[1] 10w Microglial TNF-a Mediates Enhancement of Dopaminergic Degeneration by Brain Angiotensin

[1] 228w In vitro and in vivo models of Parkinson's disease were used to investigate whether TNF-a plays a major role in the enhancement of the microglial response and dopaminergic degeneration induced by brain angiotensin hyperactivity. Treatment of primary mesencephalic cultures with low doses of the neurotoxin MPP 1 induced a significant loss of dopaminergic neurons, which was enhanced by cotreatment with angiotensin II and inhibited by TNF-a inhibitors. Treatment of primary cultures with angiotensin induced a marked increase in levels of TNF-a, which was inhibited by treatment with angiotensin type-1-receptor antagonists, NADPH-oxidase inhibitors and NFK-b inhibitors. However, TNF-a levels were not significantly affected by treatment with angiotensin in the absence of microglia. The microglial origin of the angiotensin-induced increase in TNF-a levels was confirmed using dopaminergic (MES 23.5) and microglial (N9) cell lines. Inhibition of the microglial Rho-kinase activity also blocked the AII-induced increase in TNF-a levels. Treatment of the dopaminergic cell line with TNF-a revealed that NFKb activation mediates the deleterious effect of microglial TNF-a on dopaminergic neurons. Treatment of mice with MPTP also induced significant increases in striatal and nigral TNF-a levels, which were inhibited by angiotensin type-1-receptor antagonists or NFK-b inhibitors. The present results show that microglial TNF-a plays a major role in angiotensin-induced dopaminergic cell death and that the microglial release of TNF-a is mediated by activation of angiotensin type-1 receptors, NADPHoxidase, Rho-kinase and NFK-b.

[1] 294w T he renin-angiotensin system (RAS) was classically believed to act as a circulating system and to regulate blood pressure. However, many tissues possess local (tissue or paracrine) RAS (Ganong et al., 1994;Re, 2004) that contain the same components as the circulating RAS, and that are far more important than the circulating angiotensins for regulating tissue functions and degenerative processes. The activity of angiotensin II (AII), which is the most important effector of the RAS, is mediated by two main cell receptors, AII type 1 and 2 receptors (AT1 and AT2) (Jones et al., 2008;Oro et al., 2007). The AT1 receptor mediates most of the classical actions of AII, whereas AT2 receptors appear to antagonize AT1 receptors. Angiotensin interacts with AT1 receptors acti-vating the NADPH-oxidase complex, which in turn mediates several of the key aspects of oxidative stress and inflammation involved in major degenerative diseases. In animal models of Parkinson's disease (PD), hyperactivation of the local RAS exacerbates the inflammatory microglial response, oxidative stress and dopaminergic degeneration, all of which are inhibited by angiotensin receptor blockers (see for review Labandeira-Garcia et al., 2013;Mertens et al., 2009;Wright and Harding, 2012). Increased RAS activity has also been observed in the substantia nigra in animal models of aging, menopause and chronic cerebral hypoperfusion (Rodriguez-Perez et al., 2010, 2012, 2013;Villar-Cheda et al., 2012), i.e., in animals that are also more vulnerable to dopaminergic degeneration. Recent studies have shown that AII-induced activation of the NADPH-oxidase complex (Joglar et al., 2009;Rodriguez-Pallares et al., 2008;Zawada et al., 2011) and activation of the microglial RhoA/Rho kinase pathway (Rodriguez-Perez et al., 2013;T€ onger et al., 2012;Villar-Cheda et al., 2012) are involved in AII-induced oxidative stress and neuroinflammation. However, the mechanisms involved in the AII-enhanced neuroinflammatory response and dopaminergic degeneration have not been totally clarified.

[2] 171w Tumor necrosis factor a (TNF-a) is a prototypic proinflammatory cytokine that has been shown to play a major role in several neurodegenerative disorders (Fillit et al., 1991;Hofman et al.,1989;Mogi et al., 1994). Data from different experimental models of PD suggest that dopaminergic neurons are extremely sensitive to TNF-a (Harms et al., 2012;McGuire et al., 2001;Qin et al., 2007). In addition, increased levels of TNF-a have been observed in the cerebrospinal fluid, blood and post-mortem brains of PD patients, and a major role for TNFa in PD has been suggested (Scalzo et al., 2009;Verani et al., 2010). However, it is not known whether TNF-a plays a major role in the enhancement of dopaminergic neuron death induced by AII/AT1 hyperactivity. In the present study, we used in vitro (primary mesencephalic cultures, dopaminergic neuron, and microglial cell lines) and in vivo (MPTP lesioned mice) models of PD to investigate whether TNF-a plays a major role in the AII-induced enhancement of dopaminergic neuron death and to help elucidate the possible mechanisms involved in this process.

[1] 179w Treatment of primary cultures with AII (50, 100, or 150 nM) induced a significant increase in levels of TNF-a (Fig. 2A). The AII-induced increase was blocked by treatment with the AT1 receptor antagonist ZD-7155, but not by the AT2 receptor antagonist PD-123319 (Fig. 2B). No significant change in TNF-a levels was observed after treatment with ZD-7155 alone or PD-123319 alone (Fig. 2C). Consistent with this, TNF-a levels were significantly increased by the AT1 receptor agonist L-162313 but not the AT2 receptor agonist CG-42112A (Fig. 2D). The AII-induced increase in TNF-a levels was also blocked by the inhibitor of NADPHoxidase activity apocynin, which indicates the involvement of NADPH-derived superoxide in the AII-induced TNF-a increase (Fig. 3A). The AII-induced increase in TNF-a was blocked by the NFK-b inhibitor PDTC (Fig. 3B), which was confirmed by administration of a second NFK-b inhibitor SS (Fig. 3C). Simultaneous treatment of cultures with NFK-b inhibitors and the AT2 receptor antagonist PD-123319 did not affect TNF-a levels. However, co-treatment with NFK-b inhibitors and the AT1 receptor antagonist ZD-7155 led to a further decrease in TNF-a levels.

[2] 382w AII does not Increase TNF-a Levels in the Absence of Microglia: Role of the Microglial ROCK and NFK-b In primary cultures treated with LME to eliminate the microglial population, administration of AII did not induce any significant increase in the levels of TNF-a. In addition, TNF-a levels were not significantly affected by treatment with AT1 or AT2 receptor antagonists (Fig. 4A). The lack of a significant effect of AII treatment in the absence of microglia was confirmed by using the dopaminergic cell line (MES 23.5). Treatment of dopaminergic neurons with AII did not induce significant changes in TNF-a levels. Consistent with this, TNF-a levels were not significantly affected by treatment with AT1 or AT2 antagonists, or with the NADPH-oxidase inhibitor apocynin (Fig. 4B,C). The microglial origin of AII-induced increase in TNF and the role of the microglial NFK-b in this effect were confirmed using cultures of the N9 microglial cell line (Fig. 5). Treatment of the microglial cell line with AII significantly increased the levels of TNF-a, which was FIGURE 2: Effects of treatment with angiotensin II (AII) and angiotensin receptor agonists and antagonists on TNF-a levels in primary mesencephalic cultures. A: Effects of treatment with different doses of AII on TNF-a levels. B: The increase in TNF-a levels induced by AII (100 nM) was significantly inhibited by simultaneous treatment with the AT1 receptor antagonist ZD 7155 but was not significantly affected by simultaneous treatment with the AT2 receptor antagonist PD 123319. C: TNF-a levels were not significantly affected by treatment with ZD 7155 alone or PD 123319 alone. D: TNF-a levels were significantly increased by treatment with the AT1 agonist L-162313, but were not significantly affected by treatment with the AT2 receptor agonist CGP-42112A. The data are expressed as percentages of the TNF-a levels obtained in the respective control cultures (100%). Data are means 6 SEM. $ P < 0.05 compared with control group (untreated cells), 1 nM AII and 10 nM AII; A P < 0.05 compared with AII 50 nM; *P < 0.05 for comparison with control group; # p < 0.05 for comparison with AII alone, u P < 0.05 compared with AII1 ZD; & P < 0.05 compared with L-162313 (one-way ANOVA and Holm Sidak post hoc test). AII, angiotensin II; PD, PD 123319; ZD, ZD 7155.

[3] 11w blocked by simultaneous treatment with the NFK-b inhibitor PDTC (Fig. 5A).

[4] 72w In previous studies, we have shown the major role of the microglial RhoA/ROCK activation in the AII-induced increase in dopaminergic neuron death (for details see Rodriguez-Perez et al., 2013;Villar-Cheda et al., 2012). We therefore investigated here whether the microglial RhoA/ROCK pathway is involved in the AII-induced increase in microglial TNF-a. Treatment of the N9 microglial cell line with the ROCK inhibitor Y-27632 significantly inhibited the AII-induced increase in TNF-a levels (Fig. 5B).

[5] 112w In Vitro The dopaminergic cell line MES 23.5 was used to study the effects of TNF-a on dopaminergic neurons (Fig. 6A-F). Treatment of dopaminergic cells with MPP 1 (20 mM) induced a significant loss of cells (around 40%). Note that much lower doses of MPP 1 (0.25 mM) induced a similar decrease in the number of dopaminergic neurons in primary cultures (i.e., in the presence of glia). Treatment of dopaminergic cells with MPP 1 and TNF-a (40 ng mL 21 ) led to significant increase in cell death (around 65% loss), which was significantly inhibited by the NFK-b inhibitor PDTC. Interestingly, treatment with TNF-a alone induced a significant loss of dopaminergic cells.

[6] 290w Block MPTP-induced Increase of TNF-a in the Striatum and Substantia Nigra of Mice Treatment of mice with MPTP induced an early (i.e., 24 h after injection) and significant increase in TNF-a levels (both protein and mRNA expression) in the striatum, which decreased 72 h after MPTP injection (Fig. 7A,B). In the substantia nigra, the increase in TNF-a protein and mRNA levels was observed later (i.e., 72 h after MPTP injection; Fig. 7C,D). The MPTP-induced increase in TNF-a levels was inhibited both in the substantia nigra and striatum by treatment of mice with the AT1 receptor antagonist candesartan (Fig. 7A-D). Treatment of mice with the NFK-b inhibitor PDTC significantly reduced the increase in levels of TNF-a-mRNA and protein in the striatum and substantia nigra of mice (Fig. 8A-D). itors and angiotensin receptor antagonists on TNF-a levels in primary mesencephalic cultures. A: The AII-induced increase in TNFa levels was significantly inhibited by simultaneous treatment with the NADPH-oxidase inhibitor apocynin. B, C: Treatment with AII induced a significant increase in TNF-a levels, which was significantly inhibited by simultaneous treatment with NFK-b inhibitors. TNF-a levels were further decreased by the AT1 receptor antagonist ZD 7155 but not by simultaneous treatment with the AT2 receptor antagonist PD 123319. The data are expressed as percentages of the TNF-a levels obtained in the respective control cultures (100%). Data are means 6 SEM. *P < 0.05 for comparison with control group (untreated cells); # P < 0.05 for comparison with AII alone, u P < 0.05 compared with AII1 PDTC, & P < 0.05 compared with AII1PDTC1ZD, $ P < 0.05 compared with AII1SS1ZD (one-way ANOVA and Holm Sidak post hoc test). AII, angiotensin II; APO, apocynin; PD, PD 123319; PDTC, ammonium pyrrolidinedithiocarbamate; SS, sulfasalazine; ZD, ZD 7155.

[1] 475w The present results show that microglial TNF-a plays a major role in AII-induced dopaminergic cell death, and that AT1 receptors mediate the microglial release of TNF-a, via NADPH-oxidase, ROCK activation and NFK-b translocation. Although a local RAS has been observed in different areas of the brain, the role of angiotensin in the nigrostriatal system is of particular interest. An important interaction between dopamine and angiotensin receptors has recently been demonstrated, particularly regarding the regulation of renal sodium excretion and cardiovascular function (Gildea, 2009;Khan et al., 2008;Zhen et al., 2006). In peripheral cells, dopamine and angiotensin systems directly counteract each other (Gildea, 2009;Padia et al., 2012), and we have recently shown similar counterregulatory mechanisms in the rodent striatum and substantia nigra (Labandeira-Garcia et al., 2013;Villar-Cheda et al., 2010). Our data suggest that a decrease in dopaminergic activity (e.g., initial stages of dopaminergic degeneration or aging) may induce a compensatory upregulation of local RAS function in both dopaminergic neurons and glia. The resulting overactivation of the RAS may exacerbate the microglial inflammatory response and produce oxidative stress, further contributing to the progression of dopaminergic neuron loss (Labandeira-Garcia et al., 2013). Previous studies have shown that activation of AT1 receptors induces NADPH-oxidase-derived superoxide and subsequent enhancement of mitochondrial-derived ROS, which leads to oxidative stress (Rey et al., 2007;Rodriguez-Pallares et al., 2012;Zawada et al., 2011). AT1 and AT2 receptors and the NADPH-oxidase complex have been found in both dopaminergic neurons and glial cells of rodents (Grammatopoulos et al., 2007;Joglar et al., 2009;Rodriguez-Pallares et al., 2008) and primates (Garrido-Gil et al., 2013), and the above mentioned AII-induced activation of the NADPH-oxidase complex may occur in neurons and/or glial cells. However, experiments with neuron-enriched cultures showed that microglial activation and microglial NADPHoxidase derived superoxide are essential for the AII-induced enhancement of the neuronal death triggered by low doses of neurotoxins (Joglar et al., 2009;Rodriguez-Pallares et al., 2008). The present study shows that microglia-derived TNFa plays a major role in AII-induced enhancement of dopaminergic cell death, and that activation of microglial NADPHoxidase is involved in the AII-induced increase in microglial TNF-a. Furthermore, we have recently shown that AIIinduced activation of the microglial RhoA/ROCK pathway also plays a major role in dopaminergic neuron death (Villar-Cheda et al., 2012). It is known that RhoA/ROCK is an important regulator of the actin cytoskeleton, which is particularly important for migration of inflammatory cells, including microglia (Barcia et al., 2013;Yan et al., 2012), into inflamed areas (Greenwood et al., 2003;Honing et al., 2004). It has been shown that during activation of inflammatory cells Rho/ROCK induces changes in the actin cytoskeleton that result in process retraction, cell spreading and changes in cell motility, all of which are characteristic of activation of inflammatory cells such as microglia (Bernhart et al., 2010). The present study has also revealed that AIIinduced activation of ROCK is involved in microglial release of TNF-a.

[2] 326w The present in vivo and in vitro experiments show that NFK-b mediates the increase in TNF-a levels. This is consistent with our observation that both NADPH-oxidase and ROCK inhibition block the AII-induced increase in TNF-a levels, because previous studies in renal and vascular cells have shown that both NADPH-oxidase derived ROS (Han et al., 1999;Pan et al., 2009) and RhoA/ROCK activation (Cui et al., 2006;Kobori et al., 2011) may act as signal transduction messengers for several transcription factors, including NFK-b. Although the exact relationship between ROCK activation and NADPH oxidase activation is not clear, it has been suggested that there may be some interaction, as ROCK inhibitors have been shown to suppress AII-induced NADPH oxidase activation in some types of cells (Budzyn et al., 2006;Hiroki et al., 2004). However, some data suggest that AII may activate ROCK independently of NADPH oxidase in vascular smooth muscle cells (Ohtsu et al., 2006). The present data suggest that the following pathway occurs in microglial cells: AII/AT1, NADPH/ROS and ROCK activation, NFK-b translocation and TNF-a production. In primary cultures treated with AII, we observed that simultaneous inhibition of NFK-b and AT1 receptors produces significantly lower levels of TNF-a than those observed after inhibition of NFK-b alone or AT1 receptors alone. This suggests that AT1 may also induce TNF-a by complementary NFK-b-independent pathways, or that NFK-b may also induce TNF-a by complementary AT1-independent pathways. It is also possible that inhibition of NFK-b is partially counteracted by AT1 receptor upregulation, which is blocked by simultaneous treatment with NFK-b and AT1 receptor inhibitors. Simultaneous treatment with the NFK-b inhibitor and AT1 receptor antagonists produced lower levels of TNF-a than control levels. It is known that primary cultures contain endogenous AII and there is a basal level of activation in cultured microglia. This may induce basal levels of TNF-a, which could be reduced by the treatment with NFK-b inhibitors plus AT1 receptor antagonists but not by inhibition of NFK-b alone or AT1 receptors alone.

[3] 105w The present findings demonstrate that microglial TNFa production is induced by microglial NFK-b activation. However, NFK-b has also been observed in noninflammatory cells, including neurons (Park and Bowers, 2010), and several studies in different cell types have shown that TNF-a may stimulate NFK-b activation (Park and Bowers, 2010;Tarabani and Schwaninger, 2004). The present results suggest that neuronal NFK-b may also be involved in the degenerative process. Microglial TNF-a (together with low doses of neurotoxins such as MPP 1 or other neuronal insults) may induce transcription of neuronal NFK-b (Sen and Packer, 1996), which may activate genes involved in neuronal death and further neuron-induced microglial activation.

[4] 540w A large amount of data from PD models and PD patients has shown that TNF-a is a key neuroinflammatory mediator of neurotoxicity and neurodegeneration in PD (Leal et al., 2013;McCoy et al., 2011). Although no robust genetic association between TNF-a and development of PD has been clearly demostrated in PD patients, several studies have revealed TNF-a gene polymorphisms associated with risk of PD (Bialecka et al., 2008;Wu et al., 2007). Furthermore, increased levels of TNF-a have been observed in the cerebrospinal fluid, blood and post-mortem brains of PD patients (Scalzo et al., 2009;Verani et al., 2010). Data from different experimental models of PD suggest that dopaminergic neurons are extremely sensitive to TNF-a (McCoy et al., 2011;McGuire et al., 2001;Qin et al., 2007) and that TNF-a enhances dopaminergic degeneration in 6hydroxydopamine (Harms et al., 2011;Mogi et al., 1999) and MPTP models of PD (Ferger et al., 2004;Sriram et al., 2006). In vitro, administration of TNF-a has been shown to be toxic to dopaminergic neurons (McGuire et al., 2001;Sriram et al., 2006), which is consistent with that observed in the present experiments. In vivo, most reports show that TNF-a induces degenerative changes in dopaminergic neurons (see above). However, some controversial findings have also been reported, which can probably be attributed to different aspects of the methodology (for review see Leal et al., 2013). First, the damage induced by TNF-a administration may be negligible compared with that induced by administration of too high doses of MPTP or 6-OHDA leading to non significant increase in dopaminergic cell loss. Secondly, the increase in striatal or nigral TNF-a levels may be non significant 1-3 weeks after the neurotoxin injection (i.e., when the dopaminergic neuron loss is complete and is usually analyzed). In the present study, we observed that MPTP injection induced an early (24 h in the striatum, 72 h in the nigra) increase in TNF-a levels that decreased thereafter to control values. Early microglial activation was also observed after a single MPTP injection in several previous studies (Joglar et al., 2009;Wu et al., 2003). This suggests that the observed increase in microglial activation and in TNF-a levels contributes to dopaminergic degeneration and it is not a consequence of the dopaminergic cell death. Previous studies have shown that a single injection of 30-40 mg kg 21 leads to non significant or low DA neuronal death (about 10-12% decrease) 7 days after injection, and that in mice treated with higher doses (e.g., 4 3 20 mg, every 2 h, acute model; 5 3 30 mg day 21 of MPTP, subacute model) the active phase of degeneration begins around 20 h after the first MPTP injection, continues during the following days and is complete 7 days after the last injection (Jackson-Lewis et al, 1995). This is consistent with the suggestion that long term or chronic (or acute but very high) increase in TNF-a is necessary to exert clear toxic effects on the nigral neurons (Leal et al., 2013). In the present study, we observed that the AT1 antagonist candesartan inhibits the MPTP-induced increase in TNF-a levels in the nigra and striatum, and previous studies have shown that chronic treatment with candesartan decreased the dopaminergic neuron loss induced by several MPTP injections (Grammatopoulos et al., 2007;Joglar et al., 2009).

[5] 112w In conclusion, the present results demonstrate an important functional interaction between AII and TNF-a and also that microglial TNF-a is a major mediator of AII-induced dopaminergic neuron degeneration. Anti-TNF-a and anti-NFK-b signaling have been suggested as targets for drug development in PD and other brain diseases with a major neuroinflammatory component, although the results of clinical trials have been disappointing. The present study shows that brain RAS modulates TNF-a levels, which can be reduced by AT1 antagonists and ROCK inhibitors. This is of particular interest because both types of drugs are currently used in clinical practice to treat vascular and renal diseases in which oxidative stress and inflammation play a major role.

[1] 158w In vitro and in vivo experiments were carried out to study the possible role of TNF-a in the enhancement of dopaminergic neuron death by RAS hyperactivity. Primary mesencephalic cultures (i.e., neuron-glia cultures) were used to study the role of TNFa in the AII-induced increase in dopaminergic neuron death triggered by low doses of MPP 1 , and to help clarify the possible mechanisms involved in these effects. The role of microglia and neurons were studied separately by using primary cultures lacking microglial cells (i.e., primary cultures treated with L-leucine methyl ester, LME), the murine N9 microglial cell line, and the MES 23.5 dopaminergic neuron cell line. In a second series of experiments, adult C57BL-6 male mice were treated with the dopaminergic neurotoxin MPTP alone, or MPTP and the AT1 receptor antagonist candesartan, or MPTP and NFK-b inhibitors to study the effects on levels of TNF-a in the substantia nigra and striatum and to confirm major in vitro observations.

[2] 97w The murine N9 microglial cell line was provided by Dr. Paola Ricciardi-Castagnoli (Singapore Immunology Network, Agency for Science, Technology and Research (A*STAR), Singapore, Singapore). The N9 cells were cultured in Roswell Park Memorial Institute medium (RPMI 1640; Invitrogen, 21875-091) supplemented with 5% FBS, 2 mM L-Glutamine (Sigma, G6392), 100 U mL 21 penicillin and 100 lg mL 21 streptomycin, and the cultures were maintained at 37 C, 95% air and 5% CO 2 in a humidified incubator (Righi et al., 1989). The cells were then seeded into 12-well plates with 0.5 3 10 6 cells/well for analysis.

[3] 123w Dopaminergic MES23.5 cells, a gift from Dr. Wei-dong Le (Baylor College of Medicine, TX), were cultured in DMEM/F12 containing Sato's components growth medium supplemented with 2% FBS, 100 U mL 21 penicillin and 100 mg mL 21 streptomycin at 37 C in a humidified CO 2 incubator (5% CO 2 , 95% air) (Crawford et al., 1992). For experiments, MES23.5 cells were plated at a density of 0.5 3 10 5 /cm 2 onto 35-mm plastic dishes, glass coverslips or 96well plates previously coated with poly-L-ornithine (P-4638, Sigma; 10 mg mL 21 ). Cells were stimulated to enhance differentiation by adding dibutyryl-cAMP (D0627, Sigma; 1 mM) to the supplemented growth medium and they were grown to 80% confluence prior to starting any treatment.

[4] 192w Total ribonucleic acid (RNA) was extracted from the nigral and striatal regions with Trizol (Invitrogen, Paisley, Scotland, UK) according to the manufacturer's instructions. The concentration of RNA was estimated using a Nanoquant plate and Infinite M200 multiwell plate reader (TECAN, Salzburg, Austria). Total RNA (2.5 mg) was reverse-transcribed to complementary DNA (cDNA) with deoxynucleotide triphosphate (dNTP), random primers, and Moloney murine leukemia virus reverse transcriptase (MMLV; 200 U; Invitrogen). The relative levels TNF-a messenger RNA (mRNA) were examined by real-time PCR. Experiments were performed with a real-time iCycler TM PCR platform (Bio-Rad, Hercules, CA). GAPDH was used as a housekeeping gene and was amplified in parallel with the genes of interest. The data were evaluated by the deltadelta Ct method (2-DDCt) where Ct is the cycle threshold. The expression of each gene was determined relative to the housekeeping transcripts. Forward (F) and reverse (R) primers were designed for each gene by use of Beacon Designer software (Bio-Rad). Primer sequences for mouse TNF-a were F 5 0 -CGTGGAACTGGCAGAA GAG-3 0 ; R 5 0 -ACAAGCAGGAATGAGAAGAGG-3 0 ; for mouse b-actin, F 5 0 -FTCGTGCGTGACATTA AAGAG -3 0 , R 5 0 -TGCCAC AGGATTCCATACC-3

[5] 68w All data were obtained from at least three independent experiments and were expressed as mean 6 SEM. Multiple comparisons were analyzed by one-way ANOVA followed by a post hoc Holm-Sidak test. The normality of populations and homogeneity of variances were tested prior to each ANOVA. Differences were considered statistically significant at P < 0.05. Statistical analyses were carried out with SigmaStat 3.0 from Jandel Scientific (San Rafael, CA).

[1] 190w Ventral mesencephalic tissue was dissected from rat embryos of 14 days gestation (E14). The tissue was incubated in 0.1% trypsin (Sigma), 0.05% DNase (Sigma), and DMEM (Gibco, Invitrogen) for 20 min at 37 C, and was then washed in DNase/DMEM and mechanically dissociated. The resulting cell suspension was centrifuged at 50g for 5 min, the supernatant was carefully removed, and the pellet was resuspended in 0.05% DNase/DMEM to the final volume required. The number of viable cells in the suspension was estimated by acridine orange/ethidium bromide staining, and cells were plated onto 35-mm culture dishes (Falcon, Becton Dickinson, Franklin Lakes, NJ) previously coated with poly-L-lysine (100 mg mL 21 ; Sigma) and laminin (4 mg mL 21 ; Sigma). The cells were seeded at a density of 1.5 3 10 5 cells cm 22 and maintained under control conditions [DMEM/HAMS F12/(1:1) containing 10% fetal bovine serum (FBS; Biochrom KG, Berlin, Germany)]. The cell cultures were maintained in a humidified CO 2 incubator (5% CO 2 ; 37 C) for 8 days in vitro (DIV; see below); the medium was totally removed on day 2 and replaced with fresh culture medium.

[2] 40w To obtain cultures lacking microglial cells, L-leucine methyl ester (LME; 1.5 mM; Sigma) was added 48 h after seeding the cells and was maintained in the cultures for 72 h to deplete microglia, as described by Gao et al. (2008).

[3] 395w In preliminary experiments, primary mesencephalic cultures were treated with different doses of AII (1,10,50,100,150 nM;Sigma) to study their effects on levels of TNF-a. The dose of 100 nM AII was chosen for the following experiments. A series of primary cultures was used to study the possible involvement of TNF-a in cell loss induced by the DA neurotoxin MPP 1 and AII. Cultures were exposed on 4 DIV to MPP 1 alone (0.25 mM for primary cultures or 20 mM for the MES 23.5 dopaminergic cell line; Sigma) or to MPP 1 plus AII (100 nM) for a further 4 days, or they were treated with the TNF-a inhibitor TAPI (TAPI 2 acetate salt; 50 mM; Sigma) for 30 min before treatment with MPP 1 and AII. The dose of TAPI was selected on the basis of doses suggested in several previous studies (Crowe et al., 1995;Fr eour et al., 2009;Kruse et al., 2004). The cultures were fixed with 4% paraformaldehyde in DPBS (pH 7.4) for 20 min, and endogenous peroxidase activity was quenched by incubation for 5 min with 3% H 2 O 2 in DPBS. The cultures were then preincubated with a blocking solution containing 10% normal serum in DPBS with 1% BSA and 0.3% Triton X-100 (Sigma) for 1 h. The cultures were then incubated at 4 C with mouse anti-TH (1:30,000; Sigma) before being washed and incubated for 1 h with biotinylated horse anti-mouse (1:500; Vector, Burlingame, CA). Finally, the cultures were washed again and incubated for 1 h with avidin-biotin-peroxidase complex (1:500; Vector). Labeling was revealed with 0.04% H 2 O 2 and 0.05% 3,3 0 -diaminobenzidine (Sigma) as a chromogen. Cells were observed by phase-contrast microscopy (Nikon Eclipse inverted microscope) and counted in five randomly chosen longitudinal and transverse microscopic fields along the diameter of the culture dish away from the curved edge by an operator who was blind to the treatment condition. The microscopic field was defined by a 0.5 3 0.5-cm reticule (1.25 cm 2 ). The average number of TH-positive cells in a control culture dish was 3644 6 264. The final results were obtained from at least three separate experiments, with a minimum sample size of four wells per group and per run. The results were expressed as percentages of the counts of the control group in the same batch to counteract possible variations among batches.

[4] 182w A second series of cultures was used to identify the cells and major mechanisms involved in AII-induced TNF-a production. Primary cultures or cultures of the microglial cell line (N9) or the dopaminergic cell line (ME 23.5) were treated with AII (100 nM) or the AT1 receptor agonist L-162313 (1 mM; Sigma), or the AT2 receptor agonist CGP-42112A (200 nM; Sigma), or AII and (16 h before adding AII) the AT1 receptor antagonist ZD-7155 (1 mM; Sigma), or AII and the AT2 receptor antagonist PD-123319, or ZD-7155 alone, or PD-123319 alone, or AII and apocynin (an inhibitor of NADPH oxidase; 1 mM; Fluka, Buchs, Switzerland; 1 lM; Sigma), or AII and the NFK-b inhibitors PDTC (ammonium pyrrolidinedithiocarbamate; 50 mM; Sigma) or SS (Sulfasalazine; 1.25 mM; Sigma), or AII1PDTC or SS1ZD-7155, or AII1PDTC or SS and PD-123319, or AII and the Rho-kinase inhibitor Y-27632 (30 mM; Sigma). In addition, the MES 23.5 dopaminergic cell line was treated with MPP 1 (20 mM; Sigma), or MPP 1 and TNF-a (40 ng mL 21 ; Sigma), or MPP 1 and TNF-a and PDTC, or TNF-a alone.

[5] 191w Male C57BL-6 mice (Charles River, France) weighing 20-25 g (i.e., 8-weeks old) were used for in vivo experiments. The mice were divided into four groups (A-D). Mice in group A (n 5 5) were used as normal (i.e., non-lesioned) controls, and received treatment with vehicle. Mice in group B (n 5 5) were injected with a single dose of MPTP (Free base, Sigma; 30 mg kg 21 , intraperitoneally; in saline). Mice in group C (n 5 5) were injected with MPTP as in group-B mice, but received oral treatment with the AT1 receptor antagonist candesartan (1 mg/kg/day; AstraZeneca; orally in "Nocilla" hazelnut-cream spread; Nutrexpa, Barcelona, Spain) from 2 weeks before MPTP treatment until they were killed. Mice in group D (n 5 5) received intraperitoneal injection of MPTP as above and simultaneous treatment with the NFK-b inhibitor PDTC (Sigma; 150 mg/kg/day; Zhao et al., 2013; orally in "Nocilla" hazelnut-cream spread, from 2 weeks before MPTP treatment until they were killed). Twenty-four or 72 h after treatment, mice were killed and the brains were processed for analysis of striatal and nigral levels of TNF-a by ELISA and RT-PCR studies (see below).

[6] 114w Cultured cells (primary mesencephalic cultures, MES23.5 cells, N9 cells) and tissue from mice striatum and ventral midbrain were homogenized in RIPA buffer containing protease inhibitor cocktail (P8340, Sigma) and PMSF (P7626, Sigma). The homogenates were centrifuged, at 12,000g for 20 min at 4 C, and the protein concentrations were determined by the Bradford protein assay. The levels of TNF-a were quantified with mice or rat (culture cells)-specific enzyme-linked immunosorbent assay (ELISA) kits according to the manufacturers' instructions (Rat TNF-a from Diaclone, 865.000.192; Mouse TNF-a from Diaclone,860.040.192). The TNF-a contents in the brain and culture samples were obtained in pg per milliliter protein and expressed as percentages of the contents in the control group samples.

[7] 79w Treatment of primary cultures with low doses of the neurotoxin MPP 1 induced a significant loss of TH-ir neurons (about 40%; Fig. 1A-E). Co-treatment with AII (MPP 1 1100 nM AII) significantly enhanced the loss of TH-ir neurons from cultures (to about 60%). This is consistent with our previous findings (see Joglar et al., 2009 for details). Treatment with the TNF-a inhibitor TAPI (MPP 1 1AII1TAPI) induced a significant reduction of the loss of TH-ir neurons (to about 25%).