Address correspondence and reprint requests to Byung K. Jin PhD, Brain Disease Research Center, Ajou University School of Medicine, Suwon 443–721, Korea. E-mail: firstname.lastname@example.org
The present study shows that activation of microglial NADPH oxidase and production of reactive oxygen species (ROS) is associated with thrombin-induced degeneration of nigral dopaminergic neurons in vivo. Seven days after thrombin injection in the rat substantia nigra (SN), tyrosine hydroxylase immunocytochemistry showed a significant loss of nigral dopaminergic neurons. This cell death was accompanied by localization of terminal deoxynucleotidyl transferase-mediated fluorecein UTP nick-end labelling (TUNEL) staining within dopaminergic neurons. This neurotoxicity was antagonized by the semisynthetic tetracycline derivative, minocycline, and the observed neuroprotective effects were associated with the ability of minocycline to suppress NADPH oxidase-derived ROS production and pro-inflammatory cytokine expression, including interleukin-1β and inducible nitric oxide synthase, from activated microglia. These results suggest that microglial NADPH oxidase may be a viable target for neuroprotection against oxidative damage.
Microglia, the principal immune cells of the brain, activate quickly in response to various insults (Kreutzberg 1996). Activated microglia (larger cell bodies with short, thick, or no processes) are generally identified by morphological changes compared with resting microglia (small cell bodies and thin, long or ramified processes); in several neurodegenerative diseases, including Parkinson's disease (PD) (McGeer et al. 1988; Vila et al. 2001; Imamura et al. 2003), activated microglia may gain neurotoxic functions through production of neurotoxic and inflammatory mediators (Kreutzberg 1996; Aloisi 2001). These activated microglial-derived neurotoxins include reactive oxygen species (ROS) produced by NADPH oxidase. The mechanisms leading to NADPH oxidase activation in microglia are unclear, but data from other cell types provide some insights. In resting neutrophils, the five subunits of NADPH oxidase are segregated into the cytosolic components (p40phox, p47phox and p67phox) and the plasma membrane components (p22phox and gp91phox); the oxidase is inactive under these conditions (Babior 1999). When the resting cells are exposed to a variety of stimuli, the cytosolic p47phox becomes heavily phosphorylated and the entire cytosolic complex migrates to the membrane, where it associates with the membrane-bound components to assemble the active oxidase (Babior 1999). Thrombin has been shown to activate the NADPH oxidase complex in cultured human vascular smooth muscle cells, which, in turn, may generate ROS and oxidative stress (Patterson et al. 1999). Regarding this, our very recent data also show that thrombin activates microglial NADPH oxidase and produces oxidative stress, leading to eventual cell death of hippocampal neurons in vivo (Choi et al. 2005). In terms of pathogenic conditions, oxidative stress has been shown to play an important role in the degeneration of dopaminergic neurons in the substantia nigra (SN) of PD brains (Fahn and Cohen 1992; Zhang et al. 1999; Beal 2002; Koutsilieri et al. 2002), and microglial NADPH oxidase-induced oxidative stress has been implicated in the degeneration of dopaminergic neurons in both in vivo and in vitro models of PD (Wu et al. 2002; Gao et al. 2003; Wu et al. 2003; Block et al. 2004; Qin et al. 2004).
We recently reported that thrombin-induced microglial activation is mediated by extracellular signal-regulated kinase 1/2 (ERK1/2) and p38 mitogen-activated protein kinase (MAPK) and results in the production of toxic and inflammatory mediators, leading to eventual cell death of nigral dopaminergic neurons (Choi et al. 2003a). In this study, we investigated whether thrombin is capable of inducing neurotoxic effects on the rat SN through activation of NADPH oxidase and production of ROS from activated microglia. We also examined whether minocycline could prevent thrombin-induced neurotoxicity in vivo.
Materials and methods
Stereotaxic surgery and minocycline administration
All experiments were carried out in accordance with the approved animal protocols and guidelines established by Ajou University (Suwon, Korea). Female Sprague–Dawley rats (260–280 g) were anaesthetized with chloral hydrate [360 mg/kg, intraperitoneal (i.p.) injection] and positioned in a stereotaxic apparatus (Kopf Instrument, Tujunga, CA, USA). Each rat received a unilateral administration of thrombin into the right SN (anteroposterior 5.3 mm, mediolateral 2.3, dorsoventral 7.6 mm from bregma), according to the atlas of Paxinos and Watson (1998). All injections used a Hamilton syringe equipped with a 26S-gauge bevelled needle and attached to a syringe pump (K. D. Scientific, Holliston, MA, USA). Infusions were made at a rate of 0.2 μL/min for thrombin [20 U in 4 μL sterile phosphate-buffered saline (PBS); Sigma, St Louis, MO, USA] and for vehicle [PBS or bovine serum albumin (BSA; 200 μg in 4 μL sterile saline] as controls. Because PBS had the same effects as BSA, the data of PBS were used as controls in the current study, as we recently described (Choi et al. 2005). After injection, the needle was left in place for an additional 5 min before being slowly retracted. For histological studies, animals received i.p. injections of either vehicle (saline) or minocycline (25 or 50 mg/kg per day; Sigma) dissolved in vehicle, starting 1 day before thrombin injection and continued daily until day 6 post-injection. For studies of early gene and protein expression, animals were pretreated with minocycline (25 or 50 mg/kg) or saline vehicle 1 day before thrombin injection and again 1 h prior to the thrombin injection. Animals were humanely killed and their brains harvested at the indicated time points for the various analyses. To eliminate lipopolysaccharide (LPS) contamination on thrombin used in the present study, we determined whether polymyxin B, LPS inhibitor, altered effects of thrombin. Consistent with our recent data (Lee et al. 2005), polymyxin B attenuated the LPS-induced expression of iNOS, TNF-α and IL-6 mRNA, whereas polymyxin B had little effect on inducible nitric oxide synthase (iNOS), TNF-α, and IL-6 expression induced by thrombin pre-incubated with hirudin or thrombin alone (data not shown). These data indicate that LPS could not be a significant contaminant.
Tissue preparation and immunohistochemistry
Animals were transcardially perfused with a saline solution containing 0.5% sodium nitrate and heparin (10 U/mL) and fixed with 4% paraformaldehyde dissolved in 0.1 m phosphate buffer. Brains were removed from the skull, post-fixed overnight in buffered 4% paraformaldehyde at 4°C and stored in 30% sucrose solution at 4°C until they sank, at which time samples were frozen sectioned on a sliding microtome. Coronal sections (40-μm thick) were collected in six separate series and processed for immunohistochemical staining as previously described (Choi et al. 2003a,b). In brief, sections were rinsed in PBS and then incubated overnight at room temperature with the following primary antibodies: mouse anti-OX-6 (specific for major histocompatibility complex class II antigens, 1 : 200; BD Biosciences, San Diego, CA, USA) for microglia, mouse anti-NeuN (neuronal nucleus, 1 : 200; Serotec, Oxford, UK) for general neurons, and rabbit anti-tyrosine hydroxylase (TH; 1 : 2000; Pel-Freez Biologicals, Rogers, AR, USA) for dopaminergic neurons. The following day, the brain sections were rinsed with PBS containing 0.5% bovine serum albumin (BSA), incubated with the appropriate biotinylated secondary antibody and processed with an avidin-biotin complex kit (Vectastain ABC Kit; Vector Laboratories, Burlingame, CA, USA). Immunostaining was visualized with 3,3′-diaminobenzidine-HCl (DAB), and tissue sections were mounted on gelatin-coated slides and analysed under a bright-field microscope (Olympus, Tokyo, Japan). For Nissl staining, SN tissue samples were mounted on gelatin-coated slides, dried for 1 h at room temperature, stained in 0.5% cresyl violet (Sigma), dehydrated, coverslipped, and then analysed with a bright-field microscope (Olympus).
The total number of TH-positive neurons was counted in the various animal groups at 7 days post-injection (thrombin or PBS) using the optical fractionator method performed on an Olympus CAST (Computer Assisted Stereological Toolbox) system version 2.1.4 (Olympus Denmark A/S, Ballerup, Denmark) as previously described (Choi et al. 2003a,b). This unbiased stereological method of cell counting is not affected by either the reference volume (SNpc) or the size of the counted elements (neurons) (West et al. 1991).
As previously described (Shin et al. 2004), the amount of apoptosis in thrombin-treated neurons was determined by double-staining of TUNEL with TH antibody. TUNEL assays were performed using the Apoptag fluorescein in situ detection kit (Intergen, NY, USA) that detects the 3′-OH region of cleaved DNA during apoptosis and the protocol recommended by the manufacturer. Briefly, brain tissue microsections were incubated for 30 min in a permeabilization solution (0.2% Triton X-100 and 0.5% BSA in PBS, pH 7.4). TUNEL reaction mixture was added, tissue incubated in a humidified chamber for 1 h at 37°C, and washed in PBS. For counter staining with dopaminergic neurons, brain sections were incubated with TH antibodies (1 : 2000; Pel-Freez Biologicals) overnight at 4°C. The brain tissues were washed, incubated with Texas Red-labelled anti-rabbit IgG (1 : 200), washed, mounted with Vectashield mounting medium, and viewed using an Olympus IX71 confocal laser scanning microscope (CLSM; Olympus). Images of TUNEL staining with TH immunofluorescence were created from the same tissue section and merged using interactive software.
Double-immunofluorescence staining was used to identify the cellular localization of p67phox expression, as previously described (Choi et al. 2003a,b). Briefly, free-floating sections were mounted on gelatin-coated slides and dried for 1 h at room temperature. Sections were washed in PBS, incubated in 0.2% Triton X-100 for 30 min and rinsed three times with PBS containing 0.5% BSA. The sections were then incubated with monoclonal anti-p67phox (1 : 200; BD Transduction) with or without polyclonal anti-TH (1: 2000; Pel-Freez Biologicals) for dopaminergic neurons overnight at 4°C, washed in PBS and incubated with Texas Red-conjugated goat anti-mouse IgG (1 : 100; Vector Laboratories) and fluorescein-conjugated lycopersicon esculentum (tomato) lectin (TL; 1 : 200; Vector Laboratories) for microglia or fluorescein-conjugated anti-rabbit igG (1 : 100; Vector Laboratories) for 1 h at room temperature. Finally, the slides were coverslipped with Vectorshield medium (Vector Laboratories) and viewed using an IX71 confocal microscope (Olympus).
In situ detection of O2– production
Numerous methods have been reported for in vivo detection of O2– production. Among these, we felt that hydroethidine conversion to ethidium might be a useful tool for detecting production of O2– in our experiments, because hydroethidine is selectively oxidized to ethidium by O2–, but not by other ROS (Bindokas et al. 1996; Wu et al. 2003). For these experiments, hydroethidine (Molecular Probes, Eugene, OR, USA; 1 mg/mL in 1% dimethylsulfoxide with PBS) was intraperitoneally administered 24 h after thrombin injection and animals were killed 15 min later by transcardial perfusion with heparin (10 U/mL) and 4% paraformaldehyde in saline. Brains were removed, post-fixed, sectioned (40 μm), and mounted on gelatin-coated slides. Ethidium accumulation, which represented generation of the oxidized hydroethidine products, was examined by confocal microscopy (Olympus).
Animals treated with or without minocycline (25 or 50 mg/kg, i.p. twice) were decapitated 12 h after intranigral injection of thrombin (20 U), and the ipsilateral SN regions were immediately isolated. Total RNA was prepared with RNAzol B (Tel-Test, Friendwood, TX, USA) and RT was carried out using the Superscript II reverse transcriptase (Life Technologies, Rockville, MD, USA) according to the manufacturer's instructions. The primer sequences used in this study were as follows: 5′-TGATGTTCCCATTAGACAGC-3′ (forward) and 5′-GAGGTGCTGATGTACCAGTT-3′ (reverse) for interleukin-1β (IL-1β); 5′-GCAGAATGTGACCATCATGG-3′ (forward), and 5′-ACAACCTTGGTGTTGAAGGC-3′ (reverse) for inducible nitric oxide synthase (iNOS); and 5′-TCCCTCAAGATTGTCAGCAA-3′ (forward) and 5′-AGA TCCACAACGGATACATT-3′ (reverse) for glyceraldehyde-3-phosphate dehydrogenase. Real-time PCRs were performed in a reaction volume of 20 μL including 1 μL RT product as a template, 10 μL of SYBR Green PCR master mix (Applied Biosystems, Warrington, UK) and 20 pmol of each primer described above. The PCR amplifications were performed with 40 cycles of 95°C for 30 s and 60°C for 60 s using ABI 7500 (Applied Biosystems). Average Ct values of IL-1β and iNOS from triplicate PCR reactions were normalized from average Ct values of glyceraldehyde-3-phosphate dehydrogenase. The ratios of expression levels of IL-1β and iNOS between animals treated with PBS and thrombin only or thrombin with minocycline were calculated as 2–(meanΔΔCt).
Western immunoblot analysis
For p38 MAPK and ERK1/2 analyses, ipsilateral SN tissues were dissected 4 h after animals treated with or without minocycline (25 or 50 mg/kg, i.p. twice) were subjected to intranigral injection of PBS or thrombin (20 U). For TH analyses, ipsilateral SN tissues were dissected 4, 8, 12 and 24 h after intranigral injection of thrombin (20 U). SN samples were homogenized with ice-cold lysis buffer containing 20 mm Tris-HCl, pH 7.5, 1 mm EDTA, 5 mm MgCl2, 1 mm dithiothreitol, 0.1 mm phenylmethylsulfonyl fluoride and protease inhibitor cocktail (Sigma) in a Dounce homogenizer (Wheaton, Millville, NJ, USA). For p67phox analysis, cytosolic and membrane fractions from ipsilateral SN were prepared as previously described, with some modifications (Patterson et al. 1999). Proteins (50 μg/lane) were separated by 10–12% sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDS–PAGE), transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA) using an electrophoretic transfer system (Bio-Rad Laboratories, Hercules, CA, USA), and subjected to immunoblotting with the following specific primary antibodies: rabbit anti-phospho-p38 MAPK (1 : 500; Cell Signaling Technology, Beverly, MA, USA), rabbit anti-phospho-ERK1/2 (1 : 1000; Cell Signaling Technology), rabbit anti-TH (1 : 4000; Pel-Freez Biologicals) and mouse anti-p67phox (1 : 500; BD Transduction). The blots were subsequently stripped and re-probed with antibodies against total p38 MAPK (1 : 500; Cell Signaling Technology), total ERK1/2 (1 : 1000; Cell Signaling Technology), rabbit anti-calnexin (1 : 1000; Stressgen, Victoria, BC, Canada) and actin (1: 2000; Santa Cruz Biotechnology, Santa Cruz, CA, USA). For semiquantitative analyses, the densities of the immunoblot bands were measured with the Computer Imaging Device and accompanying software (Fuji Film).
All values are expressed as mean ± SEM. Statistical significance (p < 0.05 for all analyses) was assessed by anova using the Instat 3.05 software package (GraphPad Software Inc., San Diego, CA, USA), followed by Student–Newman–Keuls analyses.
Minocycline protects dopaminergic neurons from thrombin-induced neurotoxicity via inhibition of microglial activation
Thrombin (20 U) or PBS as a control was unilaterally injected into the SN of rats. Seven days later, brains were removed and sections were processed for Nissl staining, immunostaining for NeuN to detect general neurons, or immunostaining for TH to specifically detect dopaminergic neurons. Consistent with our recent study (Choi et al. 2003a), a considerable loss of Nissl-stained (Fig. 1b), NeuN-immunopositive (NeuN-ip) (Fig. 1e), and TH-ip cells (Fig. 1h) was evident at 7 days in the thrombin-injected SN compared with PBS-treated controls (Figs 1a, d and g). Highly magnified microphotographs (Fig. 1, parts labelled with Roman numerals) revealed that thrombin treatment induced marked loss of Nissl substances with gliosis (Fig. 1ii) and neuronal degeneration, characterized by shrunken neuronal cell bodies (Figs 1e and v). In contrast, the PBS-treated control SN samples showed large, healthy dopaminergic neurons with long, branched neuritic processes (Fig. 1iv). To corroborate neuronal death, animals were killed 24 h after intranigral injection of thrombin (20 U). Brain tissue was processed for double-immunofluorescence staining with TH and TUNEL to visualize nuclear DNA cleavage during cell death of dopaminergic neurons. Thrombin-induced TUNEL-positive cells were detected within TH-ip cells (Fig. 1j), whereas PBS-treated cells in SN were TUNEL-negative (data not shown).
When neuronal degeneration was quantified and expressed as a percentage of TH-ip neurons on the ipsilateral SN, we noted that thrombin treatment decreased the number of TH-ip neurons by 57% (Fig. 1k; p < 0.05) as compared with the PBS-injected SN. In good agreement with our recent findings (Choi et al. 2003a), additional immunostaining also showed a significant loss of glutamic acid decarboxylase-ip neurons GABA-ergic neurons in the substantia nigra reticulate after intranigral injection of thrombin (data not shown).
To investigate whether minocycline altered thrombin-induced neurotoxicity of nigral neurons, minocycline was administered for 7 days, starting 1 day before thrombin injection. The results of Nissl staining (Figs 1ciii) and NeuN immunohistochemistry (Fig. 1f) showed that minocycline treatment effectively reduced thrombin-induced neuronal death in the SN. Furthermore, minocycline significantly attenuated the loss of dopaminergic neurons (Fig. 1i), and relatively healthy dopaminergic neurons with branched processes were observed in minocycline/thrombin-treated SN samples (Fig. 1vi). When quantified and expressed as the percentage of TH-ip neurons on the ipsilateral SN, administration of 25 or 50 mg/kg minocycline was found to increase the number of TH-ip neurons by 18% (Fig. 1k; p < 0.05) and 27% (Fig. 1k; p < 0.01), respectively, compared with saline-treated thrombin-injected SN samples (negative control). When animals received saline (in the absence of thrombin and minocycline) or minocycline alone (25 or 50 mg/kg), no remarkable reduction of dopaminergic neurons was evident in the SN (Fig. 1k).
We recently reported that 7 days after intranigral injection of thrombin results in degeneration of nigral dopaminergic neurons in vivo, in association with microglial activation as early as 4 h post-thrombin (Choi et al. 2003a). Accompanying these results, there was no substantial loss of nigral dopaminergic neurons as determined by western blot analysis of TH in the SN at 4, 8, 12 and 24 h after intranigral injection of thrombin (Fig. 2), indicating that microglial activation preceded neurodegeneration in the SN in vivo. In addition, increasing evidence suggests that the neuroprotective effects of minocycline on nigral dopaminergic neurons in vivo are related to inhibition of microglial activation (Du et al. 2001; He et al. 2001; Wu et al. 2002). Thus, we next examined whether the observed neuroprotective effects of minocycline resulted from inhibition of thrombin-induced microglial activation. For this purpose, minocycline (25 or 50 mg/kg, i.p.) was administered twice (1 day and 1 h) before intranigral thrombin (20 U) injection. Sections were prepared for immunohistochemical staining with OX-6 antibody to detect microglial activation at 24 h after intranigral injection of thrombin. Consistent with our recent report (Choi et al. 2003a), OX-6-ip cells were seen exclusively along the needle tract in PBS-treated SN samples (Fig. 3a), whereas thrombin-treated SN samples showed a large number of OX-6-ip cells (Fig. 3b). In contrast, pretreatment with 50 mg/kg minocycline was found to dramatically mitigate the number of OX-6-ip cells in the thrombin-injected SN (Fig. 3c). Treatment with 25 mg/kg minocycline had relatively little effect (data not shown), and minocycline alone (50 mg/kg, i.p.) had no effects on microglial activation as evidenced by OX-6 immunostaining (Fig. 3i).
Thrombin-induced O2– production via microglial NADPH oxidase is inhibited by minocycline
Thrombin was found to induce production of hydrogen peroxide and O2– in human aortic smooth muscle cells via NADPH oxidase (Patterson et al. 1999), and O2– originating from microglia is thought to mediate loss of dopaminergic neurons in the SN (Gao et al. 2003; Wu et al. 2003; Block et al. 2004; Qin et al. 2004). Thus, we examined whether thrombin induced O2– production in the SN, and whether minocycline mediated neuronal survival by inhibiting the thrombin-induced production of O2–. To test this, hydroethidine histochemistry was performed on sections adjacent to those used for OX-6 immunostaining (Figs 3a–c) for in situ visualization of thrombin-induced O2– production (Wu et al. 2003). The fluorescent products of oxidized hydroethidine (i.e. ethidium accumulation) were significantly increased 24 h after thrombin injection in the SN (Fig. 3e) compared with the PBS-injected SN control (Fig. 3d). In contrast, thrombin-induced O2– was dramatically decreased in SN samples co-treated with minocycline (50 mg/kg, i.p.) and thrombin (Fig. 3f). These observations are consistent with the data obtained from the OX-6 immunostaining (Fig. 3c).
As NADPH oxidase is the source of O2– production in phagocytic cells such as microglia, we investigated whether thrombin altered NADPH oxidase by measuring the levels and localization of p67phox, one of the cytosolic components of NADPH oxidase. Western blot analyses showed that, in thrombin-treated SN samples, levels of p67phox were significantly increased in the cytosolic (Fig. 4a, upper panel) and membrane (Fig. 4a, lower panel) fractions 12 h after thrombin injection compared with PBS-treated SN samples (Fig. 4b), indicating that there was elevated expression and translocation of this subunit. Moreover, the p67phox-ip cells (Fig. 4c; upper panel, red) were observed 12 h after thrombin injection in the SN and appeared localized within tomato lectin (TL)-positive microglia (Fig. 4c; upper panel, green). Additional double-immunofluorescence staining demonstrated in vivo that, in thrombin-treated SN, the p67phox-ip cells (Fig. 4c; lower panel, red) also were localized within TH-ip neurons (Fig. 4c; lower panel, green), but not astrocytes (data not shown).
We next examined whether treatment with minocycline altered the effects of thrombin on NADPH oxidase. In the minocycline-treated SN samples, the levels of p67phox protein were reduced in both the cytosolic and membrane fractions, as compared with saline-treated negative controls (Fig. 4a). When minocycline (50 mg/kg, i.p.) was given before thrombin injection, translocation of p67phox from the cytosol to the membrane was significantly decreased in the SN by 82% (Fig. 4b; p < 0.05). Minocycline (50 mg/kg, i.p., data not shown)-only-treated controls had no effects on the level of p67phox in both the cytosolic and membrane fractions.
Effects of minocycline on thrombin-induced expression of IL-1β and iNOS
It has been shown that transgenic mice expressing a dominant negative inhibitor of IL-1β converting enzyme or deficient in iNOS are resistant to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced neurotoxicity (Klevenyi et al. 1999; Liberatore et al. 1999; Dehmer et al. 2000). Previous studies also reported that minocycline inhibited induction of iNOS and IL-1β in animal models of ischaemia (Yrjänheikki et al. 1998) and PD (Wu et al. 2002). Thus, we examined whether thrombin-induced expression of IL-1β and iNOS in the SN could be affected by minocycline, leading to neuronal survival. For this purpose, minocycline (25 or 50 mg/kg, i.p.) was administered twice (1 day and 1 h) prior to intranigral injection of thrombin (20 U). Animals were humanely killed 12 h after thrombin injection and brain tissues were dissected and prepared for real-time PCR analysis. We chose 12 h post-thrombin because thrombin-induced expression of IL-1β and iNOS in the SN was evident at this time point, described in our recent report (Choi et al. 2003a). The results of the real-time PCR showed that minocycline attenuated thrombin-induced expression of IL-1β and iNOS mRNA in the SN (Fig. 5). When minocycline (50 mg/kg, i.p.) was given before thrombin injection, real-time PCR analysis revealed that mRNA expression of both IL-1β and iNOS was significantly reduced by 73% (Fig. 5; p < 0.001) and 46% (Fig. 5; p < 0.05), respectively. Treatment with 25 mg/kg minocycline also reduced IL-1β by 75% (Fig. 5; p < 0.001), but had no effect on iNOS (Fig. 5; p = 0.07). These results indicate that thrombin-induced expression of IL-1β and iNOS in the SN could be affected by minocycline, perhaps accounting, at least in part, for the observed minocycline-induced neuroprotection. Minocycline (50 mg/kg, i.p., data not shown)-only-treated controls had no effects on the mRNA expression of IL-1β and iNOS.
Effects of minocycline on the phosphorylation of p38 MAPK and ERK1/2
We recently reported that thrombin induces activation of ERK1/2 and p38 MAPK in microglia and inhibition of these MAPKs rescues dopaminergic neuron in the SN in vivo (Choi et al. 2003a). Thus, we next examined whether minocycline inhibited thrombin-induced activation of p38 MAPK and ERK1/2 in the SN, leading to neuronal survival. In agreement with our recent study (Choi et al. 2003a), western blot analysis showed increased levels of P-p38 MAPK (Fig. 6a) and P-ERK1/2 (Fig. 6c) in the ipsilateral SN 4 h after thrombin injection. In contrast, administration of minocycline (50 mg/kg, i.p.) reduced P-p38 MAPK by 28% (Fig. 6b; p < 0.05) and P-ERK1/2 by 67% (Fig. 6d; p < 0.01), respectively, compared with thrombin-elevated levels. Treatment with 25 mg/kg minocycline also reduced P-ERK1/2 by 61%, but had no effect on P-p38 MAPK. Minocycline (50 mg/kg, i.p., data not shown)-only-treated controls had no effects on the activation of ERK and p38 MAPK.
In this study, we demonstrated that activation of NADPH oxidase and generation of NADPH oxidase-derived O2– from activated microglia is an important determinant in thrombin-induced neurotoxicity in the SN in vivo.
Increasing evidence suggests that activated microglia induce or exacerbate neurotoxicity by generating oxidative damage to neurons. Because of their well-known features such as depletion of glutathione (Sofic et al. 1992; Sian et al. 1994; Bharath et al. 2002) or increased accumulation of total iron (Dexter et al. 1989b; Bharath et al. 2002), dopaminergic neurons in the SN are particularly vulnerable to oxidative stresses. Furthermore, the number of microglia is higher in the SN compared with other brain regions (Lawson et al. 1990; Kim et al. 2000); these microglia are quickly activated by various insults (Kreutzberg 1996), leading to production of various microglia-derived neurotoxins. Two of these deleterious microglia-derived factors, NADPH oxidase and iNOS, are thought to be key players in mediation of neuronal cell death, as inhibition of their activity resulted in reduced losses of dopaminergic neurons in the SN (Gao et al. 2003; Wu et al. 2003). NADPH oxidase is a multicomponent enzyme responsible for generating extracellular O2– in phagocytic cells (Maher and Schubert 2000). This O2– may be converted to membrane-permeable hydrogen peroxide or other downstream toxic products (i.e. hydroxyl radicals) in the extracellular space; these ROS can cross the cell membrane and cause cellular damage to neighbouring neurons (Cadet and Brannock 1998). In addition to O2–, NO generated from iNOS is also thought to contribute to the oxidative stress associated with the neurodegeneration observed in PD (Hunot et al. 1996; Dawson and Dawson 1998). The neurotoxic effects of NO are usually attributed to its reaction with O2– to form peroxynitrite, which can cause neuronal injuries, including non-specific protein oxidation, lipid peroxidation, DNA damage, and inhibition of the mitochondrial respiratory chain, all of which are seen in PD brains (Dexter et al. 1989a; Alam et al. 1997a,b; Beal 2002; Buhmann et al. 2004). The present study showed that thrombin-induced increases in O2– production occurred via activation of microglial NADPH oxidase in the SN, and could be reduced by minocycline. Consistent with our recent study (Choi et al. 2003a), we also found that the observed thrombin-induced increase in iNOS expression in the SN was reversed by minocycline. Collectively, these results indicate that minocycline inhibits thrombin-induced microglial activation along with NADPH oxidase activation and iNOS expression, which may result in reduction of oxidative damage to dopaminergic neurons in the SN and increased neuronal survival. This is consistent with a recent report in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced animal models of PD, showing that minocycline protected dopaminergic neurons in the SN; this was attributed to inhibition of microglial-derived superoxide via activation of NADPH oxidase and NO production (Du et al. 2001; Wu et al. 2002). While our results point to a likely role of microglial NADPH oxidase, recent findings demonstrate that the amyloid protein precursor-induced loss of cortical neurons in cultures is mediated by neuronal NADPH oxidase (Niikura et al. 2004) and that β-amyloid-induced neurotoxicity in cultured hippocampal neurons is mediated by astrocyte NADPH oxidase (Abramov et al. 2004). Therefore, it is likely that NADPH oxidase originating from neurons (Serrano et al. 2003) or astrocytes (Noh and Koh 2000; Tammariello et al. 2000) may also participate in thrombin-induced neurotoxicity. This hypothesis is supported by our current results that NADPH oxidase is expressed in neurons in thrombin-treated SN in vivo.
We recently reported that the p38 MAPK and ERK1/2 pathways play important roles, not only in the process of microglial activation, but also in the transcriptional regulation of gene expression by thrombin (Ryu et al. 2000; Choi et al. 2003b). Increasing evidences both in vivo and in vitro have also suggested that the neuroprotective effects of minocycline resulted in the inhibition of p38 MAPK activity, which is expressed in microglia (Du et al. 2001; Lin et al. 2001; Tikka and Koistinaho 2001; Tikka et al. 2001; Kriz et al. 2002). However, these studies showed little or no effect of minocycline on ERK1/2 activation in microglia. In contrast, we found increased p38 MAPK and ERK1/2 phosphorylation after thrombin injection, and decreased phosphorylation of these two MAPKs following administration of minocycline. Interestingly, minocycline reduced thrombin-induced ERK1/2 activation more markedly than that of p38 MAPK. In addition, our recent study also showed that inhibitors of p38 MAPK and ERK1/2 suppressed thrombin-induced transcriptional expression of iNOS from microglia, but had no effect on that of IL-1β (Choi et al. 2003b). In contrast, the current study revealed that minocycline inhibited transcriptional expression of, not only iNOS, but also IL-1β. Taken together, these results indicate that ERK1/2 and/or IL-1β may be regulated by different intracellular mechanisms in response to minocycline under our current experimental conditions.
Recently, we have shown that thrombin is neurotoxic to dopaminergic neurons in mesencephalic cultures in the absence of microglia (Choi et al. 2003b). Thus, it is speculated that thrombin could directly damage nigral dopaminergic neurons in vivo and damaged neurons could send signals that activate microglia. This microglia-mediated effect would be a secondary event in addition to the direct neurotoxicity of thrombin on dopaminergic neurons in vivo. However, this possibility can be eliminated, at least in part, by the present results that there is no substantial loss of nigral dopaminergic neurons as determined by western blot analysis of TH in the SN at 4, 8, 12 and 24 h after intranigral injection of thrombin (Fig. 2), indicating that microglial activation occurred before the lesion of nigral dopaminergic neurons.
In summary, the present study showed that thrombin-induced degeneration of dopaminergic neurons was mediated by oxidative stress derived from activated microglial NADPH oxidase. Our data suggests that targeting this enzyme may be beneficial for novel treatments of oxidative stress-mediated neurodegenerative conditions such as PD.
This work was supported by funds from KOSEF (BDRC), grant no. 1999-2-210-002-5 from KOSEF and the Neurobiology Research Program from the Korean Ministry of Science and Technology, and a grant from the BRC of the 21st Century Frontier Research Program from the Korea Ministry of Science and Technology.