Card10, Card11, and Cflar) were induced as well as genes for receptors of cell activation and NF-κB stimulation including Tnfrsf1a and Cd40. These results were similar but lesser in magnitude than stimulation of microglia with LPS (Fig. 5b, Table 3). Consistent with microarray analyses, quantitative RT-PCR analyses of Tnf, Il1β, and Ccl2 genes indicated very high levels of transcripts for these cytokines during stimulation by N-α-syn (10-, 3097-, and 16-fold increases, respectively) over pre-stimulatory levels (Fig. 5c). Verification of gene expression during stimulation of other, less abundant, NF-κB-related genes were achieved, including Tnfrsf1a (6.2-fold increase), Stat1 (2.3-fold increase), and Rela (3.6-fold increase). N-α-syn stimulation also increased expression of Nos2 (inducible nitric oxide synthase) and Ifng (data not shown), both regulated by NF-κB activation. Expression of the neurotrophins Bdnf and Gdnf were also increased following N-α-syn stimulation.
|Gene||Common name||NCBIb||N-α-syn (h)||LPS (h)|
|Tgfbr2||TGF-beta receptor 2||21813||−3.91||−2.8||−2.6|
|Tlr2||Toll-like receptor 2||24088||7.38||2.43|
|Tlr3||Toll-like receptor 3||142980||−2.2|
|Tlr8||Toll-like receptor 8||170744||2.38|
|Ccl2||Chemokine ligand 2 (MCP-1)||20296||36.43||3.83||2.68||15.1||14.7||14.5|
|Il1b||Interleukin 1-beta||16176||> 20||3.02||2.27||10.7||10.1||9.9|
|Ripk1||Receptor (TNFRSF)-interacting serine-threonine kinase 1||19766||−2.6||2.1|
|C3||Complement component 3||12266||2.02|
|Irak1||Interleukin-1 receptor-associated kinase 1||16179||2.19|
|Protein name||Mw (Da)b||PIc||Accession numberd||Time (h)e||Number of peptidesf||Volume ratiog|
|Proteins increased in N-α-syn-stimulated microglia cell lysates when compared with controlsa|
|10 kDa Heat-shock protein, mitochondrial||10 825||8.18||Q64433||8||6||3.45|
|S100 calcium-binding protein A13||11 151||5.89||P97352||8||2||5.57|
|Apoptosis-associated speck-like protein containing a CARD||21 459||5.26||O88597||2||6||1.58|
|Cystatin B||11 039||6.82||Q62426||8||2||3.06|
|Dynein light chain 2A||10 852||6.86||P62627||8||3||3.73|
|Eukaryotic initiation factor 5A isoform I variant D||16 821||5.08||Q7L7L3||4||2||1.51|
|Fatty acid-binding protein||14 996||6.18||Q05816||2||3||3.17|
|Heat-shock 70 kDa protein 1A||70 052||5.48||Q9EPB4||4||2||1.43|
|Heat-shock 70 kDa protein 1B||70 167||5.53||P17879||4||2||1.79|
|Heat-shock 70 kDa protein 1L||70 637||5.91||P16627||8||2||1.79|
|Histone H2B F||13 805||10.32||P10853||2||2||4.07|
|Kinesin light chain 4||68 613||5.76||Q5JQI4||8||2||1.85|
|Mitogen-activated protein-binding protein-interacting protein||13 472||5.3||Q9JHS3||8||2||4.75|
|Nucleophosmin 1||32 558||4.62||Q5U438||8||9||1.68|
|SH3 domain-binding glutamic acid-rich-like protein 3||10 470||5.02||Q91VW3||8||4||3.12|
|SWIPROSIN 1/EF hand domain containing protein 2 (Efhd2)||26 800||5.07||Q8C845||8||2||2.06|
|Capg protein||39 240||6.73||P24452||8||11||1.71|
|Capping protein||38 691||6.73||Q3TNN6||4||2||1.24|
|Myosin heavy||223 083||5.64||P02564||4||2||1.27|
|Talin||110 842||5.94||Q3TBC3||2, 4||4||1.58|
|Tubulin alpha-1 chain||50 152||4.94||P68361||8||7||1.92|
|Isovaleryl-CoA dehydrogenase||46 325||8.53||Q9JHI5||8||7||1.71|
|Cytochrome c oxidase, subunit Vb||13 838||8.34||Q9D881||8||3||3.98|
|Superoxide dismutase||24 603||8.8||P09671||2||2||1.42|
|Cathepsin C||52 347||6.41||Q3TIF1||4||2||1.29|
|Cathepsin Z||34 175||6.13||Q9ES94||8||4||1.69|
|Ferritin heavy chain||20 935||5.53||P09528||2||4||1.44|
|Hexosaminidase B||61 115||8.28||P20060||2||8||2.67|
|Peptidyl-prolyl cis-trans isomerase A||17 960||7.74||P17742||8||14||4.77|
|Ubiquitin-conjugating enzyme E2N||17 127||6.13||P61089||8||5||3.26|
|Ubiquitin-conjugating E2 Q2||42 192||4.9||Q7YQJ9||4||5||1.3|
|Vacuolar ATP synthase subunit G1||13 362||5.52||Q9D1K2||8||3||4.23|
|Annexin A5||35 752||4.83||P48036||8||4||1.7|
|39S ribosomal protein L12||21 695||9.34||Q9DB15||4||3||1.26|
|Apolipoprotein A||35 853||5.56||Q6GTX3||8||5||1.73|
|Fatty acid-binding protein (E-FABP)||15 006||6.18||Q05816||8||2||3.52|
|Fibulin 2||126 414||4.55||Q3TGL4||8||2||5.57|
|Heterogenous nuclear ribonucleoprotein K||50 976||5.39||P61979||8||8||1.69|
|Proteins decreased in N-α-syn-stimulated microglial cell lysatesa|
|14-3-3 Protein epsilon||29 155||4.63||P62259||8||8||−3.33|
|26S protease regulatory subunit 7||48 517||5.72||P46472||8||2||−2.4|
|Acyl-CoA binding protein||9863||8.78||P31786||4||8||−3|
|Adenylyl cyclase-associated protein 1||51 444||7.3||P40124||8||9||−1.9|
|Centromere protein F||367 594||5.03||P49454||2||3||−1.68|
|Chloride intracellular channel protein 1||26 865||5.09||Q9Z1Q5||8||3||−3.28|
|Eukaryotic translation initiation factor 3||35 586||5.69||Q3THA0||8||2||−3.32|
|Galectin 3||27 384||8.5||P16110||4||13||−2.34|
|Heterogenous nuclear ribonucleoproteins A2/B1||35 992||8.67||O88569||2||5||−1.61|
|Macrophage Migration Inhibitory factor||12 373||7.28||P34883||8||3||−2.03|
|Nuclear migration protein nudC||38 334||5.17||O35685||8||7||−4.13|
|Programmed cell death 6-interacting protein||96 010||6.15||Q9WU78||4||8||−1.25|
|SH3 domine-binding Glutamic acid-rich-like protein||10 477||5.02||Q91VW3||8||12||−1.81|
|Synaptotagmin-like protein 2||106 806||6.14||Q99N50||2||2||−1.46|
|Gamma actin-like protein||43 572||5.11||Q9QZ83||8||27||−4.03|
|L Plastin||70 018||5.2||Q61233||8||15||−2.58|
|MFLJ00343 protein||205 326||5.6||Q6KAM8||8||17||−2.12|
|Tubulin alpha 4||49 761||4.95||Q3TY31||8||7||−3.23|
|Tubulin alpha 6||49 907||4.96||QTIZ0||8||9||−3.25|
|Glutaredoxin 1||11 732||8.69||Q91V76||4||2||−1.7|
|Thioredoxin domain containing 5||46 386||5.51||Q3TEE8||8||38||−4.24|
|2′-5′ olygoadenylate synthetase 1F||42 270||7.09||Q8K465||2||7||−1.68|
|3-ketoacyl-CoA thiolase A||43 935||8.74||Q921H8||8||3||−2.66|
|ATP synthase||59 751||9.16||Q53XX6||2||2||−1.55|
|ATP synthase e chain||8099||9.34||Q06185||4||5||−1.52|
|Carbonyl reductase||30 394||6.15||Q8K354||4||3||−1.27|
|Cathepsin B||37 256||5.57||P10605||8||15||−3.34|
|Cathepsin D||44 925||6.71||P18242||8||34||−3.68|
|Cathepsin S||38 707||6.51||Q8BSZ5||4||3||−1.29|
|F1-ATPase alpha subunit||44 144||7.07||O78824||2||9||−1.55|
|Fructose-bisphosphate aldolase A||39 225||8.4||P05064||8||33||−2.66|
|Glutamate oxaloacetate transaminase 2||47 183||9.05||Q3TIP6||4||3||−1.23|
|Phosphomannomutase 2||27 625||6.01||Q9D1M5||4||3||−1.3|
|Succinyl-CoA ligase beta chain, mitochondrial||50 082||6.57||Q9Z219||8||3||−4.12|
|Transitional endoplasmic reticulum ATPase||53 524||4.14||Q01853||8||19||−3.31|
|Annexin A1||38 603||7.15||P10107||8||8||−1.9|
|Annexin A3||36 240||5.33||O35639||8||5||−1.93|
|28S ribosomal protein S12, mitochondrial||15 437||10.72||O35680||2||2||−1.63|
|Arcn1 protein||47 964||5.61||Q8R1S6||8||7||−1.93|
|Beta-galactoside-binding lectin||15 914||9.01||Q61357||4||7||−1.84|
|Centrosomal protein of 27 kDa||26 839||6.06||Q9CQS9||2||3||−1.68|
|Clathrin light chain B||25 171||4.56||Q6IRU5||2||3||−1.68|
|Density regulated protein||22 152||5.21||Q9CQJ6||8||2||−3.48|
|Fatty acid binding protein||14 996||6.18||Q05816||4||4||−1.27|
|Glycoprotein (transmembrane) nmb||63 577||7.88||Q3TAV1||8||2||−3.13|
|Gsn protein||80 763||5.52||Q6PAC1||8||13||−1.66|
|Histone H2A type 1||14 004||11.05||P22752||8||2||−5.46|
|Protective protein for beta-galactosidase||53 795||5.56||Q9D2D1||4||3||−3.47|
N-α-syn-activated microglial proteome shows a reactive inflammatory phenotype
Analysis of the N-α-syn microglial transcriptome showed differential gene regulation and induction of the NF-κB pathway, indicative of an inflammatory microglial phenotype. Activation of this pathway influences downstream expression of proteins involved in processes including inflammation, immune regulation, survival, and proliferation. Protein expression obtained from cell lysates were analyzed following 2, 4, and 8 h of stimulation with 100 nmol/L N-α-syn to assess the translation of differences in gene induction to intracellular protein expression. Two-dimensional DIGE was used to compare protein expression profiles of unstimulated microglia (control) and N-α-syn-stimulated microglia (Fig. 6). A complete listing of all proteins identified by LC-MS/MS is contained within Table 3.
Stimulation with N-α-syn resulted in differential expression of several proteins that are likely a consequence of NF-κB-related signaling pathways (Table 3) as soon as 2 h after stimulation. Many proteins differentially expressed could be attributed to oxidative stress, including the down-regulation of aconitase as well as the up-regulation of peroxiredoxin-1, -4, -5, superoxide dismutase, and heat-shock protein 70.
After 4 h of N-α-syn-stimulation, proteins decreased included several cytoskeletal proteins including β-actin, cofilin-1, profilin-1, tropomysin-3, and vimentin. The putative functions of other proteins decreased in N-α-syn-stimulated microglial lysates were found to be involved in cell adhesion and actin microfilament attachment to the plasma membrane (vinculin, coronin-1A, and adenylyl cyclase-associated protein 1), glycolysis and growth control (α-enolase), and migration (galectin 3 and macrophage migration inhibitory factor) (Walther et al. 2000; Chandrasekar et al. 2005). Annexin A3 is an inhibitor of phospholipase A2 and a promoter of apoptosis of inflammatory cells (Parente and Solito 2004), and was also down-regulated. The antioxidant glutaredoxin-1 was also decreased in cell lysates compared with unstimulated controls (Table 3). Four of the proteins increased in stimulated cell lysates affect intracellular calcium signaling, storage, and cell cycle regulation (swiprosin 1, calmodulin, calreticulum, and nucleophosmin 1) (Parente and Solito 2004; Vuadens et al. 2004; Meini et al. 2006).
By 8 h, 73 proteins were differentially expressed. Thirty-three proteins were decreased including all cytoskeletal proteins down-regulated at 4 h, vimentin and β-actin. Up-regulated proteins included the antioxidants superoxide dismutase, thioredoxin, and cytochrome c reductase. Oxidative stress can also lead to dysfunction of the proteasome and is implicated in PD pathogenesis (Gu et al. 2005). Indeed, as a result of N-α-syn stimulation the proteasome 26S subunit was decreased in these cell lysates, although ubiquitin and the ubiquitin conjugating enzyme E2N were increased, suggesting that the microglia may be compensating for decreased proteasomal activity (Table 3).
Neuroinflammatory Parkinson’s disease phenotype
Analysis of the proteome of N-α-syn-stimulated microglia revealed the induction of NF-κB-related signaling pathways and initiation of several proteins involved in the cellular response to inflammation and oxidative stress. To investigate whether differential expression of proteins identified in our proteomic analyses of in vitro stimulated microglia was reflected in PD, protein expression of lysates prepared from the SN and BG (data not shown) of control and PD brains were assessed by western blot assays (Fig. 7). Proteins increased in abundance within the secretome as a result of N-α-syn stimulation (A. D. Reynolds, I. Kadiu, S. G. Garg, J. G. Glanzer, T. Nordgen, R. Banerjee, P. Ciborowski and H. E. Gendelman) were cross-validated in PD patients including calmodulin and the redox-associated secreted proteins biliverdin reductase and thioredoxin; whereas, secretion of the regulatory proteins glatectin-3 and 14-3-3σ, structural protein actin, and the redox protein glutathione-S-transferase were decreased following N-α-syn stimulation. These analyses verified the increased expression of calmodulin as well as the antioxidant biliverdin reductase in the SN of PD compared with age-matched controls without neurological disease. Actin expression appeared decreased in PD brains relative to controls, which coincided with our analysis of the N-α-syn-stimulated microglia secretome. In contrast to our in vitro results, expression of 14-3-3σ and galectin 3 were increased in PD brains. Glutathione-S-transferase expression was decreased in PD brains relative to control. Although expression of thioredoxin did not appear to be different within the SN, expression in the BG was significantly decreased in PD (data not shown). Proteins that were identified in the proteome of N-α-syn-stimulated microglia were, in part, also cross-validated in SN of PD and control brains. Akin to our laboratory model, expression of calmodulin was increased whereas expression of α-enolase (data not shown), l-plastin, α-tubulin, and actin were decreased in PD relative to control. The discrepancies between the cellular model and expression in the human tissue underscore the complexity of human disease and the multiple cell components that are involved. Indeed, comparing non-affected brains to PD brains may be misleading as already the proportion of cellular components are different, especially at end stage where greater than 80% of the dopaminergic neurons have died and substantial gliosis is present. However, overall these results support that the molecular and biochemical analyses of N-α-syn microglial activation appear, in part, applicable to human PD.
Recent investigations (Biasini et al. 2004; Zhang et al. 2005; Zhou et al. 2005; Thomas et al. 2007) demonstrated that aggregated N-α-syn induces a neurotoxic inflammatory microglial phenotype that accelerates the demise of dopaminergic neurons, and as such, may contribute, in part, to PD progression. However, the mechanisms underlying N-α-syn-mediated microglial neurotoxicity remain obscure. To investigate the means by which N-α-syn-mediated microglial activation affects dopaminergic neurodegeneration, the molecular and biochemical signatures of N-α-syn-stimulated microglia were investigated. This report now demonstrates that microglial stimulation with aggregated, nitrated α-syn leads to a neuroinflammatory phenotype capable of mediating neuronal toxicity. These observations are consistent with the notion that release of this protein from injured neurons can lead to microglial activation and nigrostriatal degeneration, reflective of PD pathobiology. A key component of this study was the integration of physiologic, genomic, and proteomic techniques to develop a fingerprint of microglial cell activation following its interactions with N-α-syn. This microglia phenotype was characterized by morphological changes, as well as alterations in both the transcriptome and proteome that result in reactive microgliosis and secretion of bioactive factors, which were neurotoxic. Moreover, our examination shows human correlates of disease while permitting an integrated cross-disciplinary approach for describing a microglial ‘fingerprint’ that may be reminiscent of inflammatory processes in PD. The inflammatory microglial phenotype now shown follows its interaction with N-α-syn and may affect dopaminergic neurodegeneration.
Microglia normally function as debris scavengers, killers of microbial pathogens, regulators of the immune responses, and supporters of neuronal functions (Vilhardt 2005); all necessary for host defense. However, during neurodegenerative diseases, their phenotype is altered by uncontrolled activation. Substantial evidence for reactive microglia in and around dead or dying dopaminergic neurons in the SN of PD patients suggests that microglial activation and concomitant secretion of neurotoxic factors play a role in the nigrostriatal degeneration that occurs in PD. Stimulation by environmental cues that include aggregated proteins and inflammatory factors often results in the robust secretion of toxic factors that accelerate neuronal injury and death (McGeer and McGeer 1998; Liu and Hong 2003). Cytokines released from activated microglia bind their cognate receptor on dopaminergic neurons to activate signal transduction pathways resulting in apoptosis or necrosis. The persistent activation of microglia in response to dopaminergic neuron injury has been investigated extensively using the neurotoxin, MPTP. Work from several laboratories has documented that microglia activation accounts for ∼90% of MPTP-induced neuronal death (Wu et al. 2003; McGeer and McGeer 2004). Furthermore, studies show MPTP neurotoxicity may be attenuated in mice unable to mount pro-inflammatory responses (Feng et al. 2002; Sriram et al. 2002; Teismann et al. 2003; Wu et al. 2003), by treatment with anti-inflammatory drugs (Liu et al. 2006) and antioxidants (Zbarsky et al. 2005), blockage of the NF-κB pathway (Ghosh et al. 2007), or by induction of a regulatory T-cell response (Benner et al. 2004; Laurie et al. 2007; Reynolds et al. 2007); all converging on attenuating microglial activation. In contrast, exacerbation of microglial activation by infiltrating effector lymphocytes to modified self-peptides may worsen MPTP-induced neurodegeneration (E. J. Benner, R. Banerjee, A. D. Reynolds, S. Sherman, V. M. Pisarev, V. Tsiperson, C. Nemacheck, P. Ciborowski, S. Przedborski, R. L. Mosley, and H. E. Gendelman, unpublished data).
Generation of reactive molecular species by microglia, as well as changes that occur during dopamine metabolism and mitochondrial function can result in oxidation and nitration of proteins, DNA modifications, and lipid peroxidation. Oxidation and nitration of α-syn leads to formation of aggregates and the stabilization of assembled filaments found to be a major component of LBs, the hallmark lesions of PD. The results of the present study, as well as research performed by others (Biasini et al. 2004; Zhang et al. 2005; Zhou et al. 2005; Thomas et al. 2007) support the hypothesis that microglia activated by N-α-syn is a component of an inflammatory cascade that perpetuates nigrostriatal degeneration in PD. First, nitrated α-syn was identified in extracts from the SN of PD patients in copious concentrations relative to control and AD brains. Second, α-syn aggregates released from LB during dopaminergic neuronal death can interact with adjacent microglial cells found in abundance within the SNpc of PD patients (Spillantini et al. 1997; McGeer and McGeer 1998; Croisier et al. 2005). Alternatively, α-syn may also be released or secreted from the cytosol of dopaminergic cells into the extracellular environment where it is more prone to aggregation and oxidative damage (Kakimura et al. 2001; Lee et al. 2005; Sung et al. 2005). Third, microglial activation is associated with degenerating dopaminergic neurons and deposition of α-syn in the SN of PD patients (Croisier et al. 2005). Fourth, native and nitrated α-syn activate microglia with release of ROS and induce neurotoxicity as shown herein and by others (Zhang et al. 2005; Thomas et al. 2007). Fifth, recent evidence supports that N-α-syn drains to the cervical lymph nodes, availing it for processing and presentation by antigen-presenting cells to the adaptive immune system, which in turn can circumvent or break immunological tolerance to direct immune responses that possibly contribute to prolonged microglial activation and neurodegeneration (E. J. Benner, R. Banerjee, A. D. Reynolds, S. Sherman, V. M. Pisarev, V. Tsiperson, C. Nemacheck, P. Ciborowski, S. Przedborski, R. L. Mosley, and H. E. Gendelman, unpublished data). Moreover, α-syn and its modified forms are present in extracellular biological fluids including human plasma (El-Agnaf et al. 2003) and are proposed as biomarkers for disease (Fjorback et al. 2007). Evidence of systemic complications associated with PD including abnormal gastrointestinal function (Bassotti et al. 2000), cardiac denervation, and orthostatic hypotension (Taki et al. 2000; Goldstein et al. 2005) further suggest a peripheral component in disease. Taken together, this work extends those of others (Wang et al. 2005; Zhou et al. 2005; McLaughlin et al. 2006). By using proteomic analyses for examination of the activated microglial proteome it provides human disease correlates while permitting an integrated cross-disciplinary proteomic approach towards elucidating a PD microglial ‘fingerprint’. The work provides evidence that such a profile is inflammatory and may be linked to neurotoxicity. However, one must exert caution in over interpreting these experimental results. Nonetheless, whether or not these findings are directly linked to the pathogenesis of PD will certainly require further study. Although clear evidence is provided that microglial activation is part of PD whether this process is a secondary by-product of ongoing neurodegeneration or a primary inducer of disease remains uncertain.
Parallels are demonstrated herein between N-α-syn and LPS for microglial activation and support a commonality for innate immune responses in disease. The findings suggest that pro-inflammatory processes may be common amongst mononuclear phagocytes that see disparate activators. LPS is a strong activator of microglia both in vivo and in vitro. A single systemic exposure to LPS can lead to neuroinflammation associated with increased expression of pro-inflammatory cytokines, NADPH oxidase-mediated release of superoxide (Gao et al. 2002), and activation of the NF-κB pathway (Qin et al. 2007) resulting in neurodegeneration. Microglial activation by LPS and N-α-syn were associated with induction of the ΝF-κB and mitogen activation pathways, characteristic of an inflammatory phenotype. However, key differences in the inflammatory responses induced by LPS or N-α-syn were identified, including genes involved in signal transduction and apoptosis, as well as induction of an inflammatory response that was greater in magnitude after stimulation with LPS compared with that of N-α-syn. It is possible that the differences in transcription may be dose or pathway dependent as the stimulatory capacity and pathways activated by the two stimuli may differ significantly. Having demonstrated that N-α-syn stimulation induces microglial activation, this model may be used to study PD and reflect the unique molecular changes that occur during disease progression.
Our finding suggests that modifications to α-syn may be a common denominator for microglial activation in sporadic and familial PD. These observations also identify prospective pathways that are associated with PD and as such uncover potential targets for therapeutic intervention. For instance, identification of NF-κB activation by microglia in response to divergent stimuli suggests that activation of NF-κB and its related signaling pathways may be a key component in the inflammatory response leading to neuronal death. The microglial response to N-α-syn was linked to neurotoxicity. Nonetheless, we also showed proteomic fingerprints that were potentially protective. These include antioxidants and growth factors. Although considered to be a necessary component of CNS homeostasis, the potentially protective microglia mechanisms are lost as PD progresses uncontrolled. Furthermore, the similarities and differences found between the acute in vitro model of N-α-syn microglial stimulation and what is present in PD brains support that ongoing inflammatory responses present in disease may affect CNS protective responses. The observations of sustained NF-κB activation and differential expression of regulatory, structural, and redox proteins at end-stage disease support, in part, a persistent inflammatory process that could affect dopaminergic loss.
In summary, a mechanistic role for aggregated N-α-syn in stimulating a neurotoxic microglial phenotype was observed. In this ‘potential’ scheme, a pathogenic paracrine loop of immune activation occurs consisting of dopaminergic neuronal injury or death, release of aggregated N-α-syn from the cytosol or LB either through exocytosis or neuronal degeneration into the extracellular milieu, microglial activation with release of toxic factors which, may ultimately lead to further neuronal injury and sustained α-syn release. Regardless of concurrent pathogenic events in PD, which initiate dopaminergic degeneration or the mechanisms associated with activation by N-α-syn, the resultant activation of resident microglia could, in part, perpetuate neuronal injury and subsequent disease progression. Thus, the perpetual presence of activated microglia, left uncontrolled, may consistently confound PD therapies. Our investigations also provide a novel approach towards elucidating cellular immune responses for neurodegeneration and suggest potential molecular targets to slow disease processes.
We thank Dr E. Benner for providing recombinant mouse α-synuclein, Dr S. Appel for providing the MES23.5 cell line, and L. Shlyahtenko for preparing the AFM images and the cell analysis facility for flow cytometric analyses. We are deeply indepted to Drs Susan Morgello and Eliezer Masliah and the National Research Tissue Consortium for help in providing the human brain autopsy tissues. The work was supported by the Frances and Louis Blumkin Foundation, the Community Neuroscience Pride of Nebraska Research Initiative, and the Alan Baer Charitable Trust (to HEG), a University of Nebraska Medical Center Graduate Student Excellence Award (to ADR), Michael J. Fox Foundation (to RLM), and NIH grants 1T32 NS07488 (to ADR and HEG), 1R21 NS049264 (to RLM), 2R37 NS36136, PO1 NS43985, PO1 MH64570, R01 MH79886 (to HEG), and R24 NS45491 and R01 MH79886 (to BG).