Address correspondence and reprint requests to A. Machado, Departamento de Bioquímica, Bromatología, Toxicología y Medicina Legal, Universidad de Sevilla, c/Prof García González s/n, 41012 Sevilla, Spain. E-mail: firstname.lastname@example.org
It is becoming widely accepted that the inflammatory response is involved in neurodegenerative disease. In this context, we have developed an animal model of dopaminergic system degeneration by the intranigral injection of lipopolysaccharide (LPS), a potent inductor of inflammation. To address the importance of the inflammatory response in the LPS-induced degeneration of nigral dopaminergic neurones, we carried out two different kinds of studies: (i) the possible protective effect of an anti-inflammatory compound, and (ii) the effect of the intranigral injection of inflammatory cytokines (TNF-α, IL-1β and IFN-γ) on dopaminergic neurones viability. Present results show that dexamethasone, a potent anti-inflammatory drug that interferes with many of the features characterizing pro-inflammatory glial activation, prevented the loss of catecholamine content, Tyrosine hydroxylase (TH) activity and TH immunostaining induced by LPS-injection and also the bulk activation of microglia/macrophages. Surprisingly, injection of the pro-inflammatory cytokines failed to reproduce the LPS effect. Taken together, our results suggest that inflammatory response is implicated in LPS-induced neurodegeneration. This damage may be due, at least in part, to a cascade of events independent of that described for TNF-α/IL-1β/IFN-γ.
Despite intensive research into the cause of idiopathic Parkinson's disease (PD), its aetiology remains a mystery. Furthermore, the initial degeneration of dopaminergic neurones that occurs at early stages of the disease may activate secondary phenomena, aggravating the pathologic process and accounting for the continuous progression of the disease. Some cases of PD are associated with head trauma and encephalitis, suggesting that the inflammatory process could be a component of this disease. The glial reaction observed in neurodegenerative disorders is generally considered to be simply a consequence of nerve cell death. However, recent evidence suggests that glial reaction might be also involved in the evolution of the disease. This is supported by several works describing different kinds of inflammatory features present in parkinsonian brains: a dramatic proliferation of reactive ameboid macrophages and microglia in the substantia nigra (SN; McGeer et al. 1988a, 1988b; Akiyama and McGeer 1989), along with the increase in glial cells expressing different pro-inflammatory cytokines such as tumour necrosis factor (TNF)-α, interleukin (IL)-1β, and interferon (IFN)-γ (Boka et al. 1994; Mogi et al. 1994; Hunot et al. 1999). In this context, we have developed an animal model of PD by the injection of lipopolysaccharide (LPS), a potent inductor of inflammation (Benveniste 1992; Lieberman et al. 1989), into the SN of rat (Castaño et al. 1998; Herrera et al. 2000). LPS induced an acute inflammatory response with a strong macrophage/microglial reaction in the SN, and an impairment of dopaminergic markers. This structure was far more sensitive than the striatum to the inflammatory stimulus. Moreover, only the dopaminergic neurones of the SN were affected, with no detectable damage to either GABAergic or serotoninergic neurones, suggesting the inflammatory reaction could be involved in the degenerative process of dopaminergic neurones. Neurotoxicity after LPS intraparenchymal injection has also been reported by other authors (Kim et al. 2000; Liu et al. 2000; Lu et al. 2000).
Glucocorticoids are potent anti-inflammatory drugs that have long been used in clinical neurology for the treatment of brain inflammation (Anderson and Cranford 1979; Norris and Hachinski 1986) and spinal cord injury (Bracken et al. 1990). So, dexamethasone prevents the induction of cyclooxygenase (COX)-2 mRNA and prostaglandins in the lumbar spinal cord following intraplantar injection of Freund's complete adjuvant, in parallel with inhibition of oedema (Hay and Belleroche 1998). It is also well known that glucocorticoids have numerous effects on cells of the macrophage/microglia lineage. They are potent inhibitors of the IFN-γ-induced activation of microglia and macrophages in vitro (Loughlin et al. 1993) and down-regulate the expression of the major histocompatibility complex (MHC) class II molecules on macrophages, both in vivo and in vitro (Snyder and Unanue 1982). Moreover, dexamethasone down-regulates the axotomy- and the IFN-γ-induced MHC class II expression on rat microglia (Kiefer and Kreutzberg 1991; Loughlin et al. 1993) and reversibly inhibits microglial proliferation in vitro (Ganter et al. 1992).
In order to point out the importance of the inflammatory response in the LPS-induced degeneration of nigral dopaminergic neurones, we carried out two different kinds of studies: (i) the possible protective effect of an anti-inflammatory compound (dexamethasone), and (ii) the effect of the intranigral injection of inflammatory cytokines (TNF-α, IL-1β and␣IFN-γ) on dopaminergic neurones viability, though it has␣been suggested these compounds are involved in degenerative processes and are putative mediators of LPS-induced damage.
Materials and methods
Animals and surgery
Female Wistar rats (200–250 g) housed in our laboratory were used for these studies. Four groups of animals were established for immunohistochemistry: (i) the vehicle-injected group received a single intranigral injection of Monastral blue inert tracer [Sigma, St Louis, MO, USA; 1% in phosphate-buffered saline (PBS)] and the animals were killed 7 days later; (ii) the LPS-injected group received a single intranigral injection of LPS (2 µg; from Escherichia coli, serotype 026:B6; Sigma) disolved in vehicle and the animals were killed 7 days later; (iii) the dexamethasone-LPS-treated group received a daily subdermal dose of dexamethasone (2 mg/kg; Sigma) for 8 or 15 days, and a single intranigral injection of LPS (2 µg) disolved in vehicle at day 2, the animals being killed at the end of the dexamethasone treatment; (iv) the cytokine-injected group received a single intranigral injection of either rh-TNF-α (20, 100 or 1000 U; Bender Medsystems, Vienna, Austria), IL-1β (50, 100 or 1000 U), a combination of both, or a cocktail containing TNF-α (1000 U), IL-1β (1000 U) and IFN-γ (100 U); the animals were killed 7 days later. The rats were anaesthetized with 400 mg/kg chloral hydrate and positioned in a stereotaxic apparatus (Kopf Instruments, Tujunga, CA, USA) to conform with the brain atlas of Paxinos and Watson (1986). LPS and cytokines were dissolved in vehicle solution and 2.0 µL injected into the left SN. The injection needle was lowered through a drill hole 5.5 mm posterior, 1.5 mm lateral and 8.3 mm ventral to the bregma. The injections were delivered over a period of about 2 min and after each the needle was left in situ for an additional 5 min to avoid reflux along the injection track. Animals were decapitated and brains quickly removed. Striatum and SN were dissected for HPLC as described (Castaño et al. 1993) and frozen at − 80°C until analysis. Experiments were carried out in accordance with the Guidelines of the European Union Council (86/609/EU), following the Spanish regulations (BOE 67/8509–12, 1988) for the use of laboratory animals and approved by the Scientific Committee of the University of Sevilla.
Measurement of DA, 5-HT and their metabolites
HPLC with electrochemical detection was used for the determination of DA and its metabolites 3,4-dihydroxyphenylacetic acid␣(DOPAC), 3-metoxytyramine (3-MT) and homovanillic acid (HVA), and 5-HT and its metabolite 5-hydroxyindolacetic acid (5-HIAA). Analyses were performed as described (Castaño et al. 1993).
In vivo tyrosine hydroxylase activity
TH activity from striatal tissue was measured according to a modification of a previously published procedure (Reinhard et al. 1986). Briefly, the tissue of interest was homogenized in 30 mm Tris buffer containing 0.1% Triton X-100, pH 6.5. An aliquot of the homogenate was incubated with 2.5 nmol of tyrosine–HCl (containing 0.4 µCi/nmol of L-[ring-3,5–3H]tyrosine; NEN, Wellesley, MA, USA), 50 nmol of the cofactor 6(R)-L-erythro-5,6,7,8-tetrahydrobiopterin (Schirck Laboratories, Jona, Switzerland), 5000 units of catalase and 5 mm of dithiothreitol in 100 mm potassium phosphate, pH 6.0. The released [3H]OH was separated by an aqueous slurry of activated charcoal, and the radioactivity was determined by liquid scintillation counting.
Animals were perfused through the heart under deep anaesthesia (chloral hydrate) with 100 mL of PBS containing 10 U/mL heparin followed by 150–200 mL of 4% paraformaldehyde in phosphate buffer, pH 7.4. Brains were removed and then immersed in sucrose in PBS, pH 7.4, first in 10% sucrose for 24 h and then in 30% sucrose until sunk (2–5 days). Tissues were then frozen in isopentane at − 15°C and 25-µm sections were cut on a cryostat and mounted on gelatin-coated slides. The primary antibodies used were monoclonal mouse-derived antityrosine hydroxylase (anti-TH; Boehringer-Mannheim, Mannheim Germany; 1 : 200), antiglial fibrillary acidic protein (anti-GFAP; Chemicon, Temecula, CA, USA; 1 : 300), OX-42 (Serotec, Oxford, UK; 1 : 100) and OX-6 (Serotec; 1 : 200). Sections were incubated with monoclonal primary antibodies for 24 h. Antibody binding was detected by the avidin–biotin–peroxidase method as previously described (Venero et al. 1997).
Quantification of OX-42 and OX-6 reactive cells and areas lacking GFAP immunostaining
Number of OX-42- and OX-6-positive cells were counted using five sections per animal, and five fields per section. Both sections and fields within the sections were chosen randomly. A single number per animal was averaged. Quantification of the areas lacking GFAP immunostaining was done at the injection level by using the AnalySIS image software.
Results are expressed as mean ± SD. Means were compared. Results were analysed by the Student's t-test or one-way anova followed by the LSD test for posthoc multiple range comparisons.
Animals injected with vehicle into the left SN did not show any change in the levels of DA, 5-HT and their metabolites between the right and the left side in both SN and striatum (data not shown). Immunoreactivities were similar to those observed in the contralateral side, except for a slight OX-6 immunoreactivity and a slight up-regulation in the OX-42 staining confined both to the injection track (photographs not shown).
LPS-injected and Dexamethasone-LPS-treated groups
Effects on levels of monoamines and their metabolites in SN
The systemic treatment with dexamethasone did not change the concentrations of monoamines and their metabolites in␣the SN (Table 1). On the contrary, the single intranigral injection of LPS caused a significant decrease in the levels of DA (65.3% and 66.5% for 7 and 14 days, respectively) and its metabolites DOPAC (79.3% and 79.5% for 7 and 14 days, respectively) and 3-MT (63.6% and 69.8% for 7 and 14 days, respectively) compared to the control values (p < 0.001 in all cases). No change was found for HVA. The level of 5-HT decreased (79.8% and 78.8% for 7 and 14 days, respectively) in opposition to a 5-HIAA increase (131.3% and 115.2% for 7 and 14 days, respectively) compared to control values (p < 0.001 in both cases). Dexamethasone treatment prevented the LPS-induced damage. So, no statistically significant differences were found comparing the dexamethasone-LPS group to the dexamethasone group.
Table 1. Concentrations of DA, 5-HT and their metabolites measured in substantia nigra and striatum, and in vitro tyrosine hydroxylase activity measured in striatum after the injection of 2 µg of lipopolysaccharide into the left substantia nigra, with or without subcutaneous dexamethasone treatment
Dexamethasone + LPS
Twenty animals were injected daily with a subdermal dose of dexamethasone (2 mg/kg) for 8 or 15 days, while another 20 were injected daily subdermal saline. At day 2 of their respective treatments, all animals received a single injection of LPS (2 µg/2 µL) into the left SN. At the end of the treatments, animals were killed and the structures dissected out and processed for monoamines and TH-activity quantification, as described under Materials and methods. The structures dissected [substantia nigra (SN) and striatum] fall into four different groups: Control, the right-side structures of animals subdermally injected with saline; LPS, the left-side structures of animals subdermally injected with saline; Dexamethasone, the right-side structures of animals subdermally injected with dexamethasone; dexamethasone + LPS, the left-side structures of animals subdermally injected with dexamethasone. Numbers are expressed as ng/g wet tissue for monoamines quantification and pmol/h/µg prot. for TH activity quantification, and are mean ± SD of 10 and nine independent experiments for monoamines concentrations and TH activity, respectively. Statistical signification (calculated by using One-way anova followed by LSD test for post hoc multiple range comparisons, using α= 0.01 for all comparisons): acompared with the control group; bcompared with the dexamethasone group, p < 0.001.
Effects on levels of monoamines and their metabolites in striatum
Results are similar to those described for SN. No changes in monoamines concentrations were found after systemic treatment with dexamethasone (Table 1), but a decrease in their concentrations appeared after a single intranigral injection of LPS, with values ranging between 63.4% (DA) and 74.9% (3-MT) of control values (p < 0.001 in all cases) after 7 days and between 60.1% (DA) and 74.0% (3-MT) after 14 days. An increase was found for 5-HIAA concentration (113.0% and 113.9% for 7 and 14 days, respectively, compared to control values, p < 0.001). Dexamethasone treatment prevented the LPS-induced damage in striatum. Monoamines concentrations were then similar to those found for the dexamethasone group, except for DA levels that reached 86.1% of the dexamethasone group values (p < 0.001) after 7 days.
Tyrosine hydroxylase activity in striatum
TH activity in striatum was not affected after the systemic treatment with dexamethasone (Table 1). After LPS intranigral injection, TH activity decreased to 53.4% and 55.9% of control for 7 and 14 days, respectively. The treatment with dexamethasone reduced the loss of TH activity after LPS intranigral injection. So, TH activity values of the dexamethasone-LPS group reached 84.7% of the dexamethasone group value (p < 0.001) after 7 days treatment, and there were no statistical difference after 14 days.
Changes in tyrosine hydroxylase immunoreactivity
The number of both TH-positive cell bodies and fibres decreased in the SN after the injection of LPS (Fig. 1a). The systemic treatment with dexamethasone prevented this loss (Fig. 1c). Dexamethasone alone did not produce any effect on TH immunoreactivty (Fig. 1d).
Changes in microglial cell population were shown by the OX-42 and OX-6 immunostaining. As previously described by us (Castaño et al. 1998; Herrera et al. 2000), in control SN, OX-42-positive cells were ramified resident microglia with two or three fine processes. One week after LPS injection, the number of OX-42-positive cells was highly increased. By that time, most of these OX-42-positive cells did not show ramified but a round morphology, many of them resembling tissue macrophages (photograph not shown). Therefore, OX-42-positive cells may be both blood-borne macrophages and resident microglial cells. Dexamethasone treatment reduced the number of OX-42- and OX-6-positive cells (Fig. 2) to 42.5% and 63.8% of the LPS-treated animals, respectively (p < 0.005).
It is known that although microglia do not constitutively express MHC class II, it is up-regulated after a diverse range of insults in the rat (for review see Perry 1998) and that dexamethasone down-regulates MHC class II expression in␣microglial cells both in vitro (Loughlin et al. 1993) and in␣vivo (Kiefer and Kreutzberg 1991). After LPS injection numerous OX-6-positive cells surrounded the injection site (Fig. 1e) and dexamethasone treatment clearly reduced the area of OX-6 immunostaining around the LPS injection (Fig. 1f).
Astrocytes were stained with anti-GFAP monoclonal antibody, and were absent from the area surrounding the LPS injection site (Fig. 1g). Dexamethasone treatment did not prevent this loss of anti-GFAP reactivity induced by LPS (Fig. 1h). The area lacking GFAP immunoreactivity was 1.72 ± 0.21 and 1.68 ± 0.11 mm2 for the LPS and dexamethasone-LPS-treated animals, respectively.
Effects on levels of monoamines and their metabolites in SN and striatum
Intranigral injection of cytokines did not mimic the effects of intranigral injection of LPS on SN and striatum. No changes were found in the monoamine concentrations in either the SN or the striatum compared to the control values for any of the cytokines (rh-TNF-α, IL-1β, a combination of both at a concentrations up to 1000 U, or a cocktail of rh-TNF-α, IL-1β and 100 U of IFN-γ) assayed (data not shown).
Tyrosine hydroxylase activity in striatum
TH activity in striatum was not affected after the intranigral injection of cytokines at any of the concentrations assayed (data not shown).
Changes in tyrosine hydroxylase immunoreactivity
The number of TH-positive cell bodies and fibres did not change in the SN after the injection of the different cytokines assayed (Fig. 3a).
Microglial reaction after the treatment with cytokines was restricted to the injection track (Fig. 3b). The area devoid of GFAP-positive astrocytes found after LPS injection (Fig. 1g) was not observed after cytokine treatment (Fig. 3c).
Our study shows in vivo that dopaminergic degeneration induced by a single intranigral dose of 2 µg of LPS can␣be␣prevented by dexamethasone, a widely used anti-inflammatory glucocorticoid with broad effects on the pro-inflammatory cascade. In previous works, we described the␣loss of dopaminergic markers (catecholamine content, in␣vitro TH activity and TH immunostaining) and GFAP immunoreactivity, and the induction and activation of the microglial population after LPS challenge. It is known that the inflammation has an important role in the pathogenesis of brain disease (Beal 1995; Rothwell et al. 1996; Barone and Feuerstein 1999; Dirnagl et al. 1999; McGeer and McGeer 1999; Touzani et al. 1999; Cooper et al. 2000). At the same time, inflammation is regarded as an attractive pharmacological target, because it progresses over several days after injury and because intervention with anti-inflammatory agents may not result in intolerable side-effects (Barone and Feuerstein 1999). Treatment of the LPS-injected animals with dexamethasone could point out the implications of inflammation in the dopaminergic degenerative process.
Present results show that the bulk microglial activation found after LPS injection has been prevented by the dexamethasone treatment. LPS induced not only an increase in the proliferation of microglia but also up-regulated the expression of MHC class II molecules. Resident microglia do not constitutively express MHC class II, however, it is readily up-regulated after different kind of CNS damage, such as excitotoxic-induced degeneration (Akiyama et al. 1988), ischaemic injury (Gehrmann et al. 1992), Wallerian degeneration (Rao and Lund 1993), spreading depression (Gehrmann et al. 1993) and experimental allergic encephalomyelitis (Vass et al. 1986). The expression of MHC class II antigens on microglia is considered not only as a marker of microglial activation but is often interpreted as evidence of an antigen presentation capacity. We found that dexamethasone inhibits not only the number of OX-42-positive cells (42.5% of LPS group) but also the number of microglia/macrophages expressing MHC class II antigens as revealed by staining with OX-6 (63.8% of LPS group). These observations are in good agreement with previous works reporting that glucocorticoids down regulate MHC class II (Snyder and Unanue 1982; Kiefer and Kreutzberg 1991; Loughlin et al. 1993). In contrast, dexamethasone has not been able to prevent the lost of GFAP immunostaining.
Dexamethasone has also prevented the loss of catecholamine content, TH activity and TH immunostaining. After daily subcutaneous injections of dexamethasone (up to 15 days), dopamine and its metabolites returned to control values. So, in SN, the LPS group showed levels around 65% of the control group for DA and 3-MT and 79% for DOPAC, while in the dexamethasone-LPS group, levels of DA and its metabolites returned to 86–100% of the dexamethasone group. In striatum, a similar protection was found in monoamine levels and TH activity (TH activity increased from 53.4% of control in the LPS group to 91.7% of the dexamethasone group in the dexamethasone-LPS group). Dexamethasone protection observed could be due to an inhibition of microglial-mediated cytotoxic events. It is well known that neuronal survival increases in a number of systems when microglial activation is reduced (Thanos et al. 1993; Moore and Thanos 1996; Rogove and Tsirka 1998). As we mentioned before, glucocorticoids are potent anti-inflammatories that interfere with many of the features characterizing pro-inflammatory glial activation. So, dexamethasone inhibits LPS-induced synthesis of TNF-α (Han et al. 1990) and IL-1β (Kern et al. 1988; Kimberlin et al. 1995), and abolish LPS-stimulated release of IL-6 by astrocytes (Grimaldi et al. 1998).
We mentioned above that the activation of microglia may exert detrimental effects by the release of inflammatory molecules (such as IL-1β, TNF-α and NO), and that activation of brain glial cells with LPS stimulates the expression of NO and several cytokines, mainly TNF-α, IL-1β and IL-6 (Jeohn et al. 1998). However, we showed that LPS can cause the disappearance of TH immunoreactive neurones in the SN independently of NO (Castaño et al. 1998). It is also known that TNF-α is detected in the brain of PD patients. Taking into account these considerations, we have tried to get an insight into the inflammatory components responsible for the catecholaminergic damage induced by LPS. We have injected TNF-α, IL-1β and IFN-γ at different doses to compare their effects with those described for LPS. We expected to mimic the LPS-induced neurodegeneration by the injection of these cytokines. Interestingly, they failed to mimic LPS-induced damage: none of the dopaminergic markers studied suffered a significant decrease, the activation of microglial cells was closely confined to the injection track, and even the astrocyte population, which was not protected from LPS challenge by dexamethasone, appears similar to that of control animals after cytokine injection. The␣biological␣effects elicited by TNF-α and IL-1β, including inflammation and tissue injury, are initiated by ligand-induced formation of multiprotein receptor complexes. The signal transduction cascade of TNF-α is quite complex (for review see Madge et al. 1999). Interestingly, TNF-α and Il-1β signal transduction cascades have a point of convergence that lies upstream of the NF-κB-inducing kinase. Upon induction of this signal transduction pathway, free NF-κB eventually reaches the nucleus, where it induces transcription of many genes critical in the recruitment of leucocytes into␣inflammatory reactions, including those of adhesion molecules and cytokines (Collins et al. 1995). Taking into acount our results and the data arousing from the literature on␣LPS-activation of cytokines production, it cannot be discarted that LPS-induced damage to the nigrostriatal dopaminergic system may be due, at least in part, to a cascade of events independent of that described for TNF-α/IL-1β or IFN-γ.
In conclusion, our results show how inflammation-related nigrostriatal dopaminergic neurodegeneration can be prevented by dexamethasone. Glucocorticoids therapy has long been used in clinical neurology for the treatment of brain inflammation (Anderson and Cranford 1979; Norris and Hachinski 1986) and spinal cord injury (Bracken et al. 1990). So, our results may have implications for the treatment of neurological disease, including Parkinson's disease. Further studies are necessary for elucidating how LPS induces dopaminergic damage, and how dexamethasone is able to prevent it.
This work was supported by a grant CICYT PM98/0160. A.J.H. thanks Universidad de Sevilla, Spain, for a Beca del Plan Propio de Investigacion. We thank Dr V.H. Perry for his kind gift of IL-1β, and Mr J.P. Calero and Mr E. Fontiveros for their skillful technical assistance.