Microglia-Mediated Neurotoxicity Is Inhibited by through an Receptor-Independent Reduction of NADPH Oxidase Activity This information is current as of September 28, 2021. Li Qian, Kai Soo Tan, Sung-Jen Wei, Hung-Ming Wu, Zongli Xu, Belinda Wilson, Ru-Bin Lu, Jau-Shyong Hong and Patrick M. Flood J Immunol 2007; 179:1198-1209; ; doi: 10.4049/jimmunol.179.2.1198 Downloaded from http://www.jimmunol.org/content/179/2/1198

References This article cites 62 articles, 21 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/179/2/1198.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Microglia-Mediated Neurotoxicity Is Inhibited by Morphine through an -Independent Reduction of NADPH Oxidase Activity1

Li Qian,*† Kai Soo Tan,*§ Sung-Jen Wei,¶ Hung-Ming Wu,†ʈ Zongli Xu,‡ Belinda Wilson,† Ru-Bin Lu,ʈ Jau-Shyong Hong,† and Patrick M. Flood2*

Recent studies have shown that morphine modulates the function of glia cells through both opioid receptor dependent and independent mechanisms. However, the mechanism by which morphine regulates neuronal disorders through the alteration of microglia activity remains unclear. In this study, using rat primary mesencephalic neuron-glia cultures, we report that both l-morphine and its synthetic stereoenantiomer, d-morphine, an ineffective opioid receptor agonist, significantly reduced LPS- or

1-methyl-4-phenylpyridinium-induced dopaminergic neurotoxicity with similar efficacy, indicating a nonopioid receptor-mediated Downloaded from effect. In addition, using reconstituted neuron and glia cultures, subpicomolar concentrations of morphine were found to be neuroprotective only in the presence of microglia, and significantly inhibited the production of inflammatory mediators from LPS-stimulated microglia cells. Mechanistic studies showed that both l- and d- morphine failed to protect dopaminergic neurons in cultures from NADPH oxidase (PHOX) knockout mice and significantly reduced LPS-induced PHOX cytosolic subunit p47phox translocation to the cell membrane by inhibiting ERK phosphorylation. Taken together, our results demonstrate that morphine, even at subpicomolar concentrations, exerts potent anti-inflammatory and neuroprotective effects either through the inhibition of http://www.jimmunol.org/ direct microglial activation by LPS or through the inhibition of reactive microgliosis elicited by 1-methyl-4-phenylpyridinium. Furthermore, our study reveals that inhibition of PHOX is a novel site of action for the mu-opioid receptor-independent effect of morphine. The Journal of Immunology, 2007, 179: 1198–1209.

arkinson’s disease (PD)3 is a neurodegenerative move- exact mechanism and kinetics by which microglia mediate neuro- ment disorder characterized by the progressive loss of do- toxicity and the signals that control their proinflammatory response P paminergic (DA) neurons in the substantia nigra and de- remain to be elucidated. pletion of dopamine in the striatum, which leads to pathological The naturally occurring form of morphine, l-morphine, is a po- by guest on September 28, 2021 and clinical abnormalities. The midbrain region that encompasses tent and is used in the management of moder- the substantia nigra is particularly rich in microglia (1), and recent ately severe to severe pain (4). The analgesic effect of l-morphine studies have shown that dysregulated microglia activation can re- is believed to be mediated through the mu-opioid receptor, which sult in neurodengeration (2, 3). Activation of nigral microglia and is a member of the seven-transmembrane-spanning G-protein-cou- the subsequent release of proinflammatory neurotoxic factors, in- pled receptor family (4). Increasing in vitro and in vivo evidence ␤ ␣ cluding IL-1 , TNF- , PGE2, IL-6, NO, and superoxide, are con- have demonstrated wide-spectrum effects of morphine including sidered key components of DA neuron degeneration in PD (2). The anti-inflammatory, antifibrotic, antitumor, cardioprotective, and renoprotective (5–9). Morphine has also been shown to possess *Comprehensive Center for Inflammatory Disorders, University of North Carolina, immunomodulation function, regulating the activities of immune Chapel Hill, NC 27599; †Neuropharmacology Section, Laboratory of Pharmacology cells including NK cells, T cells, B cells, monocytes, and polymor- and Chemistry, ‡Epidemiology Branch, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC 27709; §Depart- phonuclear leukocytes (10, 11). In addition, morphine has been re- ment of Biochemistry, Yong Loo Lin School of Medicine, National University of ported to exert anti-inflammatory effects by reducing adhesion mol- Singapore, Kent Ridge, Singapore; ¶The National Center for Toxicogenomics, Na- ecule expression, inhibiting cell trafficking and respiratory burst tional Institute of Environmental Health Sciences, Research Triangle Park, NC 27709; and ʈDepartment of Psychiatry, College of Medicine and Hospital, National Cheng- activity, as well as suppressing the activation of transcription fac- Kung University, Tainan, Taiwan tor NF-␬B (12). Morphine can exert its potent anti-inflammatory Received for publication November 13, 2006. Accepted for publication May 5, 2007. effects at very low concentrations. Concentrations of morphine as Ϫ Ϫ The costs of publication of this article were defrayed in part by the payment of page low as 10 10–10 12 M can cause significant suppression of PMA charges. This article must therefore be hereby marked advertisement in accordance or opsonized zymosan-stimulated superoxide release in monocytes with 18 U.S.C. Section 1734 solely to indicate this fact. 1 (13, 14), whereas inhibition of phagocytosis of Cryptococcus neo- This work was supported by National Institutes of Health Grant DE-13079 from the Ϫ16 National Institute for Dental and Craniofacial Research and in part by the Intramural formans by microglia was observed at 10 M (11). Research Program of the National Institutes of Health/National Institute of Environ- These observations suggest that morphine has the potential to be mental Health Sciences. used as an immune regulator to treat chronic inflammatory condi- 2 Address correspondence and reprint requests to Dr. Patrick M. Flood, Comprehen- tions, provided that these treatments do not lead to chronic depen- sive Center for Inflammatory Disorders, University of North Carolina, Chapel Hill, NC 27599-7455. E-mail address: pat_fl[email protected] dence. Consequently, there is an urgent need to identify and un- 3 Abbreviations used in this paper: PD, Parkinson’s disease; MPPϩ, 1-methyl-4-phe- derstand the molecular mechanisms of morphine that function nylpyridinium; DA, dopaminergic; ROS, reactive oxygen species; iNOS, inducible independently of the mu-opioid receptor, to prevent narcotic ad- nitric oxide synthase; TH-IR, tyrosine hydroxylase immunoreactive; PHOX, NADPH oxidase; DCFH-DA, dichlorodihydrofluorescein diacetate; SOD, superoxide dis- diction resulting from treatment for these chronic inflammatory mutase; MFI, mean fluorescent intensity. conditions. Recent studies have shown that morphine can work in www.jimmunol.org The Journal of Immunology 1199

5 both opioid receptor-dependent and -independent manners (11, 10 /well in poly-D-lysine-coated 96-well and 24-well plates, respectively. 15–19). Further, our laboratory and others have found that nalox- Cells were maintained at 37°C in a humidified atmosphere of 5% CO2 and one and , which are mu-opioid receptor antagonists, also 95% air, in MEM containing 10% FBS, 10% horse serum, 1 g/L glucose, 2mML-glutamine, 1 mM sodium pyruvate, 100 ␮M nonessential amino show significant anti-inflammatory properties both in vitro and in acids, 50 U/ml penicillin, and 50 ␮g/ml streptomycin. Seven-day-old cul- vivo (20, 21). In addition, these compounds provide neuroprotec- tures were used for drug treatments. At the time of treatment, immunocy- tion in ischemic injury, spinal cord injury, PD, and brain ischemia tochemical analysis indicated that the rat neuron-glia cultures were made animal models (22–24). Although the analgesic effect of l-mor- up of 11% microglia, 48% astrocytes, 41% neurons, and 1% tyrosine hy- droxylase immunoreacitve (TH-IR) neurons. The composition of the neu- phine is believed to be mediated through the mu-opioid receptor, ron-glia cultures of NADPH oxidase-deficient mice was very similar to that whether the effects of the receptor antagonists and nal- of the wild-type mice, which consists of 12% microglia, 48% astrocytes, trexone are mediated through the mu-opioid receptor remains con- 40% neurons, and 1% TH-IR neurons. troversial. Therefore, we used both l-morphine, which has a high Primary mesencephalic neuron-enriched cultures affinity to the mu-opioid receptor, and its synthetic enantiomer d-morphine, which does not bind to the mu-opioid receptor (25), to Midbrain neuron-enriched cultures were established as described previ- ␤ explore the underlying molecular mechanism of the protective ef- ously (28). Briefly, 24 h after seeding, cytosine -D-arabinoside was added to a final concentration of 10 ␮M to suppress glial proliferation. Three days fect of morphine on DA neurons in two in vitro models of PD. In later, the medium was removed and replaced with maintenance medium. this study, we show that subpicomolar concentrations of both l- Cells were used for drug treatments 7 days after initial seeding. Routinely, morphine and d-morphine significantly down-regulate LPS- or the 7-day-old neuron-enriched cultures, which normally contain Ͻ0.1% 1-methyl-4-phenylpyridinium (MPPϩ)-induced microglia activa- microglia, and Ͻ3–5% astrocytes, were used for treatment. Among the neuronal population (Neu-N immunoreactive neurons), 2.7–3.9% were do- tion with similar efficacy, clearly showing that the cellular mech- Downloaded from paminergic neurons (TH-IR-positive neurons). anism of morphine-mediated neuroprotection is independent of the mu-opioid receptor pathway. Additionally, we show that the mor- Primary microglia-enriched cultures phine-mediated molecular effects result in the inhibition of Rat microglia-enriched cultures with a purity of Ͼ98%, were prepared NADPH oxidase (PHOX) activity in activated microglia via the from whole brains of 1-day-old F344 rat pups as described previously (29). inhibition of the ERK signaling pathway. For superoxide assays, 105 cells were grown overnight in 96-well culture plates before use. http://www.jimmunol.org/ Materials and Methods Rat astroglial cultures Animals Mixed-glia cultures were first prepared from brains of 1-day-old F344 rat NADPH oxidase-deficient (gp91phoxϪ/Ϫ) and wild-type C57BL/6J pups, as described previously (30). Briefly, mechanically dissociated brain (gp91phoxϩ/ϩ) mice were obtained from The Jackson Laboratory. The cells (5 ϫ 107) were seeded onto 150-cm2 culture flasks in DMEM con- Ϫ/Ϫ PHOX mutation is maintained in the C57BL/6 background; thus, C57BL/6 taining 10% heat-inactivated FBS, 2 mM L-glutamine, 1 mM sodium pyru- mice were used as control animals. Breeding of the mice was performed to vate, 100 ␮M nonessential amino acids, 50 U/ml penicillin, and 50 ␮g/ml achieve timed pregnancy with the accuracy of Ϯ0.5 days. Timed-pregnant streptomycin. The cultures were maintained at 37°C in a humidified atmo-

F344 rats were obtained from Charles River Laboratories. Housing and sphere of 5% CO2 and 95% air, and medium was replenished 4 days after breeding of the animals were performed in strict accordance with the Na- the initial seeding. Upon reaching confluence (usually 12–14 days later), tional Institutes of Health guidelines. microglia were detached from astrocytes by shaking the flasks at a speed of by guest on September 28, 2021 180 rpm for 5 h. Astrocytes were then detached with trypsin-EDTA and Reagents seeded in the same culture medium. After five or more consecutive pas- sages, cells were seeded onto 24-well plates (105/well) for experiments. l-Morphine sulfate and d-morphine base were obtained from National In- Immunocytochemical staining of the astroglial cultures with either antiglial stitute of Drug Abuse. LPS (Escherichia coli strain O111:B4) was pur- fibrillary acidic protein or anti-OX42 Ab indicated an astrocyte purity of chased from Calbiochem. Cell culture reagents were obtained from Invitro- Ͼ98% and Ͻ2% of microglia contamination. gen Life Technologies. [3H]DA (30 Ci/mmol) was obtained from PerkinElmer Life Sciences, and the mAb against the complement 3 recep- [3H]DA uptake assay tor (OX42) was purchased from Chemicon International. The polyclonal anti-tyrosine hydroxylase Ab was a gift from Dr. John Reinhard (Glaxo, [3H]DA uptake assays were performed as described (29). Briefly, cells Durham, NC). The Vectastain ABC kit and biotinylated secondary Abs were incubated for 20 min at 37°C with 1 ␮M[3H]DA in Krebs-Ringer were purchased from Vector Laboratories. The fluorescence probe dichlo- buffer (16 mM sodium phosphate, 119 mM NaCl, 4.7 mM KCl, 1.8 mM rodihydrofluorescein diacetate (DCFH-DA) was obtained from Calbio- CaCl2, 1.2 mM MgSO4, 1.3 mM EDTA, and 5.6 mM glucose; pH 7.4). chem. Rabbit anti-p47phox was purchased from Upstate. FITC-conjugated Cells were washed with ice-cold Krebs-Ringer buffer three times, after goat anti-rabbit IgG was obtained from Jackson ImmunoResearch Labo- which the cells were collected in 1 N NaOH. Radioactivity was determined ratories. Rabbit anti-GAPDH was obtained from Abcam. Mouse anti- by liquid scintillation counting. Nonspecific DA uptake observed in the gp91phox was purchased from BD Transduction Laboratories. presence of mazindol (10 ␮M) was subtracted. Microglial cell line Immunocytochemistry The rat microglia HAPI cell line was a gift from Dr. James R. Connor Immunostaining was performed as previously described (27). Briefly, (Department of Neuroscience and Anatomy, M. S. Hershey Medical Cen- formaldehyde (3.7%)-fixed cultures were treated with 1% hydrogen per- ter, Hershey, PA) (26). Briefly, cells were maintained at 37°C in DMEM oxide followed by sequential incubation with blocking solution, after supplemented with 10% FBS, 50 U/ml penicillin, and 50 ␮g/ml strepto- which the cells were incubated overnight at 4°C with Abs against tyrosine mycin in a humidified incubator with 5% CO2 and 95% air. The HAPI cell hydroxylase (1/20,000). Cells were incubated with biotinylated secondary was derived from rat primary microglia-enriched cultures and retains the Ab for 2 h followed by incubation with ABC reagents for 40 min. Color phenotypic and morphological characteristics of microglia, including was developed with 3,3Ј-diaminobenzidine. For morphological analysis, phagocytic activity, expression of the receptor for isolectin B4 from Grif- the images were recorded with an inverted microscope (Nikon) connected fonia simplicifolia, as well as the specific microglial markers OX42 (com- to a charge-coupled device camera (DAGE-MTI) operated with the Meta- plement 3 receptor) and glucose transport protein 5. Gene expression of Morph software (Universal Imaging). For visual counting of TH-IR neu- TNF-␣ and inducible NO synthase (iNOS) can be induced by treatment rons, nine representative areas per well of the 24-well plate were counted with LPS. under the microscope at ϫ100 by three individuals. The average of these scores was reported. Primary mesencephalic neuron-glia cultures Superoxide assay Neuron-glia cultures were prepared from the ventral mesencephalic tissues of embryonic day 14–15 rats or day 13–14 mice, as described previously The production of superoxide was determined by measuring the superoxide (24, 27). Briefly, dissociated cells were seeded at 1 ϫ 105/well and 5 ϫ dismutase (SOD)-inhibitable reduction of the tetrazolium salt WST-1 (31, 1200 MORPHINE PROTECTS DA NEURON AGAINST OXIDATIVE STRESS Downloaded from

FIGURE 1. l- and d- protect neuron against LPS-induced neurotoxicity with similar efficacy. Rat primary mesencephalic neuron-glia cultures http://www.jimmunol.org/ were seeded in a 24-well culture plate at 5 ϫ 105 cells/well and then pretreated with various concentrations of morphine for 30 min before the addition of 10 ng/ml LPS. Seven days later, the LPS-induced dopaminergic neurotoxicity was quantified by the [3H]DA uptake assay (A), the immunocytochemical analysis, including TH-IR neuron counts (B), and the representative pictures of immunostaining (C). Results were expressed as a percentage of the p Ͻ 0.01 compared with the ,ءء ;p Ͻ 0.05 ,ء .vehicle-treated control cultures and were the means Ϯ SE. from three independent experiments in triplicate LPS-treated cultures.

32). Neuron-glia or microglia-enriched cultures in 96-well culture plates final concentration) were added to each well along with 50 ␮l of WST-1 (1 were washed twice with HBSS without phenol red. Cultures were then mM) in HBSS and with 50 ␮l of vehicle or LPS (10 ng/ml). To measure incubated at 37°C for 30 min with vehicle control (water) or morphine in superoxide production induced by MPPϩ, 7-day-old mesencephalic neu- by guest on September 28, 2021 HBSS (50 ␮l/well). Then, 50 ␮l of HBSS with and without SOD (50 U/ml, ron-glia cultures grown in 96-well plates were treated with morphine in the

FIGURE 2. l- and d-morphine protects neuron against MPPϩ induced neurotoxicity. Different doses of morphine were added to the neuron-glial cultures for 30 min before the addition of 0.1 ␮M MPPϩ treatment. The [3H]DA uptake measurements were performed 8 days after MPPϩ treatment (A); the immunocytochemical analysis, including TH-IR neuron counts (B), and the representative pictures of immunostaining (C) were shown. Results were ;p Ͻ 0.05 ,ء .expressed as a percentage of the vehicle-treated control cultures and were the means Ϯ SE. from four independent experiments in triplicate .p Ͻ 0.01 compared with the LPS-treated cultures ,ءء The Journal of Immunology 1201

FIGURE 3. Microglia are essential for morphine-elicited neuroprotec- tion. l- and d-morphines (10Ϫ14 M) were added to the following cultures before treatment with 0.1 ␮M MPPϩ. N, Neuron-enriched cultures; (N ϩ 10%MG): reconstituted cultures by adding 10% of microglia to the neuron- enriched cultures; and (N ϩ 50%AS): reconstituted cultures by adding 50% of astroglia to the neuron-enriched cultures. The [3H]DA uptake mea- surements were performed 8 days after the treatment. Results were ex-

pressed as a percentage of the vehicle-treated control cultures and were the Downloaded from p Ͻ 0.05 ,ء .means Ϯ SE from four independent experiments in triplicate compared with the LPS-treated cultures.

presence and absence of MPPϩ or with vehicle alone in 150 ␮l of phenol red-free treatment medium. Four days after treatment, 50 ␮l of HBSS with

and without SOD (50 U/ml, final concentration) were added to each well http://www.jimmunol.org/ ␮ along with 50 l of WST-1 (1 mM) in HBSS. Fifteen minutes later, OD450 was read with a SpectraMax Plus microplate spectrophotometer (Molecular Devices). The difference in optical density observed in the presence and absence of SOD was considered to be the amount of superoxide produced, and results were expressed as the percentage of vehicle-treated control cultures. Intracellular reactive oxygen species (ROS) assay Intracellular ROS were determined by using a DCFH-DA assay as de-

scribed previously with minor modifications (33). DCFH-DA enters cells by guest on September 28, 2021 passively and is deacetylated by esterase to nonfluorescent DCFH. DCFH reacts with ROS to form dichlorodihydrofluorescein; the fluorescent prod- ϩ FIGURE 4. Morphine inhibits both LPS- and MPP -induced ROS pro- uct DCFH-DA was dissolved in methanol at 10 mM and was diluted 500- fold in HBSS to give DCFH-DA at 20 ␮M. The cells were exposed to duction by activated microglia. Microglia-enriched cultures were seeded at 5 DCFH-DA for 1 h and then treated with HBSS containing the correspond- a density of 1 ϫ 10 cells/well. Cells were pretreated with different con- ing concentrations of LPS for 2 h. The fluorescence was read immediately centrations of morphine for 30 min followed by the addition of LPS (10 at wavelengths of 485 nm for excitation and 530 nm for emission using a ng/ml). The production of ROS included extracellular superoxide (A) and SpectraMax Gemini XS fluorescence microplate reader (Molecular De- intracellular ROS (B). Extracellular superoxide was measured as SOD- vices). The experimental value minus the value of the control group was inhibitable reduction of WST-1, and intracellular ROS was determined by interpreted as the increase in intracellular ROS. probe DCFH-DA. Primary rat mesencephalic neuron-glia cultures were Real-time RT-PCR analysis pretreated for 30 min with vehicle or morphine before treatment with 0.25 ␮M MPPϩ. Two and four days after MPPϩ treatment, morphine was added For real-time quantitative PCR, DNA-free RNA was obtained using the again to the morphine-treated cultures. On day 4, the release of superoxide RNeasy Mini Kit (Qiagen). We reverse-transcribed 1 ␮g of total RNA was determined as described above (C). Superoxide production was ex- using random hexamers and Moloney murine leukemia virus reverse tran- pressed as a percentage of the vehicle-treated control cultures and as the ,ءء ;p Ͻ 0.05 ,ء .scriptase. The SYBR green DNA PCR kit (Applied Biosystems) was used for mean Ϯ SE. of three experiments performed in triplicate real-time PCR analysis. The relative differences in expression between groups ϩ p Ͻ 0.01 compared with corresponding LPS or MPP -treated cultures. were expressed using cycle time values normalized to GADPH. The relative differences between control and treatment groups were calculated and ex- pressed as relative increases setting control as 100%. The sequence of the oligonucleotide primers used were from rat: TNF-␣,5Ј-TCGTAGCAAACC ACCAAGCA-3Ј and 5Ј-CCCTTGAAGAGAACCTGGGAGTA-3Ј; iNOS, p47phox (0.5 ␮g/ml). Cells were then washed and incubated with FITC- 5Ј-ACATCAGGTCGGCCATCACT-3Ј and 5Ј-CGTACCGGATGAGCTG conjugated goat anti-rabbit Ab. Focal planes spaced at 0.4-␮m intervals TGAATT-3Ј; IL-6, 5Ј-GCCCACCAAGAACGATAGTCA-3Ј and 5Ј-GA were imaged with a Zeiss 510 laser scanning confocal microscope (ϫ63 AGGCAACTGGATGGAAGTCT-3Ј;IL-1␤,5Ј-CTGGTGTGTGACGTTC PlanApo 1.4 numerical aperture objective) equipped with LSM510 digital CCATTA-3Ј and 5Ј-CCGACAGCACGAGGCTTT-3Ј; MIP-1␣,5Ј-GCTT imaging software. The signal of p47phox (FITC-p47phox; green) and the GAGCCCCAGAACATTC and GATGTGGCTACTTGGCAGCAA-3Ј; merge view of cell morphology and p47phox (phase plus FITC-p47phox) are GP91, 5Ј-TTTGCCGGAAACCCTCCTAT-3Ј and 5Ј-GCACCAGCCTCT shown. CACACAGA-3Ј; GAPDH, 5Ј-CCTGGAGAAACCTGCCAAGTAT-3Ј and 5Ј-AGCCCAGGATGCCCTTTAGT-3Ј. Membrane fractionation and Western blot analysis Confocal microscopy Membrane fractionation was performed as described (34). HAPI cells were lysed in hypotonic lysis buffer (1 mM EGTA, 1 mM EDTA, 10 mM HAPI cells seeded in dishes at 5 ϫ 104 cells/well were treated with LPS for ␤-glycerophosphate, 10 mM NaF, 1 mM sodium orthovanadate, 2 mM ␮ 10 min in the presence and absence of morphine pretreatment for 30 min. MgCl2, 10 mM DTT, 1 mM PMSF, and 10 g/ml each leupeptin, aproti- Cells were fixed with 3.7% paraformaldehyde in PBS for 10 min. After a nin, and pepstatin A), incubated on ice for 30 min, and then subjected to washing with PBS, cells were incubated with rabbit polyclonal Ab against Dounce homogenization (20–25 stokes, tight pestle A). The lysates were 1202 MORPHINE PROTECTS DA NEURON AGAINST OXIDATIVE STRESS Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 5. Effects of morphine on LPS-induced mRNA expression of proinflammatory factors and microglial surface molecular expression. A, Mi- croglia-enriched cultures seeded at 1 ϫ 106/well were pretreated with morphine for 30 min followed by the addition of LPS (5 ng/ml). The effect of morphine on LPS-induced iNOS, TNF-␣, MIP-1␣, IL-6, IL-1␤, gp91 RNA mRNA expression was detected by real-time PCR 3 h after LPS stimulation as described in Materials and Methods. B, HAPI cells were pretreated with morphine (10Ϫ14 M) for 30 min followed by stimulation with LPS for 24 h. Expression of OX18 (MHC class I), OX6 (MHC class II), and costimulatory molecules CD40 and CD80 were detected by flow cytometry. The cells were analyzed on a FACSCalibur, and the MFI of experimental groups was determined by subtracting the MFI of the isotype control from the MFI of the p Ͻ 0.01 compared with corresponding ,ءء ;p Ͻ 0.05 ,ء .unstimulated or LPS-stimulated microglia. Data are representative of four separate experiments LPS-treated cultures; #, p Ͻ 0.05, ##, p Ͻ 0.01 compared with vehicle-treated control cultures. loaded onto sucrose in lysis buffer (final concentration, 0.5 M) and centri- Nonidet P-40 hypotonic lysis buffer were used as membranous fraction. fuged at 1600 ϫ g for 15 min; the supernatant above the sucrose gradient Equal amounts of protein (20 ␮g/lane) were separated by 4–12% Bis-Tris was centrifuged at 150,000 ϫ g for 30 min. The pellets solubilized in 1% Nu-PAGE gel and transferred to polyvinylidene difluoride membranes The Journal of Immunology 1203

(Novex). Membranes were blocked with 5% nonfat milk and incubated with rabbit anti-p47phox Ab (1/2000 dilution) or or mouse anti-gp91phox (1/2000 dilution) for 1 h at 25°C. HRP-linked anti-rabbit or anti-mouse IgG (1/3000 dilution) for1hat25°C, ECLϩPlus reagents (Amersham Biosciences) were used as a detection system. For detection of the phos- phorylation of MAPK, membranes were blocked with 5% nonfat milk and washed three times followed by incubation with either anti-phospho- ERK1/2 Ab, anti-ERK1/2 Ab, anti-phospho-p38 Ab, anti-p38 Ab, anti- phospho-JNK Ab, or anti-JNK Ab at 1/1000 dilution (Cell Signaling Tech- nology) overnight at 4°C. Anti-rabbit HRP-linked secondary Ab (1/2000 dilution) was incubated for1hat25°C. The same detection system as above was used. Flow cytometry Microglial cells (HAPI) were used for the flow cytometry study. Cells were treated with LPS for 24 h in the presence and absence of morphine pre- treatment for 30 min. The cells were removed from culture and washed twice with cold FACS buffer (saline with 5% normal goat serum). Micro- glia were incubated with FcR block for 30 min at 4°C and then incubated for 45 min at 4°C with FITC-conjugated Abs specific for OX18 (MHC class I), CD40, and PE-conjugated Abs specific for OX6 (MHC class II) and CD80 or the appropriate isotype control Abs (BD Pharmingen). After Downloaded from Ab binding, cells were washed and fixed. Fluorescence was analyzed on a FACSCalibur (BD Biosciences). The mean fluorescent intensity (MFI) of experimental groups was determined by subtracting the MFI of the isotype control from the MFI of the unstimulated or LPS-stimulated microglia. FIGURE 6. Microglial PHOX is the target of morphine inhibition in Neutralization of TNF-␣ or IL-6 LPS-induced neurotoxicity. PHOXϩ/ϩ and PHOXϪ/Ϫ murine neuron-glia Ϫ14 Neuron-glial cultures were used for Ab blocking experiments 7 days after cultures were pretreated with vehicle or morphine (10 M) for 30 min initial seeding. Indicated concentrations of goat anti-rat TNF-␣ and anti-rat followed by LPS treatment. Neurotoxicity was assessed by DA uptake (A). http://www.jimmunol.org/ IL-6 Abs, or the isotype-matched goat IgG control Ab (R&D Systems) TNF-␣ production (B) was measured by ELISA. Results are expressed as were added to the neuron-glia cultures concurrently with 10 ng/ml LPS, percent of the control culture (A), and picograms per milliliter (B), respec- and [3H]DA uptake assays were then performed as described above 8 days tively, and are the means Ϯ SE of three individual experiments in triplicate p Ͻ 0.01 compared with LPS treated ,ءء ;p Ͻ 0.05 ,ء .after addition of Abs and LPS to measure the neurotoxicity. in each experiment Statistical analysis culture.

Data were presented as the means Ϯ SE. For multiple comparisons of ϩ groups, two-way ANOVA was used. Statistical significance between induced by MPP as well as reactive microgliosis induced by toxic groups was assessed by paired Student’s t test, followed by Bonferroni factors released by dying neurons (2). After treatment with 0.2 ␮M

ϩ by guest on September 28, 2021 correction using the JMP program (SAS Institute). A value of p Ͻ 0.05 was MPP for 7 days, DA uptake was reduced to 30% in neuron-glia considered statistically significant. cultures (Fig. 2). However, both l- and d- forms of morphine were capable of significant but not complete protection from MPPϩ- Results mediated neurotoxicity (Fig. 2). Both l- and d-morphine are equipotent in protecting DA neurons ϩ Microglia, but not astroglia, are the target of the neuroprotective against LPS- and MPP -induced neurotoxicity effect of morphine on MPPϩ-induced neurotoxicity. To discern whether both stereoenantiomers of morphine possess The results in Fig. 2 are consistent with the idea that the direct neuroprotective properties, primary mesencephalic rat neuron-glia toxicity of MPPϩ on DA neurons is resistant to the neuroprotective cultures were pretreated with either l-ord-morphine (10Ϫ18–10Ϫ7 effects of morphine, whereas toxicity mediated by reactive micro- M) for 30 min and then stimulated with LPS for 7 days. The gliosis is inhibited by both the l and d forms of morphine with [3H]DA uptake assay showed that LPS treatment reduced the ca- equal potency. To further investigate the cell type that mediates the pacity of the cultures to take up [3H]DA by 47% compared with neuroprotective effect of morphine, we treated cultures of rat mid- control cells treated with vehicle alone. This LPS-induced reduc- brain lacking glial cells (neuron-enriched cultures, N) with MPPϩ tion in DA uptake was significantly inhibited by the addition of and measured DA uptake of neurons 7 days after treatment. DA either the l or the d forms of morphine at concentrations as low as uptake was reduced to ϳ45% in these neuron-enriched cultures 10Ϫ14 M (Fig. 1A). Microscopic analysis of immunostained TH-IR (Fig. 3). However, both l and d forms of morphine did not signif- neurons revealed that LPS treatment not only decreased the num- icantly protect DA neurons from MPPϩ-induced toxicity. Because ber of TH-IR neurons (Fig. 1B) but also caused a loss of neuronal our standard mesencephalic neuron-glia cultures contain ϳ10% processes, including the loss of neurites (Fig. 1C), which was re- microglia and ϳ50% astrocytes in addition to DA and other neu- versed when cells were pretreated with either l- or d-morphine at rons, we reconstituted these neuron-enriched cultures by adding concentrations of 10Ϫ7,10Ϫ10, and 10Ϫ14 M. Taken together, either ϳ10% microglia (Fig. 3, N ϩ 10% MG) or ϳ50% astro- these results indicate that both l- and d-morphines protect DA neu- cytes (Fig. 3, N ϩ 50% AS) back to these cultures (Fig. 3). We rons from LPS-induced neuronal cell death and loss of function found that both l- and d-morphines failed to protect DA neurons with similar efficacies. from MPPϩ-induced toxicity in either neuron-enriched cultures or To investigate whether morphine mediates its protective func- neuron-astroglia cultures. Interestingly, the presence of 50% as- tion directly on DA neurons, rather than inhibiting LPS-induced trocytes in neuron-enriched cultures showed some protective ef- inflammatory response by glial cells, we treated neuron-glial cul- fects from MPPϩ-induced cytotoxicity when compared with cul- tures with MPPϩ, the active metabolite of 1-methyl-4-phenyl- tures containing only neurons or neurons with only microglial 1,2,3,6-tetrahydropyridine which is known to kill DA neurons di- cells, but this protective effect was morphine independent. In con- rectly. Our previous work has shown that neuronal death induced trast, addition of either the l or d isomers of morphine to neuron- by MPPϩ is the result of both direct cytotoxic effects on the neuron enriched cultures reconstituted with 10% of microglia showed 1204 MORPHINE PROTECTS DA NEURON AGAINST OXIDATIVE STRESS Downloaded from http://www.jimmunol.org/

FIGURE 7. Effects of morphine on cytosolic p47phox protein translocation. A, HAPI cells seeded in dish at 5 ϫ 104 cells/well were treated with LPS for 10 min in the absence or presence of morphine pretreatment for 30 min. Cells were fixed with 3.7% paraformaldehyde in PBS for 10 min. After a phox washing with PBS, cells were incubated with rabbit polyclonal Ab against p47 . Cells were then washed and incubated with FITC-conjugated goat by guest on September 28, 2021 anti-rabbit Ab. Focal planes spaced at 0.4-␮m intervals were imaged with a Zeiss 510 laser scanning confocal microscope (ϫ63 PlanApo 1.4 numerical aperture objective) equipped with LSM510 digital imaging software. Three adjacent focal planes were averaged using Metamorph software. The signal of p47phox (FITC-p47phox; green), and the merge view of cell morphology and p47phox (Phase plus FITC-p47phox) are shown. B, HAPI cells were pretreated with vehicle or morphine (10Ϫ14 M) for 30 min, followed by LPS treatment for 10 min. Membrane fractions were isolated to perform Western blot analysis, gp91phox as an internal membrane control. Each experiment has been performed three times.

significant protective effects. These results demonstrate that mi- directly by LPS stimulation or indirectly by MPPϩ-mediated croglia, but not astroglia or neurons, serve as the target of mor- reactive microgliosis. phine-mediated neuroprotection against MPPϩ-induced toxicity, We also investigated the effect of morphine on other inflamma- and that both l- and d-morphine are capable of inhibiting this re- tory markers by measuring mRNA expression of proinflammatory active microgliosis. mediators in microglia-enriched cultures stimulated with LPS. We used LPS to activate microglia because reactive microgliosis in- Effects of morphine on the expression of proinflammatory duced by MPPϩ does not result in measurable NO or proinflam- factors and microglial surface molecules after matory cytokine production in neuron-glial mixed cultures. Both LPS stimulation l- and d-morphine at subpicomolar concentrations significantly in- Overactivation of microglia through the direct effects of LPS on hibited LPS-induced increases in the expression of iNOS, TNF-␣, microglia or by reactive microgliosis via the death of neurons pro- MIP-1␣, IL-6, IL-1␤, and gp91 mRNA as measured by real-time duce an array of proinflammatory mediators, including ROS, PCR (Fig. 5A). Because it is known that LPS up-regulates MHC which is the pivotal product mediating inflammation-related neu- class I, MHC class II, and costimulatory molecules CD40 and rotoxicity (35, 36). To test the effect of l- and d-morphines on the CD80 in microglia (37), we also examined whether morphine can generation of ROS by microglia, cells were pretreated with mor- regulate these surface markers expression by flow cytometry. Fig. phine, then exposed to either LPS or MPPϩ. Both l- and d- 5B shows that 24 h after LPS stimulation, significant increases of morphine significantly reduced LPS-mediated extracellular MHC class I, class II, CD40, and CD80 expression were detected. superoxide production (Fig. 4A) and intracellular ROS concen- Pretreatment with either l- or d-morphine (10Ϫ14 M) significantly trations (Fig. 4B). Likewise, treatment of cultures with mor- inhibited LPS-induced MHC class I expression. In addition, both phine significantly inhibited the MPPϩ-induced superoxide pro- isomers slightly reduced CD40 expression, but the reduction did duction resulting from reactive microgliosis (Fig. 4C). Taken not reach a statistically significant level. On the other hand, mor- together, these results demonstrated that morphine can inhibit phine treatment with either enantiomer did not show significant ROS free radical production by microglial cells induced either effects on LPS-induced MHC class II and CD80 expression. The Journal of Immunology 1205

FIGURE 8. Western blot analysis of MAPK phos- phorylation. Enriched microglia were treated with LPS (10 ng/ml) in the presence or absence of morphine Ϫ14 (10 M) for 5 or 30 min, cells were then harvested, Downloaded from and the levels of phosphorylated MAPK (ERK1/2, p38, and JNK1/2) relative to total MAPK were determined by Western blot using specific Abs against phosphory- lated or total MAPK, respectively. Representative Western blots for ERK1/2 (A), p38, and JNK (B) phos- phorylation are shown from three independent experiments. http://www.jimmunol.org/ by guest on September 28, 2021

PHOX plays an important role in morphine-mediated protection ducing TNF-␣ production in PHOXϪ/Ϫ mice (Fig. 6B). These re- against LPS-induced neuron degeneration sults demonstrate that LPS can induce some degree of Ϫ/Ϫ The results mentioned above indicate that morphine significantly neurotoxicity in PHOX cell cultures, presumably through ␣ reduced both LPS- and MPPϩ-induced production of superoxide. the production of other toxic mediators such as TNF- , IL-6, To further investigate the role of ROS in morphine-elicited neu- and NO. Therefore, our data support the idea that inhibition of ␣ roprotection, we used neuron-glia cells from mice deficient in the ROS production and subsequent TNF- production is associ- catalytic subunit of PHOX, which is the key enzyme required for ated with the neuroprotective effect of morphine at subpicomo- the production of ROS. Neuron-glia cultures were prepared from lar concentrations. PHOXϪ/Ϫ knockout and PHOXϩ/ϩ wild-type mice and, as shown Morphine inhibits LPS-induced translocation of cytosolic in Fig. 6A, LPS treatment of neuron-glia cultures prepared from phox PHOXϩ/ϩ mice substantially reduced [3H]DA uptake. Similarly, subunit of PHOX p47 to the cell membrane both l- and d-morphine at 10Ϫ14 M significantly attenuated the It has been reported that activation of PHOX requires phosphor- decrease in [3H]DA uptake (Fig. 6A). In contrast, although LPS ylation and subsequent translocation of the cytosolic component treatment also showed a significant albeit smaller reduction in p47phox, together with other PHOX cytoplasmic subunits to the 3 Ϫ/Ϫ [ H]DA uptake capacity in PHOX mice, both 1- and d-mor- cell membrane, where they complex with cytochrome b558 to as- phines failed to show any protective effect on DA neurons from semble into an active enzyme (38). Because morphine significantly these mice. We also observed that LPS-induced TNF-␣ production inhibits superoxide production induced by either LPS or MPPϩ, in PHOXϪ/Ϫ mice is significantly less than that seen in PHOXϩ/ϩ we sought to determine whether morphine inhibits PHOX activa- mice (Fig. 6B) and that both enantiomers of morphine were able to tion by preventing the translocation of p47phox from the cytosol to significantly reduce TNF-␣ production in PHOXϩ/ϩ mice. In con- the membrane after LPS stimulation. Using confocal microscopy, trast, neither l- nor d-morphines were capable of significantly re- we observed that LPS initiated the translocation of cytosolic 1206 MORPHINE PROTECTS DA NEURON AGAINST OXIDATIVE STRESS

ylation of p38 and JNK at the time points studied (Fig. 8B). These results suggest that morphine-mediated inhibition of p47phox phos- phorylation and translocation occurs in an ERK (most likely ERK1)-dependent manner in microglia.

TNF-␣ and IL6 partially contributes to LPS-induced neurotoxicity The above-mentioned results indicate that PHOX plays an impor- tant role in LPS-induced neurotoxicity. However, there is a highly significant PHOXindependent neurotoxicity after LPS stimulation in neuron-glia cultures (Fig. 6A). Thus, we further studied the role of other proinflammatory factors in LPS-induced neuronal death. Various concentrations of anti-TNF-␣ or IL6 Abs were added to the neuron-glia cultures to neutralize these two proinflammatory cytokines, respectively. Our results (Fig. 9) indicated that both anti-TNF-␣ Ab (10 ␮g/ml) and anti-IL6 Ab (3 and 10 ␮g/ml) significantly reduced LPS-induced neurotoxicity. Therefore, TNF-␣ and IL-6 partially contribute to LPS-induced neuronal Downloaded from damage.

FIGURE 9. TNF-␣ and IL6 partially contribute to LPS-induced neuro- toxicity. Various concentrations of anti-TNF-␣ (A) anti-IL6 (B) Abs (1–10 Discussion ␮g/ml), or the isotype control Ab were added to 7-day neuron-glia cultures This study demonstrates that both isomers of morphine, at subpi- concurrently with 10 ng/ml LPS. Eight days after LPS stimulation, [3H]DA comolar concentrations, exhibit potent anti-inflammatory and neu- uptake assays were performed to measure neurotoxicity. Results were ex- roprotective effects and operate through a novel mechanism that is http://www.jimmunol.org/ pressed as a percent of the vehicle-treated control cultures and were the independent of conventional opioid receptors. Using the well-es- p Ͻ 0.05 tablished LPS- and MPPϩ-mediated PD in vitro models, we found ,ء .means Ϯ SE. from three independent experiments in triplicate compared with the LPS-treated cultures. that: 1) morphine shows strong efficacy as a neuroprotective agent; 2) both enantiomers are equally potent in neuroprotection at fem- tomolar concentrations; 3) morphine is neuroprotective in both the p47phox to the membrane (Fig. 7A, III and IV), and pretreatment of LPS and MPPϩ models of neurotoxicity, and that this protection is cells with either l-morphine (Fig. 7A, V and VI)ord-morphine mediated through the inhibition of microglia overactivation rather Ϫ (Fig. 7A, VII and VIII)at10 14 M significantly attenuated this than via a direct effect on neurons or astrocytes; 4) mechanisti- translocation. In cells treated with vehicle alone (Fig. 7A, I and II) cally, one novel and important finding from this paper is that by guest on September 28, 2021 Ϫ or 10 14 M morphine (data not shown) in the absence of LPS PHOX is the site of action mediating the anti-inflammatory effect stimulation, p47phox remained localized primarily in the cytosol. of morphine; 5) finally, we have elucidated the signaling pathway Consistent with the result of the confocal microscopy study, West- by which morphine affects PHOX activation by showing that mor- ern blot assay clearly showed an increase in the immunoreactivity phine reduced the LPS-induced PHOX cytosolic subunit p47phox of p47phox in the membrane fraction of HAPI cells 10 min after translocation to the cell membrane by inhibiting ERK phosphor- LPS treatment, and this increase in the immunoreactivity was sig- ylation. Our studies strongly suggest that anti-inflammatory prop- nificantly blocked in the presence of morphine (Fig. 7B). There- erty underlies the neuroprotective effect of morphine. This notion fore, one mechanism by which morphine inhibits superoxide pro- was supported by the finding that both l- and d-morphine showed duction in microglial cells is through the inhibition of p47phox potent anti-inflammatory effects in the subpicomolar doses, which translocation to the cell membrane after LPS stimulation. is much lower than the nanomolar concentrations required for the activation of mu-opioid receptor (42). Morphine suppresses LPS-induced ERK1/2 phosphorylation Although the analgesic effect of l-morphine is believed to be It has been shown that translocation of p47phox to the cell mem- mediated through the mu-opioid receptor, it is not clear if the anti- brane in activated neutrophils requires activation of the MAPK inflammatory functions of morphine are mu-opioid receptor de- pathway, resulting in the phosphorylation and subsequent translo- pendent. Our previous studies have shown that other cation of p47phox to the membrane (39–41). Thus, we investigated compounds, such as (43), 3-hydroxymethor- whether the inhibitory effect of morphine on the activation of phan (44), as well as morphine antagonists naloxone and naltrex- PHOX enzyme by LPS is mediated through the ERK1/2, p38, or one, are also highly neuroprotective both in vivo and in vitro (20, JNK pathways. Microglia cells were pretreated with subpicomolar 21). The results from this study strongly suggest that the anti- concentration of morphine followed by stimulation with LPS for inflammatory effects of morphine are mu-opioid receptor indepen- 5–30 min. ERK1/2, p38, and JNK were all phosphorylated in mi- dent given that d-morphine, which does not bind to the mu-opioid croglia afater LPS stimulation, and the phosphorylation of p44 receptor, was equally and in some cases more effective than l- ERK (ERK1) but not p42 ERK (ERK2) was significantly reduced morphine in protecting DA neurons against LPS or MPPϩ-induced by pretreatment with 10Ϫ14 M of either l-ord-morphine (Fig. 8A) neurotoxicity. More recently, Tseng et al. (19) showed that both l- at 5 min after LPS treatment. Pretreatment with either l-ord- and d-morphines also possess potent antianalgesic effects on morphine significantly reduced both ERK1 and ERK2 phosphor- mouse spinal cord responses at extremely low concentrations. Al- ylation 30 min after LPS treatment (Fig. 8A). Interestingly, though the nature and activity of this novel receptor have yet to be d-morphine showed a stronger inhibitory effect on ERK1/2 phor- characterized, results from our studies and those of Tseng et al. phorylation compared with l-morphine at this time point. In con- suggest that the target of the activity of morphine is mediated trast, neither l- nor d-morphine inhibited LPS-induced phosphor- through the alteration of glial cell activity rather than on neurons. The Journal of Immunology 1207

Interestingly, our results showed that morphine has protective induced neurodegeneration (58). Moreover, we recently found that effects in both the LPS and the MPPϩ model of PD even though a systemic injection of TNF-␣ can cause neuronal death (59) and the target of these two agents is different. LPS leads to the direct the production of superoxide (L. Qin and J.-S. Hong, unpublished activation of microglia, which results in death of DA neurons observations) in the brain. Furthermore, our studies indicated that through the production of inflammatory mediators. Our previous after neutralization of endogenous TNF-␣ or IL-6, LPS-induced study has demonstrated that removal of microglia but not astroglial neurotoxicity was significantly decreased (Fig. 9). These observa- cells from neuron-glia cultures prevents the neurotoxicity of LPS tions are consistent with the notion that LPS-elicited neuronal (28). However, more compelling data for the crucial role of mi- damage is resulted from the combined toxic effects from an array croglia cells in neurotoxicity and the target of morphine action of pro-inflammatory factors, such as free radicals, cytokines, and come from the use of MPPϩ as the neurotoxic agent. Even though prostaglandins released by microglia. Therefore, PHOX plays an MPPϩ directly targets DA neurons for toxicity, morphine also pro- important, but not exclusive role in mediating inflammation-re- vides significant although more modest protective effects to these lated neurotoxicity. cultures. We demonstrate that this effect is due to the fact that a While investigating the mechanism by which morphine inhibits significant amount of the MPPϩ-mediated neurotoxicity requires PHOX activity, we found that both 1- and d-morphines inhibited microglia cells, suggesting that MPPϩ works both to directly kill the phosphorylation of ERK 1/2 as early as 5 min after LPS stim- a subset of DA neurons but also indirectly, by activating microglia, ulation. Several studies have already shown that inhibitors of the which contributes to additional neurotoxicity by producing toxic ERK and p38 pathways inhibit the respiratory burst in fMLP, a inflammatory mediators. Increasing evidence has shown that the chemoattractant released from bacteria, activated neutrophils (39– microglia response to neuronal damage is a long-term self-propel- 41). In our previous study, we show that p47phox translocation to Downloaded from ling process (35, 45, 46), where a small percentage of dying or the cell membrane occurred 10 min after LPS stimulation (28). In damaged neurons release proinflammatory signals that activate mi- this study, we provide evidence that morphine down-regulates croglia through either the release of soluble factors or the loss of ERK phorphorylation within 5 min of LPS exposure, which cor- cell inhibition between neurons and microglia (47–50). The acti- related with the inhibition of superoxide production and p47phox vated microglia then damage the remaining DA neurons, thus cre- translocation. Interestingly, we observed that although morphine

ating a continuous cycle of reactive microgliosis that mimics a can inhibit ERK phosphorylation for at least 30 min after LPS http://www.jimmunol.org/ chronic progressive neurodegeneration such as observed in PD. treatment, morphine has no effect on the phorphorylation of either Reports from our laboratory and others have shown that MPPϩ can p38 or JNK. Moreover, morphine treatment can lower the basal cause reactive microgliosis and that oxidative stress is involved in levels of phospho-p38 and ERK, however, the mechanism for the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine/MPPϩ induced neu- lower expression of this basal level of phosphorylation is not rotoxicity (35, 48, 51). On the basis of our current evidence, we known. Because we repeatedly do not see a significant effect of propose that the anti-inflammatory effect of morphine is capable of morphine on the level of the LPS-induced phospho-p38 signal, inhibiting both LPS-induced microglial activation and MPPϩ-in- these results demonstrate that the major effect of morphine after duced reactive microgliosis, and its ultimate result is to suppress LPS stimulation is to inhibit ERK but not p38 phosphorylation. the inflammation that mediates chronic neurodegeneration in PD. Although it has been previously reported that protein kinase C is by guest on September 28, 2021 We further propose that morphine inhibits microglia-induced involved in the regulation of p47phox phosphorylation and its trans- neurotoxicity by inhibiting the activation of PHOX. Support for location to the cellular membrane in monocytes (60, 61), Neither this conclusion comes from the following sets of observations: 1) l- nor d-morphine exhibited any inhibitory effect on protein kinase our previous studies have shown that inhibiting microglia produc- C phorphorylation (data not shown). Taken together, our results tion of superoxide was critical in eliciting a neuroprotective effect suggest that morphine at 10Ϫ14 M inhibits LPS-induced ROS pro- against microglia-mediated inflammatory damage, indicating that duction in microglia by inhibiting p47phox translocation, which is superoxide is a key element associated with the degeneration of regulated by ERK signaling pathway. DA neurons (24); 2) morphine significantly inhibits superoxide Increasing evidence from both in vivo and in vitro studies sug- production induced by LPS by preventing translocation of the cy- gest that possess important immunomodulatory poten- tosolic subunit p47phox to the cellular membrane (Fig. 7), which is tial (62). Morphinans are a group of compounds structurally sim- required for the activity of PHOX (52); 3) we found that subpi- ilar to morphine, but lacking the E-ring found in the naturally comolar concentrations of both l- and d-morphine could signifi- occurring , as well as the 6-OH and the 7,8-double bond. cantly reduce LPS-induced DA uptake reduction in wild-type mice Our studies and those of others have shown that these compounds but did not provide significant protection to cells derived from work at very low concentrations, to inhibit both cellular activation PHOXϪ/Ϫ mice (Fig. 6A); 4) the production of intracellular ROS and microglial cell-mediated neurotoxicity. More importantly, the and TNF-␣ are reduced in PHOXϪ/Ϫ mice, suggesting that PHOX dextrorotatory form of this compound, which has no affinity or can indirectly regulate intracellular ROS concentration and TNF-␣ effect on the mi-opioid receptors and therefore does not cause ad- production, presumably through the reduction NF-␬B activity (28, diction, is equally if not more effective than its l form in mediating 53). This reduction in NF-␬B activity would likely explain our anti-inflammatory effects on microglia (62). Given the effective- results which found that morphine also reduces the transcription ness of these compounds at such low concentrations and the broad of proinflammatory genes such as iNOS, IL-1␤, IL-6, TNF-␣, spectrum of activity in inhibiting inflammatory responses, these MIP-1␣, gp91phox, and MHC class I, all of which are regulated compounds appear to have significant potential in the treatment of by NF-␬B (54–56). chronic inflammatory disorders. In addition, PHOX is an ideal tar- Although the above-mentioned results show that PHOX plays a get for the development of anti-inflammatory drugs because of its role in LPS-mediated neurotoxicity, there is a highly significant central role in tissue destruction, through the production of ROS, PHOX-independent neurotoxicity induced by LPS. We have pre- as well as its role in amplifying the production of other proinflam- ␣ ␥ ␣ viously reported that TNF- in conjunction with IFN- produced matory factors, such as TNF- , PGE2, and NO (2). In addition, the neurotoxicity in neuron-glia cultures (57) and that inhibition of expression of PHOX is highly cell type specific, with high expres- iNOS activity and NO production prevents LPS-induced neuro- sion in cells of the monocytic lineage. However, since these proin- toxicity, indicating that NO production is also involved in LPS- flammatory neurotoxic factors also exhibit immunoregulatory 1208 MORPHINE PROTECTS DA NEURON AGAINST OXIDATIVE STRESS functions necessary for normal immune responses, the use of these 23. Chen, C. J., F. C. Cheng, S. L. Liao, W. Y. Chen, N. N. Lin, and J. S. Kuo. 2000. compounds to regulate inflammatory stimuli must be carefully Effects of naloxone on lactate, pyruvate metabolism and antioxidant enzyme ac- tivity in rat cerebral ischemia/reperfusion. Neurosci. Lett. 287: 113–116. studied to determine their efficacy for long-term usage. 24. Liu, B., L. Du, and J. S. Hong. 2000. Naloxone protects rat dopaminergic neurons against inflammatory damage through inhibition of microglia activation and su- peroxide generation. J. Pharmacol. Exp. Ther. 293: 607–617. Acknowledgments 25. Jacquet, Y. F., W. A. Klee, K. C. Rice, I. Iijima, and J. Minamikawa. 1977. We thank Dr. Jie Liu at National Institute of Environmental Health Sci- Stereospecific and nonstereospecific effects of (ϩ)- and (Ϫ)-morphine: evidence ences for his help in performing the real time PCR studies and thank Dr. for a new class of receptors? Science 198: 842–845. Brian Mill at the same institution for his help in editing this paper. 26. Cheepsunthorn, P., L. Radov, S. Menzies, J. Reid, and J. R. Connor. 2001. Char- acterization of a novel brain-derived microglial cell line isolated from neonatal rat brain. Glia 35: 53–62. Disclosures 27. Gao, H. M., J. Jiang, B. Wilson, W. Zhang, J. S. Hong, and B. Liu. 2002. Mi- The authors have no financial conflict of interest. croglial activation-mediated delayed and progressive degeneration of rat nigral dopaminergic neurons: relevance to Parkinson’s disease. J. Neurochem. 81: 1285–1297. References 28. Qian, L., M. L. Block, S. J. Wei, C. F. Lin, J. Reece, H. Pang, B. Wilson, 1. Kim, W. G., R. P. Mohney, B. Wilson, G. H. Jeohn, B. Liu, and J. S. Hong. 2000. J. S. Hong, and P. M. Flood. 2006. Interleukin-10 protects lipopolysaccharide- Regional difference in susceptibility to lipopolysaccharide-induced neurotoxicity induced neurotoxicity in primary midbrain cultures by inhibiting the function of in the rat brain: role of microglia. J. Neurosci. 20: 6309–6316. NADPH oxidase. J. Pharmacol. Exp. Ther. 319: 44–52. 2. Block, M. L., and J. S. Hong. 2005. Microglia and inflammation-mediated neu- 29. Liu, B., L. Du, L. Y. Kong, P. M. Hudson, B. C. Wilson, R. C. Chang, H. H. Abel, rodegeneration: multiple triggers with a common mechanism. Prog. Neurobiol. and J. S. Hong. 2000. Reduction by naloxone of lipopolysaccharide-induced neu- 76: 77–98. rotoxicity in mouse cortical neuron-glia co-cultures. Neuroscience 97: 749–756. 3. Rosi, S., V. Ramirez-Amaya, A. Vazdarjanova, P. F. Worley, C. A. Barnes, and 30. Liu, B., and J. S. Hong. 2003. Primary rat mesencephalic neuron-glia, neuron-

G. L. Wenk. 2005. Neuroinflammation alters the hippocampal pattern of behav- enriched, microglia-enriched, and astroglia-enriched cultures. Methods Mol. Med. Downloaded from iorally induced Arc expression. J. Neurosci. 25: 723–731. 79: 387–395. 4. Yu, W., J. X. Hao, X. J. Xu, and Z. Wiesenfeld-Hallin. 1997. Comparison of the 31. Peskin, A. V., and C. C. Winterbourn. 2000. A microtiter plate assay for super- anti-allodynic and antinociceptive effects of systemic; intrathecal and intracere- oxide dismutase using a water-soluble tetrazolium salt (WST-1). Clin. Chim. Acta broventricular morphine in a rat model of central neuropathic pain. Eur. J. Pain 293: 157–166. 1: 17–29. 32. Tan, A. S., and M. V. Berridge. 2000. Superoxide produced by activated neu- 5. Grimm, M. C., A. Ben-Baruch, D. D. Taub, O. M. Howard, J. H. Resau, trophils efficiently reduces the tetrazolium salt, WST-1 to produce a soluble J. M. Wang, H. Ali, R. Richardson, R. Snyderman, and J. J. Oppenheim. 1998. formazan: a simple colorimetric assay for measuring respiratory burst activation transdeactivate chemokine receptors: delta and mu receptor-me- and for screening anti-inflammatory agents. J. Immunol. Methods 238: 59–68. http://www.jimmunol.org/ diated heterologous desensitization. J. Exp. Med. 188: 317–325. 33. Liu, J., H. M. Shen, and C. N. Ong. 2001. Role of intracellular thiol depletion, 6. Chao, C. C., T. W. Molitor, K. Close, S. Hu, and P. K. Peterson. 1993. Morphine mitochondrial dysfunction and reactive oxygen species in Salvia miltiorrhiza- inhibits the release of tumor necrosis factor in human peripheral blood mononu- induced apoptosis in human hepatoma HepG2 cells. Life Sci. 69: 1833–1850. clear cell cultures. Int. J. Immunopharmacol. 15: 447–453. 34. Yuan, Z. M., Y. Huang, S. K. Kraeft, L. B. Chen, S. Kharbanda, and D. Kufe. 7. Lysle, D. T., M. E. Coussons, V. J. Watts, E. H. Bennett, and L. A. Dykstra. 1993. 1996. Interaction of cyclin-dependent kinase 2 and the Lyn tyrosine kinase in Morphine-induced alterations of immune status: dose dependency, compartment cells treated with 1-␤-D-arabinofuranosylcytosine. Oncogene 13: 939–946. specificity and antagonism by naltrexone. J. Pharmacol. Exp. Ther. 265: 35. Gao, H. M., B. Liu, W. Zhang, and J. S. Hong. 2003. Critical role of microglial 1071–1078. NADPH oxidase-derived free radicals in the in vitro MPTP model of Parkinson’s 8. Pellis, N. R., C. Harper, and N. Dafny. 1986. Suppression of the induction of disease. FASEB J. 17: 1954–1956. delayed hypersensitivity in rats by repetitive morphine treatments. Exp. Neurol. 36. Qin, L., Y. Liu, T. Wang, S. J. Wei, M. L. Block, B. Wilson, B. Liu, and 93: 92–97. J. S. Hong. 2004. NADPH oxidase mediates lipopolysaccharide-induced neuro-

9. Weber, R. J., and A. Pert. 1989. The periaqueductal gray matter mediates opiate- toxicity and proinflammatory gene expression in activated microglia. J. Biol. by guest on September 28, 2021 induced immunosuppression. Science 245: 188–190. Chem. 279: 1415–1421. 10. McCarthy, L., M. Wetzel, J. K. Sliker, T. K. Eisenstein, and T. J. Rogers. 2001. 37. Olson, J. K., and S. D. Miller. 2004. Microglia initiate central nervous system Opioids, opioid receptors, and the immune response. Drug Alcohol Depend. 62: innate and adaptive immune responses through multiple TLRs. J. Immunol. 173: 111–123. 3916–3924. 11. Sowa, G., G. Gekker, M. M. Lipovsky, S. Hu, C. C. Chao, T. W. Molitor, and 38. Liva, S. M., M. A. Kahn, J. M. Dopp, and J. de Vellis. 1999. Signal transduction P. K. Peterson. 1997. Inhibition of swine microglial cell phagocytosis of Cryp- pathways induced by GM-CSF in microglia: significance in the control of pro- tococcus neoformans by femtomolar concentrations of morphine. Biochem. Phar- liferation. Glia 26: 344–352. macol. 53: 823–828. 39. Nick, J. A., N. J. Avdi, S. K. Young, C. Knall, P. Gerwins, G. L. Johnson, and 12. Walker, J. S. 2003. Anti-inflammatory effects of opioids. Adv. Exp. Med. Biol. G. S. Worthen. 1997. Common and distinct intracellular signaling pathways in 521: 148–160. human neutrophils utilized by platelet activating factor and FMLP. J. Clin. Invest. 13. Peterson, P. K., G. Gekker, C. Brummitt, P. Pentel, M. Bullock, M. Simpson, 99: 975–986. J. Hitt, and B. Sharp. 1989. Suppression of human peripheral blood mononuclear 40. Avdi, N. J., B. W. Winston, M. Russel, S. K. Young, G. L. Johnson, and cell function by and morphine. J. Infect. Dis. 159: 480–487. G. S. Worthen. 1996. Activation of MEKK by formyl-methionyl-leucyl-phenyl- 14. Peterson, P. K., B. Sharp, G. Gekker, C. Brummitt, and W. F. Keane. 1987. alanine in human neutrophils: mapping pathways for mitogen-activated protein Opioid-mediated suppression of cultured peripheral blood mononuclear cell re- kinase activation. J. Biol. Chem. 271: 33598–33606. spiratory burst activity. J. Immunol. 138: 3907–3912. 41. Downey, G. P., J. R. Butler, H. Tapper, L. Fialkow, A. R. Saltiel, B. B. Rubin, 15. Lipovsky, M. M., G. Gekker, S. Hu, A. I. Hoepelman, and P. K. Peterson. 1998. and S. Grinstein. 1998. Importance of MEK in neutrophil microbicidal respon- Morphine enhances complement receptor-mediated phagocytosis of Cryptococ- siveness. J. Immunol. 160: 434–443. cus neoformans by human microglia. Clin. Immunol. Immunopathol. 87: 163–167. 42. Marki, A., K. Monory, F. Otvos, G. Toth, R. Krassnig, H. Schmidhammer, 16. Simpkins, C. O., N. Ives, E. Tate, and M. Johnson. 1985. Naloxone inhibits J. R. Traynor, B. P. Roques, R. Maldonado, and A. Borsodi. 1999. Mu-opioid superoxide release from human neutrophils. Life Sci. 37: 1381–1386. receptor specific antagonist cyprodime: characterization by in vitro radioligand 35 ␥ 17. Brailoiu, E., J. Hoard, G. C. Brailoiu, M. Chi, R. Godbolde, and N. J. Dun. 2004. and [ S]GTP S binding assays. Eur. J. Pharmacol. 383: 209–214. Ultra low concentrations of morphine increase neurite outgrowth in cultured rat 43. Liu, Y., L. Qin, G. Li, W. Zhang, L. An, B. Liu, and J. S. Hong. 2003. Dextro- spinal cord and cerebral cortical neurons. Neurosci. Lett. 365: 10–13. protects dopaminergic neurons against inflammation-mediated de- 18. Wu, H. E., H. S. Sun, M. Terashivili, E. Schwasinger, I. Sora, F. S. Hall, generation through inhibition of microglial activation. J. Pharmacol. Exp. Ther. G. R. Uhl, and L. F. Tseng. 2006. dextro- and levo-morphine attenuate opioid 305: 212–218. delta and kappa receptor agonist produced analgesia in mu-opioid receptor 44. Zhang, W., L. Qin, T. Wang, S. J. Wei, H. M. Gao, J. Liu, B. Wilson, B. Liu, knockout mice. Eur. J. Pharmacol. 531: 103–107. H. C. Kim, and J. S. Hong. 2005. 3-Hydroxymorphinan is neurotrophic to do- 19. Wu, H. E., J. Thompson, H. S. Sun, M. Terashvili, and L. F. Tseng. 2005. paminergic neurons and is also neuroprotective against LPS-induced neurotox- Antianalgesia: stereoselective action of dextro-morphine over levo-morphine on icity. FASEB J. 19: 395–397. glia in the mouse spinal cord. J. Pharmacol. Exp. Ther. 314: 1101–1108. 45. Huh, Y., J. W. Jung, C. Park, J. R. Ryu, C. Y. Shin, W. K. Kim, and J. H. Ryu. 20. Liao, S. L., W. Y. Chen, S. L. Raung, and C. J. Chen. 2003. Neuroprotection of 2003. Microglial activation and tyrosine hydroxylase immunoreactivity in the naloxone against ischemic injury in rats: role of mu receptor antagonism. Neu- substantia nigral region following transient focal ischemia in rats. Neurosci. Lett. rosci. Lett. 345: 169–172. 349: 63–67. 21. Liu, B., J. W. Jiang, B. C. Wilson, L. Du, S. N. Yang, J. Y. Wang, G. C. Wu, 46. McGeer, P. L., C. Schwab, A. Parent, and D. Doudet. 2003. Presence of reactive X. D. Cao, and J. S. Hong. 2000. Systemic infusion of naloxone reduces degen- microglia in monkey substantia nigra years after 1-methyl-4-phenyl-1,2,3,6-tet- eration of rat substantia nigral dopaminergic neurons induced by intranigral in- rahydropyridine administration. Ann. Neurol. 54: 599–604. jection of lipopolysaccharide. J. Pharmacol. Exp. Ther. 295: 125–132. 47. Feng, Z. H., T. G. Wang, D. D. Li, P. Fung, B. C. Wilson, B. Liu, S. F. Ali, 22. Kunihara, T., K. Matsuzaki, N. Shiiya, Y. Saijo, and K. Yasuda. 2004. Naloxone R. Langenbach, and J. S. Hong. 2002. Cyclooxygenase-2-deficient mice are re- lowers cerebrospinal fluid levels of excitatory amino acids after thoracoabdomi- sistant to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced damage of do- nal aortic surgery. J. Vasc. Surg. 40: 681–690. paminergic neurons in the substantia nigra. Neurosci. Lett. 329: 354–358. The Journal of Immunology 1209

48. Wu, D. C., P. Teismann, K. Tieu, M. Vila, V. Jackson-Lewis, H. Ischiropoulos, 56. Molinero, L. L., M. B. Fuertes, M. V. Girart, L. Fainboim, G. A. Rabinovich, and S. Przedborski. 2003. NADPH oxidase mediates oxidative stress in the M. A. Costas, and N. W. Zwirner. 2004. NF-␬B regulates expression of the MHC 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine model of Parkinson’s disease. class I-related chain A gene in activated T lymphocytes. J. Immunol. 173: Proc. Natl. Acad. Sci. USA 100: 6145–6150. 5583–5590. 49. Teismann, P., M. Vila, D. K. Choi, K. Tieu, D. C. Wu, V. Jackson-Lewis, and 57. Jeohn, G. H., L. Y. Kong, B. Wilson, P. Hudson, and J. S. Hong. 1998. Syner- S. Przedborski. 2003. COX2 and neurodegeneration in Parkinson’s disease. Ann. gistic neurotoxic effects of combined treatments with cytokines in murine pri- NY Acad. Sci. 991: 272–277. mary mixed neuron/glia cultures. J. Neuroimmunol. 85: 1–10. 50. Sriram, K., J. M. Matheson, S. A. Benkovic, D. B. Miller, M. I. Luster, and J. P. 58. Jeohn, G. H., C. L. Cooper, B. Wilson, R. C. Chang, K. J. Jang, H. C. Kim, O’Callaghan. 2002. Mice deficient in TNF receptors are protected against dopa- B. Liu, and J. S. Hong. 2002. p38 MAP kinase is involved in lipopolysaccharide- minergic neurotoxicity: implications for Parkinson’s disease. FASEB J. 16: induced dopaminergic neuronal cell death in rat mesencephalic neuron-glia cul- 1474–1476. tures. Ann. NY Acad. Sci. 962: 332–346. 51. Jackson-Lewis, V., and R. J. Smeyne. 2005. MPTP and SNpc DA neuronal vul- 59. Qin, L., X. Wu, M. L. Block, Y. Liu, G. R. Breese, J. S. Hong, D. J. Knapp, and nerability: role of dopamine, superoxide and nitric oxide in neurotoxicity: mini- F. T. Crews. 2007. Systemic LPS causes chronic neuroinflammation and pro- review. Neurotox. Res. 7: 193–202. gressive neurodegeneration. Glia 55: 453–462. 52. Winkelstein, B. A., M. D. Rutkowski, S. M. Sweitzer, J. L. Pahl, and J. A. DeLeo. 2001. Nerve injury proximal or distal to the DRG induces similar spinal glial 60. Dang, P. M., A. Fontayne, J. Hakim, J. El Benna, and A. Perianin. 2001. Protein kinase C␨ phosphorylates a subset of selective sites of the NADPH oxidase com- activation and selective cytokine expression but differential behavioral responses phox to pharmacologic treatment. J. Comp. Neurol. 439: 127–139. ponent p47 and participates in formyl peptide-mediated neutrophil respiratory 53. Pawate, S., Q. Shen, F. Fan, and N. R. Bhat. 2004. Redox regulation of glial burst. J. Immunol. 166: 1206–1213. inflammatory response to lipopolysaccharide and interferon ␥. J. Neurosci. Res. 61. Bey, E. A., B. Xu, A. Bhattacharjee, C. M. Oldfield, X. Zhao, Q. Li, 77: 540–551. V. Subbulakshmi, G. M. Feldman, F. B. Wientjes, and M. K. Cathcart. 2004. 54. Anrather, J., G. Racchumi, and C. Iadecola. 2006. NF-␬B regulates phagocytic Protein kinase C␦ is required for p47phox phosphorylation and translocation in NADPH oxidase by inducing the expression of gp91phox. J. Biol. Chem. 281: activated human monocytes. J. Immunol. 173: 5730–5738. 5657–5667. 62. Zhang, W., J. S. Hong, H. C. Kim, and M. L. Block. 2004. Morphinan neuro- 55. Baldwin, A. S., Jr. 2001. Series introduction: the transcription factor NF-kappaB protection: new insight into the therapy of neurodegeneration. Crit. Rev. Neuro- and human disease. J. Clin. Invest. 107: 3–6. biol. 16: 271–302. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021