- Email: [email protected]
The mechanism of heme oxygenase-1 action involved in the enhancement of neurotrophic factor expression
Neuropharmacology 58 (2010) 321e329
Contents lists available at ScienceDirect
Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm
The mechanism of heme oxygenase-1 action involved in the enhancement of neurotrophic factor expressionq Shih-Ya Hung a, Houng-Chi Liou a, Wen-Mei Fu a, b, * a b
Pharmacological Institute, College of Medicine, National Taiwan University, Taipei 100, Taiwan Neurobiology and Cognition Center, National Taiwan University, Taipei 106, Taiwan
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 July 2009 Received in revised form 7 November 2009 Accepted 10 November 2009
Heme oxygenase-1 (HO-1) is up-regulated in response to oxidative stress and catalyzes the degradation of pro-oxidant heme to carbon monoxide (CO), iron and bilirubin. Bilirubin is a potent antioxidant and neuroprotectant. Neurotrophic factors of BDNF and GDNF also play important roles in survival and morphological differentiation of dopaminergic neurons. We have previously found that HO-1 induction by adenovirus containing human HO-1 gene (Ad-HO-1) in substantia nigra of rat increases BDNF and GDNF expression. We here further examined the possible mechanism of HO-1 action involved in the enhancement of neurotrophic factor expression. Treatment of anti-BDNF/GDNF antibody significantly enhanced dopaminergic neuronal death, whereas Ad-HO-1 co-treatment was able to antagonize the apoptosis-inducing effect of these antibodies. The confocal imaging shows that HO-1 induction appeared in dopaminergic neuron, astrocyte and microglia at 24 h after injecting Ad-HO-1. HO-1 induced-BDNF/ GDNF mRNA expression in substantia nigra was 26/21 folds of that of the contralateral Ad-injected side. The downstream product bilirubin increased GDNF expression through ERK and PI3K-Akt pathways, and also enhanced NFkB (p65 and p50) nuclear translocation in glia-enriched cultures. In addition, bilirubin also enhanced BDNF expression through similar pathway in cortical neuron-enriched cultures. We also examined the effect of another HO-1 product, CO, by using CO donor. [Ru(CO)3Cl2]2 increased neurotrophic factor expression via sGC-PKG pathway in both neuron and glia. These results indicate that the downstream products of HO-1, i.e. bilirubin and CO, modulate BDNF and GDNF expression in neuron and astrocyte. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: HO-1 Bilirubin BDNF GDNF CO
1. Introduction Parkinson's disease (PD) is a movement disorder characterized by the accelerated loss of dopaminergic neurons in the pars compacta of the substantia nigra (Hornykiewicz, 1988). Oxidative stress contributes to the cascade leading to dopamine cell degeneration in
Abbreviations: BDNF, brain-derived neurotrophic factor; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; GDNF, glia cell line-derived neurotrophic factor; HO-1, heme oxygenase-1; HO-2, heme oxygenase-2; moi, multiplicity of infection; MPPþ, 1-methyl-4-phenylpyridinium; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; [Ru(CO)3Cl2]2, tricarbonyldichlororuthenium (II) dimmer; TH, tyrosine hydroxylase; sGC, soluble guanylyl cyclase; ODQ, (1H-[1,2,4] oxadiazolo[4,3-a]quinoxalin-1-one); PKG, cGMP-dependent protein kinase. q This work was supported by grants from the National Science of Council and National Taiwan University (97R0066-53). * Correspondence to: Wen-Mei Fu, Pharmacological Institute, College of Medicine, National Taiwan University, No. 1, Sec. 1, Jen-Ai Road, Taipei 100, Taiwan. Tel: þ886 2 23123456x88319, fax: þ886 2 23417930. E-mail address: [email protected] (W.-M. Fu). 0028-3908/$ e see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropharm.2009.11.003
PD (Jenner, 2003). It has been reported that there is an up-regulation of heme oxygenase-1 (HO-1) in the substantia nigra of PD patients (Schipper et al., 1998). Heme oxygenase is an enzyme, which degrades intracellular heme to free iron, carbon monoxide (CO) and biliverdin. Biliverdin is subsequently converted to bilirubin by biliverdin reductase (BVR) and free iron is sequestered by ferritin (Tenhunen et al., 1969). HO-1, a 32 kDa cellular stress response protein (also known as Hsp32) can be rapidly induced under oxidative challenge and other noxious stimuli in the brain or other tissues (Hu et al., 2004; Le et al., 1999). In normal brain, the level of HO-1 is rather low, however, HO-2 is constitutively expressed and does not respond to environmental stress (Baranano et al., 2002; Kim et al., 2004; Schipper, 2000; Schipper et al., 1995). The end products from HO enzymatic activity, CO and bilirubin, exert many biological functions (Ryter et al., 2006). CO is an endogenous gaseous molecule, which activates soluble guanylyl cyclase (sGC) at a lower rate than nitric oxide and plays significant role in anti-apoptosis, anti-inflammation, anti-proliferation, neurotransmission and vasorelaxant (Otterbein et al., 2000, 2003).
322
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
The HO/CO system may play similar role as nitric oxide synthase (NOS)/nitric oxide (NO) system in the regulation of cellular function and communication (Marks, 1994). On the other hand, bilirubin acts as reactive oxygen species (ROS) scavenger and exerts strong antiinflammatory and neuroprotective function (Baranano et al., 2002; Dore et al., 1999). We have previously found that HO-1 induction in substantia nigra by adenovirus exerts a strong neuroprotective effect in Parkinsonian rat model. HO-1 induction in substantia nigra upregulates BDNF in dopaminergic neuron and GDNF in glia (Hung et al., 2008). In vitro treatment of glia with bilirubin or CO also exerts similar effect in up-regulating the expression of neurotrophic factors. One recent study indicates that HO-1-midiated cytoprotection is dependent on PI3K/Akt pathway (via p85a and Akt phosphorylation) to decrease ROS production and apoptosis (Pachori et al., 2007). In addition, many studies indicate that activation of the ERK1/2 pathway inhibits apoptosis and promotes cell survival and growth (McKinstry et al., 2002; Roberts and Der, 2007). Furthermore, Ollinger et al. (2007) have also demonstrated that bilirubin treatment in vitro induces the phosphorylation of MEK and ERK1/2 but not p38 or JNK1/2. Here we further investigated the signaling pathways involved in bilirubin- or CO-induced enhancement of neurotrophic factor expression. 2. Materials and methods 2.1. Reagents Bilirubin stock solution (10 mM) (Sigma, St. Louis, MO) was prepared in 0.1 M NaOH. [Ru(CO)3Cl2]2 (Sigma, St. Louis, MO) is a carbon monoxide (CO)-releasing molecule and each mole of [Ru(CO)3Cl2]2 releases approximately 0.7 mol of CO gas (Motterlini et al., 2002). [Ru(CO)3Cl2]2 was freshly dissolved in DMSO to make a 100 mM stock solution. KT5823, PDTC, ODQ, PD98059 and LY294002 were purchased from Sigma (Sigma, St. Louis, MO).
cultures were maintained in a humidified chamber at 37 C in a 5% CO2 atmosphere for 7e10 days, the medium was changed every 3-day before use. Primary rat glia cultures were prepared according to previous report with some modification (Lu et al., 2009). The cells were seeded at 1 107 on 75 cm2 flask. The culture medium for glia was DMEM with 10% FBS, 4500 mg/L glucose, 100 U/mL penicillin and 0.1 mg/mL streptomycin, which was changed every 4-day. On the last day, flasks were placed on a shaker platform and shaken at 220 rpm for 6 h at 37 C to remove oligodendrocytes and microglia. The glia-enriched cultures were then grown to confluence before use. All inhibitors were pre-treated for 30 min before the administration of bilirubin or [Ru(CO)3Cl2]3. 2.4. Intra-substantia nigra injection Wistar rats weighing approximately 350 g were used. Rats received bilateral injection of Ad-HO-1/empty adenovirus in the ipsilateral/contralateral side (5 107 PFU for each in 1 mL buffer containing 10 mM Tris pH 8.0, 2 mM MgCl2, 4% sucrose). For intra-substantia nigra injection, stereotaxic coordinates for SNpc were lateral (L) þ 2.0 mm, antero-posterior (AP, from the bregma point) - 5.3 mm and dorsiventral (DV) þ 7.8 mm (Sindhu et al., 2006). The solution was injected into the substantia nigra with a 10 mL Hamilton syringe coupled to a motorized injector (Stoelting) at a rate of 0.2 mL/min and the needle was left in situ for at least 5 min after injection. The rats were sacrificed 24 h later for the studies of real-time quantitative PCR or immunofluorescent staining. 2.5. Total RNA extraction, reverse transcription (RT) and real-time quantitative PCR Substantia nigra and striatal tissues dissected from rats were prepared as described previously (Roberts and Kapur, 1977). Total RNA was extracted from substantia nigra using TRIzol kit (MDBio Inc., Taipei, Taiwan). Single-strand cDNA was synthesized using SSIII reverse transcription reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. All cDNA samples were stored at e 20 C. Real-time quantitative PCR was performed to monitor the expression of HO-1, BDNF, GDNF and GAPDH. The TaqMan technology was used and the results were analyzed in a 7900 HT sequence detector system (Applied Biosystems, Foster City, CA). All primers and TaqMan probes were purchased from Applied Biosystems. The amplification reaction mixture (25 mL) contained 2.5 mL cDNA sample, TaqMan Universal PCR Master Mix and TaqMan primer/probe pre-mix. The thermal cycling conditions included 10 min at 95 C, proceeding with 40 cycles of 95 C for 15 s and 60 C for 1 min. The GDNF or BDNF mRNA levels were normalized to GAPDH mRNA and expressed relative to control using the DDCt method.
2.2. Preparation of the recombinant adenovirus 2.6. Immunofluorescence Replication-defective empty adenovirus and recombinant adenovirus carrying human HO-1 gene (Ad-HO-1) were kindly provided by Dr. Lee-Young Chau (Academia Sinica, Taiwan). Large scales of viral vectors were purified using CsCl ultracentrifugation as previously described (Juan et al., 2001). Virus was amplified in QBI-293 cells and the plaque forming unit (PFU) within 21 days was used for determining the titer of empty adenovirus and Ad-HO-1 stocks. 2.3. Cell cultures of midbrain, cortical neuron and astrocyte Neuron-glia co-cultures from the midbrain of E14 Wistar rats were prepared as previously described with some modification (Chen et al., 2006). Briefly, the ventral portion of the midbrain was removed in sterile ice-cold CaCl2, MgCl2 and MgSO4-free Hank balanced salt solution (GIBCO, Grand Island, NY). The tissues were cleaned, minced and mechanically dissociated by passage through a flame-polished Pasteur pipette. Dissociated cells were seeded in DMEM (GIBCO, Grand Island, NY) with 10% fetal bovine serum (FBS; Biological Industries, Grand Island, NY, Cat. No: 04-001-1A, Lot no: 115115), 4500 mg/L glucose, 100 U/mL penicillin and 0.1 mg/mL streptomycin. Cells were seeded at 3.5 105/well on poly-D-lysine-coated 48-well plates. The cultures were kept in a humidified chamber at 37 C in a 5% CO2 atmosphere. Twentyfour hours after plating, the cells were changed to MEM (GIBCO, Grand Island, NY) with 2% FBS, 2% horse serum (HS; Hyclone, Logan, UT),100 U/mL penicillin and 0.1 mg/ mL streptomycin. The Day-4 cultures were infected with Ad-HO-1 or empty adenovirus (10 moi for each) and co-treated with anti-BDNF or anti-GDNF antibody (1 mg/ mL IgG for each; Santa Cruz Biotechnology, CA) for 24 h. MPPþ at 5 nM was then added for another 48 h. Dopaminergic neurons were characterized by staining with a rabbit anti-tyrosine hydroxylase (anti-TH) antibody (1:10,000; Calbiochem Inc., San. Diego, CA). The cultures were then incubated for 1 h with the biotinylated secondary antibody, and then for 30 min with avidineperoxidase complex (ABC kit; Vector Laboratories. Burlingame, CA). Finally, the labeling was revealed by treatment with 0.01% hydrogen peroxide and 0.05% 3, 30 -diaminobenzidine (DAB; Sigma, St. Louis, MO). The total number of TH-immunopositive cells in each well was counted under inverted microscope at 200 magnification (Leica, Heidelberg, Germany). Cortical neuron cultures were obtained from cerebral cortex of E17 Wistar rats and glia cultures were obtained from P1 Wistar rats. For cortical neuron cultures, the cells were seeded at 1.2 106 cell/well on poly-D-lysine-coated 6-well plate. Twenty-four hours after plating, the cells were changed to DMEM containing 2% B27, 4500 mg/L glucose, 100 U/mL penicillin and 0.1 mg/mL streptomycin. The neuron
Brain tissue sections (30 mm thickness) were used for double-immunolabeling studies. The sections were incubated with 10% BSA and 0.1% Triton X-100 in phosphate-buffered saline (PBS) for 1 h, and then incubated overnight at 4 C with the following primary antibody: TH (1:3000; Calbiochem Inc., San. Diego, CA), HO-1 (1:600; Catalog # OSA-150F, Stressgen Biotechnologies, Victoria, British Columbia), GFAP (1:500) and CD11b (1:250; Abcam, Cambridge, MA) and p65 (1:500; Santa Cruz Biotechnology, Santa Cruz, CA) in PBS, and then with Alexa-488 or 543conjugated goat anti-rabbit or anti-mouse as the secondary antibody (1:600; Invitrogen, Carlsbad, CA). The confocal images were obtained at excitation wavelength of 488 nm and 543 nm, respectively (model SP2 TCS; Leica, Heidelberg, Germany). 2.7. Western blotting The protein expression levels of BDNF, GDNF, ERK, pERK, Akt, pAkt, p50 and p65 were determined by Western blotting. Samples obtained from cells were homogenized in RIPA buffer containing 1 mM PMSF, 25 mM leupeptin and 1 mg/mL aprotinin. Isolation of nuclear proteins from cells used the Pierce NE-PER extraction reagents (Pierce, Rockford, IL). Protein concentration was determined using the BCA protein assay kit (Pierce, Rockford, IL). Protein at 30e50 mg was separated by SDS-PAGE using a 10e15% resolving gel under reducing conditions and electrotransferred onto immobilon-P membrane (Millpore, Bedford, MA). After being blocked with 5% nonfat milk in TBS-T (0.5% Tween 20 in 20 mM Tris and 137 mM NaCl) for 1 h at room temperature, the membranes were incubated overnight at 4 C with anti-BDNF (1:3000), anti-GDNF (1:3000), anti-pAkt (1:5000), anti-p50 (1:1000), anti-p65 (1:1000) or anti-pERK (1:5000) primary antibodies (Santa Cruz Biotechnology) diluted in TBS-T. The membranes were probed with a mouse anti-GAPDH antibody as a standard (1:10,000; Abcam Inc., Cambridge, MA). The blots were then incubated for 1 h at room temperature with an HRP-conjugated secondary antibody (1:20,000; Amersham Life Science, Arlington Heights, IL). Protein bands were detected using the ECL Western Blotting Substrate (Pierce, Rockford, IL) and were estimated using the Image Analysis Program Labwork 4.5 (UVP, Inc., Upland, CA). 2.8. Statistics Results are expressed as mean SEM. Results were analyzed with one-way analysis of variance (ANOVA) and Neuman-Keuls post-hoc test to determine
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329 statistical significance between specific groups. Difference was considered significant when p < 0.05.
3. Results 3.1. HO-1 induction antagonizes BDNF or GDNF deficiency-induced dopaminergic neuronal death in midbrain neuron-glia co-cultures BDNF and GDNF enhance the survival and morphological differentiation of dopaminergic neuron in CNS (Lin et al., 1993; Spenger et al., 1995). Adenovirus vector-mediated GDNF gene therapy inhibits
323
the dopaminergic neuron loss in Parkinsonism patients (Choi-Lundberg et al., 1997). We here examined the role of BDNF and GDNF in dopaminergic neuronal survival in midbrain neuron-glia co-cultures. Fig. 1 showed that treatment of anti-BDNF or anti-GDNF antibody for 48 h to neutralize the function of endogenously released BDNF of GDNF in the cultures significantly enhanced dopaminergic neuronal death. The control IgG was used for comparison (the THþ neurons in anti-BDNF, anti-GDNF and control IgG groups were 53.7 5.3%, 54.8 4.5% and 93.0 3.1%, respectively). Co-treatment of Ad-HO-1 (10 moi) but not empty adenovirus with antibodies against neurotrophic factors antagonized dopaminergic neuronal death. Ad-HO-1
Fig. 1. HO-1 induction antagonizes dopaminergic neuronal death induced by BDNF or GDNF deficiency in midbrain neuron-glia co-cultures. Midbrain neuron-glia co-cultures derived from E14 rats were used to evaluate dopaminergic neuronal survival. The dopaminergic neurons were counted by the immunostaining with antibody against tyrosine hydroxylase. (A) Administration of anti-BDNF or anti-GDNF antibody but not control IgG (1 mg/mL) in midbrain neuron-glia co-cultures for 48 h enhanced dopaminergic neuronal death. Co-treatment with Ad-HO-1 (10 moi) but not empty adenovirus antagonized the neuronal death induced by BDNF or GDNF deficiency. The summarized result was shown in (B). Scale bar: 50 mm. Data are given as mean SEM (n ¼ 4). *p < 0.05 compared with no virus control; #p < 0.05 compared with anti-BDNF or anti-GDNF alone.
324
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
Fig. 2. HO-1 induction by Ad-HO-1 increases BDNF and GDNF expression in substantia nigra of adult rats. Ad-HO-1 (5 107 PFU) was locally injected into substantia nigra and immunofluorescence was performed in the slice 200 mm away from injection site. (A) The confocal image showed that HO-1 expressed in THþ neuron, GFAPþ astroglia and CD11bþ microglia in the midbrain of adult rat in response to Ad-HO-1 (5 107 PFU) but not empty adenovirus injection (co-localization is indicated by arrow). (B) Real-time PCR showed that mRNA levels of HO-1, BDNF and GDNF increased in midbrain 24 h after local administration of Ad-HO-1. Scale bar: 50 mm. Data are given as mean SEM (n ¼ 3e5). *p < 0.05 compared with empty adenovirus-injected side.
Fig. 3. Enhancement of GDNF expression by bilirubin and CO in primary glia cultures. Glia-enriched cultures derived from P1 rats were infected with Ad-HO-1 (10 moi) or treated with bilirubin or CO donor [Ru(CO)3Cl2]2. GDNF mRNA was analyzed 24 h later. (A) Real-time PCR showed that GDNF mRNA expression level increased in response to the administration of Ad-HO-1, bilirubin (10 mM) or [Ru(CO)3Cl2]2 (30 mM). (B) MTT result showed that treatment with various concentrations of bilirubin and [Ru(CO3)Cl2]2 for 24 h did not affect cell viability. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with control.
Fig. 4. Bilirubin enhances GDNF protein expression in glia cultures via MEK and PI3K pathway. Glia-enriched cultures were used and bilirubin at various concentrations was added to the culture medium. (A) Bilirubin treatment at 24 h enhanced GDNF expression in a concentration-dependent manner. HO-1 or HO-2 expression was not influenced by bilirubin treatment. (B) Bilirubin (10 mM) treatment time-dependently enhanced ERK1/2 and Akt phosphorylation. (C) Co-treatment with PI3K inhibitor LY294002 or MEK inhibitor PD98059 but not sGC inhibitor OQD (20 mM for each) antagonized bilirubin-induced GDNF expression. The summarized results were shown in the lower panels. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with bilirubin treatment alone.
326
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
or empty adenovirus alone did not affect the survival of THþ neuron (Fig. 1). These results indicate that both BDNF and GDNF are the important endogenously released neurotrophic factors for the survival of dopaminergic neuron and HO-1 induction rescued neurotrophic factor deficiency-induced neuronal death. 3.2. HO-1 induction by Ad-HO-1 enhances the expression of BDNF and GDNF in the substantia nigra of rat HO-1 over-expression can exert neuroprotection as mentioned above. We have previously shown that Ad-HO-1 injected into substantia nigra can reduce MPPþ-induced release of TNF-a and IL-1b in rat (Hung et al., 2008). We here further characterized the cells overexpressing HO-1 in substantia nigra following Ad-HO-1 application. Ad-HO-1/empty adenovirus (5 107 PFU) was injected into ipsilateral/contralateral side of rat substantia nigra. To avoid the mechanical damage of the tissue following local injection, the confocal images of substantia nigra were obtained 200 mm away from injection site 24 h later. The results showed that HO-1 only expressed in Ad-HO-1-injected side but not in contralateral side. In addition, HO-1 co-localized with THþ dopaminergic neuron, GFAPþ astroglia and CD11bþ microglia (Fig. 2A, the co-localization was indicated by arrow). Substantia nigra was dissected for real-time quantitative PCR 24 h after administration of Ad-HO-1. It was found that the level of HO-1/ BDNF/GDNF mRNA in Ad-HO-1 injected side was 16/26/21 folds of that in the contralateral side, respectively. These results demonstrated that HO-1 induction in vivo enhanced the expression of both BDNF and GDNF. 3.3. Increase of GDNF expression by HO-1, bilirubin and CO in glia cultures GDNF is found to be secreted from glia cells with potent neurotrophic effect to dopaminergic neurons (Lin et al., 1993). The summarized results of real-time PCR of GDNF mRNA expression in glia-enriched cultures by the treatment of Ad-HO-1 (10 moi), bilirubin (10 mM) or CO-releasing molecule [Ru(CO)3Cl2]2 (30 mM) were shown in Fig. 3A. Ad-HO-1 and its downstream products, bilirubin and CO, increased GDNF expression by 3.2 0.5, 3.1 0.1 and 2.8 0.8 folds, respectively. Bilirubin (3e30 mM) and [Ru(CO)3Cl2]2 (10e100 mM) did not influence the viability of glia cells (Fig. 3B). We then further examined the signaling pathway involved in bilirubin-induced enhancement of GDNF expression in glia cultures. Bilirubin treatment (3e30 mM) for 24 h concentrationdependently increased GDNF expression in glia cultures (Fig. 4A). Bilirubin by itself did not induce HO-1 expression or influence HO-2 level (Fig. 4A). In addition, bilirubin (10 mM) induced ERK and Akt phosphorylation time-dependently (Fig. 4B). The MEK inhibitor PD98059 and phosphoinositide 3-kinase (PI3K) inhibitor LY294002 but not soluble guanylyl cyclase (sGC) inhibitor ODQ antagonized bilirubin-induced GDNF expression (Fig. 4C), suggesting that MEK and PI3K but not sGC pathway are involved in
Fig. 5. NFkB is involved in the GDNF increasing action of bilirubin in glia cultures. Gliaenriched cultures derived from P1 rats were used. (A) Bilirubin (10 mM) treatment at 2 h enhanced p65 nuclear translocation as shown by confocal images. DAPI was used to stain the nucleus. Scale bar ¼ 5 mm (B) Immunoblotting result showed that bilirubin (10 mM) treatment at 2e3 h enhanced nuclear (N) translocation of p50 and p65 from cytosol (C). The summarized result was expressed as the ratio of N/C fraction of p50 or p65. (C) Co-treatment with NFkB inhibitor PDTC (50 mM) antagonized bilirubin-induced GDNF expression. The summarized results were shown in the lower panel. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with bilirubin treatment alone.
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
bilirubin-induced GDNF expression in glia cultures. We further examined the downstream signaling pathway of MEK. Treatment of bilirubin (10 mM) for 2 h enhanced p65 nuclear translocation in glia cultures as shown by confocal imaging (Fig. 5A). Western blotting results in Fig. 5B also showed that bilirubin (10 mM) treatment for 2e3 h enhanced nuclear translocation of p50 and p65 from cytosol. Co-treatment of NFkB inhibitor PDTC (50 mM) inhibited bilirubin-induced GDNF expression in glia (Fig. 5C), indicating that NFkB is involved in bilirubin-induced enhancement of GDNF expression.
327
3.4. [Ru(CO)3Cl2]2 treatment enhances GDNF expression through sGC-PKG and Akt pathways in glia cultures We then examined the effect of another HO-1 downstream product of CO on GDNF expression by using CO donor [Ru(CO)3Cl2]2. [Ru(CO)3Cl2]2 treatment (10e100 mM) increased GDNF expression concentration-dependently in glia cultures (Fig. 6A). [Ru(CO)3Cl2]2 at concentrations used did not affect the levels of HO-1 and HO-2 (Fig. 6A), as well as cell viability (Fig. 3B). [Ru(CO)3Cl2]2 (30 mM) treatment induced the phosphorylation of Akt but not ERK (Fig. 6B).
Fig. 6. CO donor [Ru(CO)3Cl2]2 treatment enhances GDNF expression through sGC-PKG in glia-enriched cultures. Glia-enriched cultures derived from P1 rats were used. (A) [Ru(CO)3Cl2]2 treatment at 24 h enhanced GDNF expression in a concentration-dependent manner. The protein levels of HO-1 and HO-2 were not affected by CO donor. (B) [Ru(CO)3Cl2]2 (30 mM) treatment time-dependently enhanced Akt but not ERK1/2 phosphorylation. (C) Co-treatment with PI3K inhibitor LY294002, sGC inhibitor ODQ or PKG inhibitor KT5823 antagonized [Ru(CO)3Cl2]2-induced GDNF expression. The summarized results of protein expression were shown in the lower panels. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with [Ru(CO)3Cl2]2 treatment alone.
328
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
Co-treatment of [Ru(CO)3Cl2]2 (30 mM) with cGMP-dependent protein kinase (PKG) inhibitor KT5823 (0.2 and 2 mM), sGC inhibitor ODQ (1 H-[1,2,4]oxadiazolo[4,3-a] quinoxalin-1-one; 10 and 30 mM) or PI3K inhibitor LY294002 (20 mM) antagonized [Ru(CO)3Cl2]2induced GDNF expression in glia cultures (Fig. 6C), indicating that sGC-PKG and PI3K signaling pathways are involved in the potentiating action of CO on the expression of GDNF in astrocytes. 3.5. Enhancement of BDNF expression by bilirubin and CO in cortical neuron cultures As mentioned above, both bilirubin and CO donor increased GDNF expression in glia. We further investigated whether bilirubin and CO influence BDNF expression in cortical neuron cultures. As shown in Fig. 7A, bilirubin (10 mM) treatment also induced ERK and Akt phosphorylation time-dependently. MEK inhibitor PD98059 or PI3K inhibitor LY294002 (20 mM for each) antagonized bilirubininduced BDNF expression (Fig. 7B). In addition, [Ru(CO)3Cl2]2 (30 mM) was also able to increase BDNF expression in cortical neuron cultures, whereas the sGC inhibitor ODQ (20 mM) antagonized the potentiating action. These results indicate that bilirubin or [Ru(CO)3Cl2]2 enhances neurotrophic factor expression in both neuron and astrocyte via similar signaling pathway. 4. Discussion Neurons in CNS are particularly vulnerable to oxidative stress and HO-1 has been shown to act as a neuronal protector during injury
(Chen et al., 2000). We have previously found that over-expression of HO-1 via replication-defective adenovirus vector encoding human HO-1 gene markedly inhibited MPPþ-induced dopaminergic neuronal death (Hung et al., 2008). Bilirubin, the downstream product of HO-1, also exerts the similar neuroprotective effect (Hung et al., 2008). Here, we further found that neutralization of endogenously released BDNF/GDNF with respective antibody resulted in 50% dopaminergic neuronal death in midbrain neuron-glia cocultures of E14 rats. Over-expression of HO-1 rescued the neuronal death. These results indicate that the neuroprotective action of HO-1 may not only due to its antioxidant and anti-inflammation action but also result from the regulation of BDNF and GDNF expression. In mammalian CNS, HO-2 is abundantly and constitutively expressed, whereas HO-1 is found to be expressed in a small population of neurons and glia or induced by cellular stress (Scapagnini et al., 2002; Schipper, 2004). Heme oxygenase in brain has been reported to be active and play a crucial role in the pathogenesis of Parkinson disease (Miwa et al., 2004; Schipper et al., 1998). Bilirubin at nanomolar concentration is also a potent antioxidant and neuroprotectant (Dore et al., 1999). In this study, the glia cells were cultured in DMEM with 10% FBS, which contained 41.5 mM albumin. Since bilirubin can bind with albumin, the free form of 10 mM bilirubin is about 99 nM, according to the calculation reported by Weisiger et al. (Weisiger et al., 2001). Bilirubin (3e30 mM) or [Ru(CO)3Cl2]2 (10e100 mM) treatment for 24 h doesn't influence the survival rate or the expression of HO-1 and HO-2 in glia-enriched cultures, indicating that the doses used in bilirubin or [Ru(CO)3Cl2]2 did not cause cellular stress.
Fig. 7. Bilirubin and [Ru(CO)3Cl2]2 increase BDNF expression in cortical neuron cultures. Primary cortical neurons were derived from E17 rats and cultured for 7e10 days before experiments. (A) Bilirubin (10 mM) time-dependently increased ERK and Akt phosphorylation. (B) Co-treatment with MEK inhibitor PD98059 or PI3K inhibitor LY294002 (20 mM for each) antagonized bilirubin-induced BDNF expression after 24 h incubation. (C) Co-treatment with sGC inhibitor ODQ (20 mM) antagonized [Ru(CO)3Cl2]2 (30 mM)-induced BDNF expression. The summarized results of protein expression were shown at right panels. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with bilirubin or [Ru(CO)3Cl2]2 treatment alone.
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
BDNF and GDNF have been reported to increase the survival and morphology differentiation of dopaminergic neurons (Kirschner et al., 1996; Lin et al., 1993). We here found that neutralization of endogenous release of BDNF or GDNF by specific antibody significantly enhanced dopaminergic neuronal death. Treatment of glia cultures with bilirubin concentration-dependently enhanced GDNF expression via ERK and Akt phosphorylation. Furthermore, bilirubin treatment also induced nuclear translocation of p50 and p65. Furthermore, NFkB inhibitor PDTC antagonized bilirubin-induced GDNF expression in glia cultures. In addition, bilirubin also enhanced BDNF expression in cortical neuron cultures via similar pathway. Fluoxetine has also been reported to induce BDNF and GDNF mRNA expression in rat astrocyte via the activation of MEKERK pathway (Mercier et al., 2004). Rasagiline (N-Propargyl-l(R)aminoindan; an anti-Parkinson drug) also increases GDNF level via the activation of ERK1/2 and p65 nuclear translocation in human neuroblastoma SH-SY5Y (Maruyama et al., 2004). Dibutyryl cGMP, an analog of cGMP, preferentially activates PKG and elevates BDNF content from dopaminergic cell line SN4741 (Chun et al., 2000). Here we also found that the CO-releasing molecule, [Ru(CO)3Cl2]2, enhanced BDNF and GDNF expression in both neuron and glia through sGC-PKG pathway. These results suggest that NO and CO may regulate BDNF and GDNF expression through cGMP-PKG pathway. In conclusion, HO-1 induction exerts neuroprotection through its downstream products, i.e. bilirubin and CO. The low concentration of free-form bilirubin can exert its protective action on dopaminergic neuron via GDNF or BDNF induction, which may be mediated through MEK, Akt and NFkB activation. CO also increases the expression of neurotrophic factor via sGC-PKG-dependent pathway. Acknowledgement We thank Dr. Lee-Young Chau for providing the Ad-HO-1. References Baranano, D.E., Rao, M., Ferris, C.D., Snyder, S.H., 2002. Biliverdin reductase: a major physiologic cytoprotectant. Proc. Natl. Acad. Sci. U S A 99, 16093e16098. Chen, K., Gunter, K., Maines, M.D., 2000. Neurons overexpressing heme oxygenase-1 resist oxidative stress-mediated cell death. J. Neurochem. 75, 304e313. Chen, P.S., Peng, G.S., Li, G., Yang, S., Wu, X., Wang, C.C., Wilson, B., Lu, R.B., Gean, P.W., Chuang, D.M., Hong, J.S., 2006. Valproate protects dopaminergic neurons in midbrain neuron/glia cultures by stimulating the release of neurotrophic factors from astrocytes. Mol. Psychiatry 11, 1116e1125. Choi-Lundberg, D.L., Lin, Q., Chang, Y.N., Chiang, Y.L., Hay, C.M., Mohajeri, H., Davidson, B.L., Bohn, M.C., 1997. Dopaminergic neurons protected from degeneration by GDNF gene therapy. Science 275, 838e841. Chun, H.S., Son, J.J., Son, J.H., 2000. Identification of potential compounds promoting BDNF production in nigral dopaminergic neurons: clinical implication in Parkinson's disease. Neuroreport 11, 511e514. Dore, S., Takahashi, M., Ferris, C.D., Zakhary, R., Hester, L.D., Guastella, D., Snyder, S.H., 1999. Bilirubin, formed by activation of heme oxygenase-2, protects neurons against oxidative stress injury. Proc. Natl. Acad. Sci. U S A 96, 2445e2450. Hornykiewicz, O., 1988. Neurochemical pathology and the etiology of Parkinson's disease: basic facts and hypothetical possibilities. Mt. Sinai J. Med. 45, 35e43. Hu, C.M., Chen, Y.H., Chiang, M.T., Chau, L.Y., 2004. Heme oxygenase-1 inhibits angiotensin II-induced cardiac hypertrophy in vitro and in vivo. Circulation 110, 309e316. Hung, S.Y., Liou, H.C., Kang, K.H., Wu, R.M., Wen, C.C., Fu, W.M., 2008. Overexpression of heme oxygenase-1 protects dopaminergic neurons against 1-methyl-4-phenylpyridinium-induced neurotoxicity. Mol. Pharmacol. 74, 1564e1575. Jenner, P., 2003. Oxidative stress in Parkinson's disease. Ann. Neurol. 53 (Suppl. 3), S26eS38. Juan, S.H., Lee, T.S., Tseng, K.W., Liou, J.Y., Shyue, S.K., Wu, K.K., Chau, L.Y., 2001. Adenovirus-mediated heme oxygenase-1 gene transfer inhibits the development of atherosclerosis in apolipoprotein E-deficient mice. Circulation 104, 1519e1525. Kim, H.P., Wang, X., Galbiati, F., Ryter, S.W., Choi, A.M., 2004. Caveolae compartmentalization of heme oxygenase-1 in endothelial cells. Faseb J. 18, 1080e1089.
329
Kirschner, P.B., Jenkins, B.G., Schulz, J.B., Finkelstein, S.P., Matthews, R.T., Rosen, B.R., Beal, M.F., 1996. NGF, BDNF and NT-5, but not NT-3 protect against MPPþ toxicity and oxidative stress in neonatal animals. Brain Res. 713, 178e185. Le, W.D., Xie, W.J., Appel, S.H., 1999. Protective role of heme oxygenase-1 in oxidative stress-induced neuronal injury. J. Neurosci. Res. 56, 652e658. Lin, L.F., Doherty, D.H., Lile, J.D., Bektesh, S., Collins, F., 1993. GDNF: a glial cell linederived neurotrophic factor for midbrain dopaminergic neurons. Science 260, 1130e1132. Lu, D.Y., Yu, W.H., Yeh, W.L., Tang, C.H., Leung, Y.M., Wong, K.L., Chen, Y.F., Lai, C.H., Fu, W.M., 2009. Hypoxia-induced matrix metalloproteinase-13 expression in astrocytes enhances permeability of brain endothelial cells. J. Cell Physiol. 220, 163e173. Marks, G.S., 1994. Heme oxygenase: the physiological role of one of its metabolites, carbon monoxide and interactions with zinc protoporphyrin, cobalt protoporphyrin and other metalloporphyrins. Cell Mol. Biol. (Noisy-le-grand) 40, 863e870. Maruyama, W., Nitta, A., Shamoto-Nagai, M., Hirata, Y., Akao, Y., Yodim, M., Furukawa, S., Nabeshima, T., Naoi, M., 2004. N-Propargyl-1 (R)-aminoindan, rasagiline, increases glial cell line-derived neurotrophic factor (GDNF) in neuroblastoma SH-SY5Y cells through activation of NF-kappaB transcription factor. Neurochem. Int. 44, 393e400. McKinstry, R., Qiao, L., Yacoub, A., Dai, Y., Decker, R., Holt, S., Hagan, M.P., Grant, S., Dent, P., 2002. Inhibitors of MEK1/2 interact with UCN-01 to induce apoptosis and reduce colony formation in mammary and prostate carcinoma cells. Cancer Biol. Ther. 1, 243e253. Mercier, G., Lennon, A.M., Renouf, B., Dessouroux, A., Ramauge, M., Courtin, F., Pierre, M., 2004. MAP kinase activation by fluoxetine and its relation to gene expression in cultured rat astrocytes. J. Mol. Neurosci. 24, 207e216. Miwa, H., Kubo, T., Morita, S., Nakanishi, I., Kondo, T., 2004. Oxidative stress and microglial activation in substantia nigra following striatal MPPþ. Neuroreport 15, 1039e1044. Motterlini, R., Clark, J.E., Foresti, R., Sarathchandra, P., Mann, B.E., Green, C.J., 2002. Carbon monoxide-releasing molecules: characterization of biochemical and vascular activities. Circ. Res. 90, E17eE24. Ollinger, R., Kogler, P., Troppmair, J., Hermann, M., Wurm, M., Drasche, A., Konigsrainer, I., Amberger, A., Weiss, H., Ofner, D., Bach, F.H., Margreiter, R., 2007. Bilirubin inhibits tumor cell growth via activation of ERK. Cell Cycle 6, 3078e3085. Otterbein, L.E., Bach, F.H., Alam, J., Soares, M., Tao Lu, H., Wysk, M., Davis, R.J., Flavell, R.A., Choi, A.M., 2000. Carbon monoxide has anti-inflammatory effects involving the mitogen-activated protein kinase pathway. Nat. Med. 6, 422e428. Otterbein, L.E., Zuckerbraun, B.S., Haga, M., Liu, F., Song, R., Usheva, A., Stachulak, C., Bodyak, N., Smith, R.N., Csizmadia, E., Tyagi, S., Akamatsu, Y., Flavell, R.J., Billiar, T.R., Tzeng, E., Bach, F.H., Choi, A.M., Soares, M.P., 2003. Carbon monoxide suppresses arteriosclerotic lesions associated with chronic graft rejection and with balloon injury. Nat. Med. 9, 183e190. Pachori, A.S., Smith, A., McDonald, P., Zhang, L., Dzau, V.J., Melo, L.G., 2007. Hemeoxygenase-1-induced protection against hypoxia/reoxygenation is dependent on biliverdin reductase and its interaction with PI3K/Akt pathway. J. Mol. Cell Cardiol. 43, 580e592. Roberts, P.J., Der, C.J., 2007. Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer. Oncogene 26, 3291e3310. Roberts, P.J., Kapur, H., 1977. Neurochemical changes in rat striatum and substantia nigra following drug administration. Neurochem. Res. 2, 485e494. Ryter, S.W., Alam, J., Choi, A.M., 2006. Heme oxygenase-1/carbon monoxide: from basic science to therapeutic applications. Physiol. Rev. 86, 583e650. Scapagnini, G., D'Agata, V., Calabrese, V., Pascale, A., Colombrita, C., Alkon, D., Cavallaro, S., 2002. Gene expression profiles of heme oxygenase isoforms in the rat brain. Brain Res. 954, 51e59. Schipper, H.M., 2000. Heme oxygenase-1: role in brain aging and neurodegeneration. Exp. Gerontol. 35, 821e830. Schipper, H.M., 2004. Heme oxygenase expression in human central nervous system disorders. Free Radic. Biol. Med. 37, 1995e2011. Schipper, H.M., Cisse, S., Stopa, E.G., 1995. Expression of heme oxygenase-1 in the senescent and Alzheimer-diseased brain. Ann. Neurol. 37, 758e768. Schipper, H.M., Liberman, A., Stopa, E.G., 1998. Neural heme oxygenase-1 expression in idiopathic Parkinson's disease. Exp. Neurol. 150, 60e68. Sindhu, K.M., Banerjee, R., Senthilkumar, K.S., Saravanan, K.S., Raju, B.C., Rao, J.M., Mohanakumar, K.P., 2006. Rats with unilateral median forebrain bundle, but not striatal or nigral, lesions by the neurotoxins MPPþ or rotenone display differential sensitivity to amphetamine and apomorphine. Pharmacol. Biochem. Behav. 84, 321e329. Spenger, C., Hyman, C., Studer, L., Egli, M., Evtouchenko, L., Jackson, C., DahlJorgensen, A., Lindsay, R.M., Seiler, R.W., 1995. Effects of BDNF on dopaminergic, serotonergic, and GABAergic neurons in cultures of human fetal ventral mesencephalon. Exp. Neurol. 133, 50e63. Tenhunen, R., Marver, H.S., Schmid, R., 1969. Microsomal heme oxygenase. Characterization of the enzyme. J. Biol. Chem. 244, 6388e6394. Weisiger, R.A., Ostrow, J.D., Koehler, R.K., Webster, C.C., Mukerjee, P., Pascolo, L., Tiribelli, C., 2001. Affinity of human serum albumin for bilirubin varies with albumin concentration and buffer composition: results of a novel ultrafiltration method. J. Biol. Chem. 276, 29953e29960.
Contents lists available at ScienceDirect
Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm
The mechanism of heme oxygenase-1 action involved in the enhancement of neurotrophic factor expressionq Shih-Ya Hung a, Houng-Chi Liou a, Wen-Mei Fu a, b, * a b
Pharmacological Institute, College of Medicine, National Taiwan University, Taipei 100, Taiwan Neurobiology and Cognition Center, National Taiwan University, Taipei 106, Taiwan
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 July 2009 Received in revised form 7 November 2009 Accepted 10 November 2009
Heme oxygenase-1 (HO-1) is up-regulated in response to oxidative stress and catalyzes the degradation of pro-oxidant heme to carbon monoxide (CO), iron and bilirubin. Bilirubin is a potent antioxidant and neuroprotectant. Neurotrophic factors of BDNF and GDNF also play important roles in survival and morphological differentiation of dopaminergic neurons. We have previously found that HO-1 induction by adenovirus containing human HO-1 gene (Ad-HO-1) in substantia nigra of rat increases BDNF and GDNF expression. We here further examined the possible mechanism of HO-1 action involved in the enhancement of neurotrophic factor expression. Treatment of anti-BDNF/GDNF antibody significantly enhanced dopaminergic neuronal death, whereas Ad-HO-1 co-treatment was able to antagonize the apoptosis-inducing effect of these antibodies. The confocal imaging shows that HO-1 induction appeared in dopaminergic neuron, astrocyte and microglia at 24 h after injecting Ad-HO-1. HO-1 induced-BDNF/ GDNF mRNA expression in substantia nigra was 26/21 folds of that of the contralateral Ad-injected side. The downstream product bilirubin increased GDNF expression through ERK and PI3K-Akt pathways, and also enhanced NFkB (p65 and p50) nuclear translocation in glia-enriched cultures. In addition, bilirubin also enhanced BDNF expression through similar pathway in cortical neuron-enriched cultures. We also examined the effect of another HO-1 product, CO, by using CO donor. [Ru(CO)3Cl2]2 increased neurotrophic factor expression via sGC-PKG pathway in both neuron and glia. These results indicate that the downstream products of HO-1, i.e. bilirubin and CO, modulate BDNF and GDNF expression in neuron and astrocyte. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: HO-1 Bilirubin BDNF GDNF CO
1. Introduction Parkinson's disease (PD) is a movement disorder characterized by the accelerated loss of dopaminergic neurons in the pars compacta of the substantia nigra (Hornykiewicz, 1988). Oxidative stress contributes to the cascade leading to dopamine cell degeneration in
Abbreviations: BDNF, brain-derived neurotrophic factor; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; GDNF, glia cell line-derived neurotrophic factor; HO-1, heme oxygenase-1; HO-2, heme oxygenase-2; moi, multiplicity of infection; MPPþ, 1-methyl-4-phenylpyridinium; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; [Ru(CO)3Cl2]2, tricarbonyldichlororuthenium (II) dimmer; TH, tyrosine hydroxylase; sGC, soluble guanylyl cyclase; ODQ, (1H-[1,2,4] oxadiazolo[4,3-a]quinoxalin-1-one); PKG, cGMP-dependent protein kinase. q This work was supported by grants from the National Science of Council and National Taiwan University (97R0066-53). * Correspondence to: Wen-Mei Fu, Pharmacological Institute, College of Medicine, National Taiwan University, No. 1, Sec. 1, Jen-Ai Road, Taipei 100, Taiwan. Tel: þ886 2 23123456x88319, fax: þ886 2 23417930. E-mail address: [email protected] (W.-M. Fu). 0028-3908/$ e see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropharm.2009.11.003
PD (Jenner, 2003). It has been reported that there is an up-regulation of heme oxygenase-1 (HO-1) in the substantia nigra of PD patients (Schipper et al., 1998). Heme oxygenase is an enzyme, which degrades intracellular heme to free iron, carbon monoxide (CO) and biliverdin. Biliverdin is subsequently converted to bilirubin by biliverdin reductase (BVR) and free iron is sequestered by ferritin (Tenhunen et al., 1969). HO-1, a 32 kDa cellular stress response protein (also known as Hsp32) can be rapidly induced under oxidative challenge and other noxious stimuli in the brain or other tissues (Hu et al., 2004; Le et al., 1999). In normal brain, the level of HO-1 is rather low, however, HO-2 is constitutively expressed and does not respond to environmental stress (Baranano et al., 2002; Kim et al., 2004; Schipper, 2000; Schipper et al., 1995). The end products from HO enzymatic activity, CO and bilirubin, exert many biological functions (Ryter et al., 2006). CO is an endogenous gaseous molecule, which activates soluble guanylyl cyclase (sGC) at a lower rate than nitric oxide and plays significant role in anti-apoptosis, anti-inflammation, anti-proliferation, neurotransmission and vasorelaxant (Otterbein et al., 2000, 2003).
322
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
The HO/CO system may play similar role as nitric oxide synthase (NOS)/nitric oxide (NO) system in the regulation of cellular function and communication (Marks, 1994). On the other hand, bilirubin acts as reactive oxygen species (ROS) scavenger and exerts strong antiinflammatory and neuroprotective function (Baranano et al., 2002; Dore et al., 1999). We have previously found that HO-1 induction in substantia nigra by adenovirus exerts a strong neuroprotective effect in Parkinsonian rat model. HO-1 induction in substantia nigra upregulates BDNF in dopaminergic neuron and GDNF in glia (Hung et al., 2008). In vitro treatment of glia with bilirubin or CO also exerts similar effect in up-regulating the expression of neurotrophic factors. One recent study indicates that HO-1-midiated cytoprotection is dependent on PI3K/Akt pathway (via p85a and Akt phosphorylation) to decrease ROS production and apoptosis (Pachori et al., 2007). In addition, many studies indicate that activation of the ERK1/2 pathway inhibits apoptosis and promotes cell survival and growth (McKinstry et al., 2002; Roberts and Der, 2007). Furthermore, Ollinger et al. (2007) have also demonstrated that bilirubin treatment in vitro induces the phosphorylation of MEK and ERK1/2 but not p38 or JNK1/2. Here we further investigated the signaling pathways involved in bilirubin- or CO-induced enhancement of neurotrophic factor expression. 2. Materials and methods 2.1. Reagents Bilirubin stock solution (10 mM) (Sigma, St. Louis, MO) was prepared in 0.1 M NaOH. [Ru(CO)3Cl2]2 (Sigma, St. Louis, MO) is a carbon monoxide (CO)-releasing molecule and each mole of [Ru(CO)3Cl2]2 releases approximately 0.7 mol of CO gas (Motterlini et al., 2002). [Ru(CO)3Cl2]2 was freshly dissolved in DMSO to make a 100 mM stock solution. KT5823, PDTC, ODQ, PD98059 and LY294002 were purchased from Sigma (Sigma, St. Louis, MO).
cultures were maintained in a humidified chamber at 37 C in a 5% CO2 atmosphere for 7e10 days, the medium was changed every 3-day before use. Primary rat glia cultures were prepared according to previous report with some modification (Lu et al., 2009). The cells were seeded at 1 107 on 75 cm2 flask. The culture medium for glia was DMEM with 10% FBS, 4500 mg/L glucose, 100 U/mL penicillin and 0.1 mg/mL streptomycin, which was changed every 4-day. On the last day, flasks were placed on a shaker platform and shaken at 220 rpm for 6 h at 37 C to remove oligodendrocytes and microglia. The glia-enriched cultures were then grown to confluence before use. All inhibitors were pre-treated for 30 min before the administration of bilirubin or [Ru(CO)3Cl2]3. 2.4. Intra-substantia nigra injection Wistar rats weighing approximately 350 g were used. Rats received bilateral injection of Ad-HO-1/empty adenovirus in the ipsilateral/contralateral side (5 107 PFU for each in 1 mL buffer containing 10 mM Tris pH 8.0, 2 mM MgCl2, 4% sucrose). For intra-substantia nigra injection, stereotaxic coordinates for SNpc were lateral (L) þ 2.0 mm, antero-posterior (AP, from the bregma point) - 5.3 mm and dorsiventral (DV) þ 7.8 mm (Sindhu et al., 2006). The solution was injected into the substantia nigra with a 10 mL Hamilton syringe coupled to a motorized injector (Stoelting) at a rate of 0.2 mL/min and the needle was left in situ for at least 5 min after injection. The rats were sacrificed 24 h later for the studies of real-time quantitative PCR or immunofluorescent staining. 2.5. Total RNA extraction, reverse transcription (RT) and real-time quantitative PCR Substantia nigra and striatal tissues dissected from rats were prepared as described previously (Roberts and Kapur, 1977). Total RNA was extracted from substantia nigra using TRIzol kit (MDBio Inc., Taipei, Taiwan). Single-strand cDNA was synthesized using SSIII reverse transcription reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. All cDNA samples were stored at e 20 C. Real-time quantitative PCR was performed to monitor the expression of HO-1, BDNF, GDNF and GAPDH. The TaqMan technology was used and the results were analyzed in a 7900 HT sequence detector system (Applied Biosystems, Foster City, CA). All primers and TaqMan probes were purchased from Applied Biosystems. The amplification reaction mixture (25 mL) contained 2.5 mL cDNA sample, TaqMan Universal PCR Master Mix and TaqMan primer/probe pre-mix. The thermal cycling conditions included 10 min at 95 C, proceeding with 40 cycles of 95 C for 15 s and 60 C for 1 min. The GDNF or BDNF mRNA levels were normalized to GAPDH mRNA and expressed relative to control using the DDCt method.
2.2. Preparation of the recombinant adenovirus 2.6. Immunofluorescence Replication-defective empty adenovirus and recombinant adenovirus carrying human HO-1 gene (Ad-HO-1) were kindly provided by Dr. Lee-Young Chau (Academia Sinica, Taiwan). Large scales of viral vectors were purified using CsCl ultracentrifugation as previously described (Juan et al., 2001). Virus was amplified in QBI-293 cells and the plaque forming unit (PFU) within 21 days was used for determining the titer of empty adenovirus and Ad-HO-1 stocks. 2.3. Cell cultures of midbrain, cortical neuron and astrocyte Neuron-glia co-cultures from the midbrain of E14 Wistar rats were prepared as previously described with some modification (Chen et al., 2006). Briefly, the ventral portion of the midbrain was removed in sterile ice-cold CaCl2, MgCl2 and MgSO4-free Hank balanced salt solution (GIBCO, Grand Island, NY). The tissues were cleaned, minced and mechanically dissociated by passage through a flame-polished Pasteur pipette. Dissociated cells were seeded in DMEM (GIBCO, Grand Island, NY) with 10% fetal bovine serum (FBS; Biological Industries, Grand Island, NY, Cat. No: 04-001-1A, Lot no: 115115), 4500 mg/L glucose, 100 U/mL penicillin and 0.1 mg/mL streptomycin. Cells were seeded at 3.5 105/well on poly-D-lysine-coated 48-well plates. The cultures were kept in a humidified chamber at 37 C in a 5% CO2 atmosphere. Twentyfour hours after plating, the cells were changed to MEM (GIBCO, Grand Island, NY) with 2% FBS, 2% horse serum (HS; Hyclone, Logan, UT),100 U/mL penicillin and 0.1 mg/ mL streptomycin. The Day-4 cultures were infected with Ad-HO-1 or empty adenovirus (10 moi for each) and co-treated with anti-BDNF or anti-GDNF antibody (1 mg/ mL IgG for each; Santa Cruz Biotechnology, CA) for 24 h. MPPþ at 5 nM was then added for another 48 h. Dopaminergic neurons were characterized by staining with a rabbit anti-tyrosine hydroxylase (anti-TH) antibody (1:10,000; Calbiochem Inc., San. Diego, CA). The cultures were then incubated for 1 h with the biotinylated secondary antibody, and then for 30 min with avidineperoxidase complex (ABC kit; Vector Laboratories. Burlingame, CA). Finally, the labeling was revealed by treatment with 0.01% hydrogen peroxide and 0.05% 3, 30 -diaminobenzidine (DAB; Sigma, St. Louis, MO). The total number of TH-immunopositive cells in each well was counted under inverted microscope at 200 magnification (Leica, Heidelberg, Germany). Cortical neuron cultures were obtained from cerebral cortex of E17 Wistar rats and glia cultures were obtained from P1 Wistar rats. For cortical neuron cultures, the cells were seeded at 1.2 106 cell/well on poly-D-lysine-coated 6-well plate. Twenty-four hours after plating, the cells were changed to DMEM containing 2% B27, 4500 mg/L glucose, 100 U/mL penicillin and 0.1 mg/mL streptomycin. The neuron
Brain tissue sections (30 mm thickness) were used for double-immunolabeling studies. The sections were incubated with 10% BSA and 0.1% Triton X-100 in phosphate-buffered saline (PBS) for 1 h, and then incubated overnight at 4 C with the following primary antibody: TH (1:3000; Calbiochem Inc., San. Diego, CA), HO-1 (1:600; Catalog # OSA-150F, Stressgen Biotechnologies, Victoria, British Columbia), GFAP (1:500) and CD11b (1:250; Abcam, Cambridge, MA) and p65 (1:500; Santa Cruz Biotechnology, Santa Cruz, CA) in PBS, and then with Alexa-488 or 543conjugated goat anti-rabbit or anti-mouse as the secondary antibody (1:600; Invitrogen, Carlsbad, CA). The confocal images were obtained at excitation wavelength of 488 nm and 543 nm, respectively (model SP2 TCS; Leica, Heidelberg, Germany). 2.7. Western blotting The protein expression levels of BDNF, GDNF, ERK, pERK, Akt, pAkt, p50 and p65 were determined by Western blotting. Samples obtained from cells were homogenized in RIPA buffer containing 1 mM PMSF, 25 mM leupeptin and 1 mg/mL aprotinin. Isolation of nuclear proteins from cells used the Pierce NE-PER extraction reagents (Pierce, Rockford, IL). Protein concentration was determined using the BCA protein assay kit (Pierce, Rockford, IL). Protein at 30e50 mg was separated by SDS-PAGE using a 10e15% resolving gel under reducing conditions and electrotransferred onto immobilon-P membrane (Millpore, Bedford, MA). After being blocked with 5% nonfat milk in TBS-T (0.5% Tween 20 in 20 mM Tris and 137 mM NaCl) for 1 h at room temperature, the membranes were incubated overnight at 4 C with anti-BDNF (1:3000), anti-GDNF (1:3000), anti-pAkt (1:5000), anti-p50 (1:1000), anti-p65 (1:1000) or anti-pERK (1:5000) primary antibodies (Santa Cruz Biotechnology) diluted in TBS-T. The membranes were probed with a mouse anti-GAPDH antibody as a standard (1:10,000; Abcam Inc., Cambridge, MA). The blots were then incubated for 1 h at room temperature with an HRP-conjugated secondary antibody (1:20,000; Amersham Life Science, Arlington Heights, IL). Protein bands were detected using the ECL Western Blotting Substrate (Pierce, Rockford, IL) and were estimated using the Image Analysis Program Labwork 4.5 (UVP, Inc., Upland, CA). 2.8. Statistics Results are expressed as mean SEM. Results were analyzed with one-way analysis of variance (ANOVA) and Neuman-Keuls post-hoc test to determine
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329 statistical significance between specific groups. Difference was considered significant when p < 0.05.
3. Results 3.1. HO-1 induction antagonizes BDNF or GDNF deficiency-induced dopaminergic neuronal death in midbrain neuron-glia co-cultures BDNF and GDNF enhance the survival and morphological differentiation of dopaminergic neuron in CNS (Lin et al., 1993; Spenger et al., 1995). Adenovirus vector-mediated GDNF gene therapy inhibits
323
the dopaminergic neuron loss in Parkinsonism patients (Choi-Lundberg et al., 1997). We here examined the role of BDNF and GDNF in dopaminergic neuronal survival in midbrain neuron-glia co-cultures. Fig. 1 showed that treatment of anti-BDNF or anti-GDNF antibody for 48 h to neutralize the function of endogenously released BDNF of GDNF in the cultures significantly enhanced dopaminergic neuronal death. The control IgG was used for comparison (the THþ neurons in anti-BDNF, anti-GDNF and control IgG groups were 53.7 5.3%, 54.8 4.5% and 93.0 3.1%, respectively). Co-treatment of Ad-HO-1 (10 moi) but not empty adenovirus with antibodies against neurotrophic factors antagonized dopaminergic neuronal death. Ad-HO-1
Fig. 1. HO-1 induction antagonizes dopaminergic neuronal death induced by BDNF or GDNF deficiency in midbrain neuron-glia co-cultures. Midbrain neuron-glia co-cultures derived from E14 rats were used to evaluate dopaminergic neuronal survival. The dopaminergic neurons were counted by the immunostaining with antibody against tyrosine hydroxylase. (A) Administration of anti-BDNF or anti-GDNF antibody but not control IgG (1 mg/mL) in midbrain neuron-glia co-cultures for 48 h enhanced dopaminergic neuronal death. Co-treatment with Ad-HO-1 (10 moi) but not empty adenovirus antagonized the neuronal death induced by BDNF or GDNF deficiency. The summarized result was shown in (B). Scale bar: 50 mm. Data are given as mean SEM (n ¼ 4). *p < 0.05 compared with no virus control; #p < 0.05 compared with anti-BDNF or anti-GDNF alone.
324
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
Fig. 2. HO-1 induction by Ad-HO-1 increases BDNF and GDNF expression in substantia nigra of adult rats. Ad-HO-1 (5 107 PFU) was locally injected into substantia nigra and immunofluorescence was performed in the slice 200 mm away from injection site. (A) The confocal image showed that HO-1 expressed in THþ neuron, GFAPþ astroglia and CD11bþ microglia in the midbrain of adult rat in response to Ad-HO-1 (5 107 PFU) but not empty adenovirus injection (co-localization is indicated by arrow). (B) Real-time PCR showed that mRNA levels of HO-1, BDNF and GDNF increased in midbrain 24 h after local administration of Ad-HO-1. Scale bar: 50 mm. Data are given as mean SEM (n ¼ 3e5). *p < 0.05 compared with empty adenovirus-injected side.
Fig. 3. Enhancement of GDNF expression by bilirubin and CO in primary glia cultures. Glia-enriched cultures derived from P1 rats were infected with Ad-HO-1 (10 moi) or treated with bilirubin or CO donor [Ru(CO)3Cl2]2. GDNF mRNA was analyzed 24 h later. (A) Real-time PCR showed that GDNF mRNA expression level increased in response to the administration of Ad-HO-1, bilirubin (10 mM) or [Ru(CO)3Cl2]2 (30 mM). (B) MTT result showed that treatment with various concentrations of bilirubin and [Ru(CO3)Cl2]2 for 24 h did not affect cell viability. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with control.
Fig. 4. Bilirubin enhances GDNF protein expression in glia cultures via MEK and PI3K pathway. Glia-enriched cultures were used and bilirubin at various concentrations was added to the culture medium. (A) Bilirubin treatment at 24 h enhanced GDNF expression in a concentration-dependent manner. HO-1 or HO-2 expression was not influenced by bilirubin treatment. (B) Bilirubin (10 mM) treatment time-dependently enhanced ERK1/2 and Akt phosphorylation. (C) Co-treatment with PI3K inhibitor LY294002 or MEK inhibitor PD98059 but not sGC inhibitor OQD (20 mM for each) antagonized bilirubin-induced GDNF expression. The summarized results were shown in the lower panels. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with bilirubin treatment alone.
326
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
or empty adenovirus alone did not affect the survival of THþ neuron (Fig. 1). These results indicate that both BDNF and GDNF are the important endogenously released neurotrophic factors for the survival of dopaminergic neuron and HO-1 induction rescued neurotrophic factor deficiency-induced neuronal death. 3.2. HO-1 induction by Ad-HO-1 enhances the expression of BDNF and GDNF in the substantia nigra of rat HO-1 over-expression can exert neuroprotection as mentioned above. We have previously shown that Ad-HO-1 injected into substantia nigra can reduce MPPþ-induced release of TNF-a and IL-1b in rat (Hung et al., 2008). We here further characterized the cells overexpressing HO-1 in substantia nigra following Ad-HO-1 application. Ad-HO-1/empty adenovirus (5 107 PFU) was injected into ipsilateral/contralateral side of rat substantia nigra. To avoid the mechanical damage of the tissue following local injection, the confocal images of substantia nigra were obtained 200 mm away from injection site 24 h later. The results showed that HO-1 only expressed in Ad-HO-1-injected side but not in contralateral side. In addition, HO-1 co-localized with THþ dopaminergic neuron, GFAPþ astroglia and CD11bþ microglia (Fig. 2A, the co-localization was indicated by arrow). Substantia nigra was dissected for real-time quantitative PCR 24 h after administration of Ad-HO-1. It was found that the level of HO-1/ BDNF/GDNF mRNA in Ad-HO-1 injected side was 16/26/21 folds of that in the contralateral side, respectively. These results demonstrated that HO-1 induction in vivo enhanced the expression of both BDNF and GDNF. 3.3. Increase of GDNF expression by HO-1, bilirubin and CO in glia cultures GDNF is found to be secreted from glia cells with potent neurotrophic effect to dopaminergic neurons (Lin et al., 1993). The summarized results of real-time PCR of GDNF mRNA expression in glia-enriched cultures by the treatment of Ad-HO-1 (10 moi), bilirubin (10 mM) or CO-releasing molecule [Ru(CO)3Cl2]2 (30 mM) were shown in Fig. 3A. Ad-HO-1 and its downstream products, bilirubin and CO, increased GDNF expression by 3.2 0.5, 3.1 0.1 and 2.8 0.8 folds, respectively. Bilirubin (3e30 mM) and [Ru(CO)3Cl2]2 (10e100 mM) did not influence the viability of glia cells (Fig. 3B). We then further examined the signaling pathway involved in bilirubin-induced enhancement of GDNF expression in glia cultures. Bilirubin treatment (3e30 mM) for 24 h concentrationdependently increased GDNF expression in glia cultures (Fig. 4A). Bilirubin by itself did not induce HO-1 expression or influence HO-2 level (Fig. 4A). In addition, bilirubin (10 mM) induced ERK and Akt phosphorylation time-dependently (Fig. 4B). The MEK inhibitor PD98059 and phosphoinositide 3-kinase (PI3K) inhibitor LY294002 but not soluble guanylyl cyclase (sGC) inhibitor ODQ antagonized bilirubin-induced GDNF expression (Fig. 4C), suggesting that MEK and PI3K but not sGC pathway are involved in
Fig. 5. NFkB is involved in the GDNF increasing action of bilirubin in glia cultures. Gliaenriched cultures derived from P1 rats were used. (A) Bilirubin (10 mM) treatment at 2 h enhanced p65 nuclear translocation as shown by confocal images. DAPI was used to stain the nucleus. Scale bar ¼ 5 mm (B) Immunoblotting result showed that bilirubin (10 mM) treatment at 2e3 h enhanced nuclear (N) translocation of p50 and p65 from cytosol (C). The summarized result was expressed as the ratio of N/C fraction of p50 or p65. (C) Co-treatment with NFkB inhibitor PDTC (50 mM) antagonized bilirubin-induced GDNF expression. The summarized results were shown in the lower panel. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with bilirubin treatment alone.
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
bilirubin-induced GDNF expression in glia cultures. We further examined the downstream signaling pathway of MEK. Treatment of bilirubin (10 mM) for 2 h enhanced p65 nuclear translocation in glia cultures as shown by confocal imaging (Fig. 5A). Western blotting results in Fig. 5B also showed that bilirubin (10 mM) treatment for 2e3 h enhanced nuclear translocation of p50 and p65 from cytosol. Co-treatment of NFkB inhibitor PDTC (50 mM) inhibited bilirubin-induced GDNF expression in glia (Fig. 5C), indicating that NFkB is involved in bilirubin-induced enhancement of GDNF expression.
327
3.4. [Ru(CO)3Cl2]2 treatment enhances GDNF expression through sGC-PKG and Akt pathways in glia cultures We then examined the effect of another HO-1 downstream product of CO on GDNF expression by using CO donor [Ru(CO)3Cl2]2. [Ru(CO)3Cl2]2 treatment (10e100 mM) increased GDNF expression concentration-dependently in glia cultures (Fig. 6A). [Ru(CO)3Cl2]2 at concentrations used did not affect the levels of HO-1 and HO-2 (Fig. 6A), as well as cell viability (Fig. 3B). [Ru(CO)3Cl2]2 (30 mM) treatment induced the phosphorylation of Akt but not ERK (Fig. 6B).
Fig. 6. CO donor [Ru(CO)3Cl2]2 treatment enhances GDNF expression through sGC-PKG in glia-enriched cultures. Glia-enriched cultures derived from P1 rats were used. (A) [Ru(CO)3Cl2]2 treatment at 24 h enhanced GDNF expression in a concentration-dependent manner. The protein levels of HO-1 and HO-2 were not affected by CO donor. (B) [Ru(CO)3Cl2]2 (30 mM) treatment time-dependently enhanced Akt but not ERK1/2 phosphorylation. (C) Co-treatment with PI3K inhibitor LY294002, sGC inhibitor ODQ or PKG inhibitor KT5823 antagonized [Ru(CO)3Cl2]2-induced GDNF expression. The summarized results of protein expression were shown in the lower panels. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with [Ru(CO)3Cl2]2 treatment alone.
328
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
Co-treatment of [Ru(CO)3Cl2]2 (30 mM) with cGMP-dependent protein kinase (PKG) inhibitor KT5823 (0.2 and 2 mM), sGC inhibitor ODQ (1 H-[1,2,4]oxadiazolo[4,3-a] quinoxalin-1-one; 10 and 30 mM) or PI3K inhibitor LY294002 (20 mM) antagonized [Ru(CO)3Cl2]2induced GDNF expression in glia cultures (Fig. 6C), indicating that sGC-PKG and PI3K signaling pathways are involved in the potentiating action of CO on the expression of GDNF in astrocytes. 3.5. Enhancement of BDNF expression by bilirubin and CO in cortical neuron cultures As mentioned above, both bilirubin and CO donor increased GDNF expression in glia. We further investigated whether bilirubin and CO influence BDNF expression in cortical neuron cultures. As shown in Fig. 7A, bilirubin (10 mM) treatment also induced ERK and Akt phosphorylation time-dependently. MEK inhibitor PD98059 or PI3K inhibitor LY294002 (20 mM for each) antagonized bilirubininduced BDNF expression (Fig. 7B). In addition, [Ru(CO)3Cl2]2 (30 mM) was also able to increase BDNF expression in cortical neuron cultures, whereas the sGC inhibitor ODQ (20 mM) antagonized the potentiating action. These results indicate that bilirubin or [Ru(CO)3Cl2]2 enhances neurotrophic factor expression in both neuron and astrocyte via similar signaling pathway. 4. Discussion Neurons in CNS are particularly vulnerable to oxidative stress and HO-1 has been shown to act as a neuronal protector during injury
(Chen et al., 2000). We have previously found that over-expression of HO-1 via replication-defective adenovirus vector encoding human HO-1 gene markedly inhibited MPPþ-induced dopaminergic neuronal death (Hung et al., 2008). Bilirubin, the downstream product of HO-1, also exerts the similar neuroprotective effect (Hung et al., 2008). Here, we further found that neutralization of endogenously released BDNF/GDNF with respective antibody resulted in 50% dopaminergic neuronal death in midbrain neuron-glia cocultures of E14 rats. Over-expression of HO-1 rescued the neuronal death. These results indicate that the neuroprotective action of HO-1 may not only due to its antioxidant and anti-inflammation action but also result from the regulation of BDNF and GDNF expression. In mammalian CNS, HO-2 is abundantly and constitutively expressed, whereas HO-1 is found to be expressed in a small population of neurons and glia or induced by cellular stress (Scapagnini et al., 2002; Schipper, 2004). Heme oxygenase in brain has been reported to be active and play a crucial role in the pathogenesis of Parkinson disease (Miwa et al., 2004; Schipper et al., 1998). Bilirubin at nanomolar concentration is also a potent antioxidant and neuroprotectant (Dore et al., 1999). In this study, the glia cells were cultured in DMEM with 10% FBS, which contained 41.5 mM albumin. Since bilirubin can bind with albumin, the free form of 10 mM bilirubin is about 99 nM, according to the calculation reported by Weisiger et al. (Weisiger et al., 2001). Bilirubin (3e30 mM) or [Ru(CO)3Cl2]2 (10e100 mM) treatment for 24 h doesn't influence the survival rate or the expression of HO-1 and HO-2 in glia-enriched cultures, indicating that the doses used in bilirubin or [Ru(CO)3Cl2]2 did not cause cellular stress.
Fig. 7. Bilirubin and [Ru(CO)3Cl2]2 increase BDNF expression in cortical neuron cultures. Primary cortical neurons were derived from E17 rats and cultured for 7e10 days before experiments. (A) Bilirubin (10 mM) time-dependently increased ERK and Akt phosphorylation. (B) Co-treatment with MEK inhibitor PD98059 or PI3K inhibitor LY294002 (20 mM for each) antagonized bilirubin-induced BDNF expression after 24 h incubation. (C) Co-treatment with sGC inhibitor ODQ (20 mM) antagonized [Ru(CO)3Cl2]2 (30 mM)-induced BDNF expression. The summarized results of protein expression were shown at right panels. Data are given as mean SEM (n ¼ 3). *p < 0.05 compared with vehicle control. #p < 0.05 compared with bilirubin or [Ru(CO)3Cl2]2 treatment alone.
S.-Y. Hung et al. / Neuropharmacology 58 (2010) 321e329
BDNF and GDNF have been reported to increase the survival and morphology differentiation of dopaminergic neurons (Kirschner et al., 1996; Lin et al., 1993). We here found that neutralization of endogenous release of BDNF or GDNF by specific antibody significantly enhanced dopaminergic neuronal death. Treatment of glia cultures with bilirubin concentration-dependently enhanced GDNF expression via ERK and Akt phosphorylation. Furthermore, bilirubin treatment also induced nuclear translocation of p50 and p65. Furthermore, NFkB inhibitor PDTC antagonized bilirubin-induced GDNF expression in glia cultures. In addition, bilirubin also enhanced BDNF expression in cortical neuron cultures via similar pathway. Fluoxetine has also been reported to induce BDNF and GDNF mRNA expression in rat astrocyte via the activation of MEKERK pathway (Mercier et al., 2004). Rasagiline (N-Propargyl-l(R)aminoindan; an anti-Parkinson drug) also increases GDNF level via the activation of ERK1/2 and p65 nuclear translocation in human neuroblastoma SH-SY5Y (Maruyama et al., 2004). Dibutyryl cGMP, an analog of cGMP, preferentially activates PKG and elevates BDNF content from dopaminergic cell line SN4741 (Chun et al., 2000). Here we also found that the CO-releasing molecule, [Ru(CO)3Cl2]2, enhanced BDNF and GDNF expression in both neuron and glia through sGC-PKG pathway. These results suggest that NO and CO may regulate BDNF and GDNF expression through cGMP-PKG pathway. In conclusion, HO-1 induction exerts neuroprotection through its downstream products, i.e. bilirubin and CO. The low concentration of free-form bilirubin can exert its protective action on dopaminergic neuron via GDNF or BDNF induction, which may be mediated through MEK, Akt and NFkB activation. CO also increases the expression of neurotrophic factor via sGC-PKG-dependent pathway. Acknowledgement We thank Dr. Lee-Young Chau for providing the Ad-HO-1. References Baranano, D.E., Rao, M., Ferris, C.D., Snyder, S.H., 2002. Biliverdin reductase: a major physiologic cytoprotectant. Proc. Natl. Acad. Sci. U S A 99, 16093e16098. Chen, K., Gunter, K., Maines, M.D., 2000. Neurons overexpressing heme oxygenase-1 resist oxidative stress-mediated cell death. J. Neurochem. 75, 304e313. Chen, P.S., Peng, G.S., Li, G., Yang, S., Wu, X., Wang, C.C., Wilson, B., Lu, R.B., Gean, P.W., Chuang, D.M., Hong, J.S., 2006. Valproate protects dopaminergic neurons in midbrain neuron/glia cultures by stimulating the release of neurotrophic factors from astrocytes. Mol. Psychiatry 11, 1116e1125. Choi-Lundberg, D.L., Lin, Q., Chang, Y.N., Chiang, Y.L., Hay, C.M., Mohajeri, H., Davidson, B.L., Bohn, M.C., 1997. Dopaminergic neurons protected from degeneration by GDNF gene therapy. Science 275, 838e841. Chun, H.S., Son, J.J., Son, J.H., 2000. Identification of potential compounds promoting BDNF production in nigral dopaminergic neurons: clinical implication in Parkinson's disease. Neuroreport 11, 511e514. Dore, S., Takahashi, M., Ferris, C.D., Zakhary, R., Hester, L.D., Guastella, D., Snyder, S.H., 1999. Bilirubin, formed by activation of heme oxygenase-2, protects neurons against oxidative stress injury. Proc. Natl. Acad. Sci. U S A 96, 2445e2450. Hornykiewicz, O., 1988. Neurochemical pathology and the etiology of Parkinson's disease: basic facts and hypothetical possibilities. Mt. Sinai J. Med. 45, 35e43. Hu, C.M., Chen, Y.H., Chiang, M.T., Chau, L.Y., 2004. Heme oxygenase-1 inhibits angiotensin II-induced cardiac hypertrophy in vitro and in vivo. Circulation 110, 309e316. Hung, S.Y., Liou, H.C., Kang, K.H., Wu, R.M., Wen, C.C., Fu, W.M., 2008. Overexpression of heme oxygenase-1 protects dopaminergic neurons against 1-methyl-4-phenylpyridinium-induced neurotoxicity. Mol. Pharmacol. 74, 1564e1575. Jenner, P., 2003. Oxidative stress in Parkinson's disease. Ann. Neurol. 53 (Suppl. 3), S26eS38. Juan, S.H., Lee, T.S., Tseng, K.W., Liou, J.Y., Shyue, S.K., Wu, K.K., Chau, L.Y., 2001. Adenovirus-mediated heme oxygenase-1 gene transfer inhibits the development of atherosclerosis in apolipoprotein E-deficient mice. Circulation 104, 1519e1525. Kim, H.P., Wang, X., Galbiati, F., Ryter, S.W., Choi, A.M., 2004. Caveolae compartmentalization of heme oxygenase-1 in endothelial cells. Faseb J. 18, 1080e1089.
329
Kirschner, P.B., Jenkins, B.G., Schulz, J.B., Finkelstein, S.P., Matthews, R.T., Rosen, B.R., Beal, M.F., 1996. NGF, BDNF and NT-5, but not NT-3 protect against MPPþ toxicity and oxidative stress in neonatal animals. Brain Res. 713, 178e185. Le, W.D., Xie, W.J., Appel, S.H., 1999. Protective role of heme oxygenase-1 in oxidative stress-induced neuronal injury. J. Neurosci. Res. 56, 652e658. Lin, L.F., Doherty, D.H., Lile, J.D., Bektesh, S., Collins, F., 1993. GDNF: a glial cell linederived neurotrophic factor for midbrain dopaminergic neurons. Science 260, 1130e1132. Lu, D.Y., Yu, W.H., Yeh, W.L., Tang, C.H., Leung, Y.M., Wong, K.L., Chen, Y.F., Lai, C.H., Fu, W.M., 2009. Hypoxia-induced matrix metalloproteinase-13 expression in astrocytes enhances permeability of brain endothelial cells. J. Cell Physiol. 220, 163e173. Marks, G.S., 1994. Heme oxygenase: the physiological role of one of its metabolites, carbon monoxide and interactions with zinc protoporphyrin, cobalt protoporphyrin and other metalloporphyrins. Cell Mol. Biol. (Noisy-le-grand) 40, 863e870. Maruyama, W., Nitta, A., Shamoto-Nagai, M., Hirata, Y., Akao, Y., Yodim, M., Furukawa, S., Nabeshima, T., Naoi, M., 2004. N-Propargyl-1 (R)-aminoindan, rasagiline, increases glial cell line-derived neurotrophic factor (GDNF) in neuroblastoma SH-SY5Y cells through activation of NF-kappaB transcription factor. Neurochem. Int. 44, 393e400. McKinstry, R., Qiao, L., Yacoub, A., Dai, Y., Decker, R., Holt, S., Hagan, M.P., Grant, S., Dent, P., 2002. Inhibitors of MEK1/2 interact with UCN-01 to induce apoptosis and reduce colony formation in mammary and prostate carcinoma cells. Cancer Biol. Ther. 1, 243e253. Mercier, G., Lennon, A.M., Renouf, B., Dessouroux, A., Ramauge, M., Courtin, F., Pierre, M., 2004. MAP kinase activation by fluoxetine and its relation to gene expression in cultured rat astrocytes. J. Mol. Neurosci. 24, 207e216. Miwa, H., Kubo, T., Morita, S., Nakanishi, I., Kondo, T., 2004. Oxidative stress and microglial activation in substantia nigra following striatal MPPþ. Neuroreport 15, 1039e1044. Motterlini, R., Clark, J.E., Foresti, R., Sarathchandra, P., Mann, B.E., Green, C.J., 2002. Carbon monoxide-releasing molecules: characterization of biochemical and vascular activities. Circ. Res. 90, E17eE24. Ollinger, R., Kogler, P., Troppmair, J., Hermann, M., Wurm, M., Drasche, A., Konigsrainer, I., Amberger, A., Weiss, H., Ofner, D., Bach, F.H., Margreiter, R., 2007. Bilirubin inhibits tumor cell growth via activation of ERK. Cell Cycle 6, 3078e3085. Otterbein, L.E., Bach, F.H., Alam, J., Soares, M., Tao Lu, H., Wysk, M., Davis, R.J., Flavell, R.A., Choi, A.M., 2000. Carbon monoxide has anti-inflammatory effects involving the mitogen-activated protein kinase pathway. Nat. Med. 6, 422e428. Otterbein, L.E., Zuckerbraun, B.S., Haga, M., Liu, F., Song, R., Usheva, A., Stachulak, C., Bodyak, N., Smith, R.N., Csizmadia, E., Tyagi, S., Akamatsu, Y., Flavell, R.J., Billiar, T.R., Tzeng, E., Bach, F.H., Choi, A.M., Soares, M.P., 2003. Carbon monoxide suppresses arteriosclerotic lesions associated with chronic graft rejection and with balloon injury. Nat. Med. 9, 183e190. Pachori, A.S., Smith, A., McDonald, P., Zhang, L., Dzau, V.J., Melo, L.G., 2007. Hemeoxygenase-1-induced protection against hypoxia/reoxygenation is dependent on biliverdin reductase and its interaction with PI3K/Akt pathway. J. Mol. Cell Cardiol. 43, 580e592. Roberts, P.J., Der, C.J., 2007. Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer. Oncogene 26, 3291e3310. Roberts, P.J., Kapur, H., 1977. Neurochemical changes in rat striatum and substantia nigra following drug administration. Neurochem. Res. 2, 485e494. Ryter, S.W., Alam, J., Choi, A.M., 2006. Heme oxygenase-1/carbon monoxide: from basic science to therapeutic applications. Physiol. Rev. 86, 583e650. Scapagnini, G., D'Agata, V., Calabrese, V., Pascale, A., Colombrita, C., Alkon, D., Cavallaro, S., 2002. Gene expression profiles of heme oxygenase isoforms in the rat brain. Brain Res. 954, 51e59. Schipper, H.M., 2000. Heme oxygenase-1: role in brain aging and neurodegeneration. Exp. Gerontol. 35, 821e830. Schipper, H.M., 2004. Heme oxygenase expression in human central nervous system disorders. Free Radic. Biol. Med. 37, 1995e2011. Schipper, H.M., Cisse, S., Stopa, E.G., 1995. Expression of heme oxygenase-1 in the senescent and Alzheimer-diseased brain. Ann. Neurol. 37, 758e768. Schipper, H.M., Liberman, A., Stopa, E.G., 1998. Neural heme oxygenase-1 expression in idiopathic Parkinson's disease. Exp. Neurol. 150, 60e68. Sindhu, K.M., Banerjee, R., Senthilkumar, K.S., Saravanan, K.S., Raju, B.C., Rao, J.M., Mohanakumar, K.P., 2006. Rats with unilateral median forebrain bundle, but not striatal or nigral, lesions by the neurotoxins MPPþ or rotenone display differential sensitivity to amphetamine and apomorphine. Pharmacol. Biochem. Behav. 84, 321e329. Spenger, C., Hyman, C., Studer, L., Egli, M., Evtouchenko, L., Jackson, C., DahlJorgensen, A., Lindsay, R.M., Seiler, R.W., 1995. Effects of BDNF on dopaminergic, serotonergic, and GABAergic neurons in cultures of human fetal ventral mesencephalon. Exp. Neurol. 133, 50e63. Tenhunen, R., Marver, H.S., Schmid, R., 1969. Microsomal heme oxygenase. Characterization of the enzyme. J. Biol. Chem. 244, 6388e6394. Weisiger, R.A., Ostrow, J.D., Koehler, R.K., Webster, C.C., Mukerjee, P., Pascolo, L., Tiribelli, C., 2001. Affinity of human serum albumin for bilirubin varies with albumin concentration and buffer composition: results of a novel ultrafiltration method. J. Biol. Chem. 276, 29953e29960.