Features of bilirubin-induced reactive microglia: From phagocytosis to inflammation

Neurobiology of Disease 40 (2010) 663–675

Contents lists available at ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r.com/ locate /ynbd i

Features of bilirubin-induced reactive microglia: From phagocytosis to inflammation

Sandra L. Silva, Ana R. Vaz, Andreia Barateiro, Ana S. Falcão, Adelaide Fernandes, Maria A. Brito,Rui F.M. Silva, Dora Brites ⁎Research Institute for Medicines and Pharmaceutical Sciences (iMed.UL), Faculty of Pharmacy, University of Lisbon, Avenida Professor Gama Pinto, 1649-003 Lisbon, Portugal

Abbreviations: BSA, Bovine serum albumin; CNS, CeCyclooxygenase-2; DMEM-Ham's F-12, Dulbecco's moF12 medium; DTT, Dithiothreitol; EDTA, EthylenediaEnzyme-linked immune sorbent assay; ERK1/2, Extrace2; FBS, Fetal bovine serum; FITC, Fluorescein isothiocyanprotein; HIV, Human immunodeficiency virus; HSA, Humacalcium-binding adaptor molecule 1; IgG, ImmunoglobuliInterleukin-6; iNOS, Inductiblenitric oxide synthase; JNK1LPS, Lipopolysaccharide; MAPK, Mitogen-activated pmetalloproteinase; NEA, Non-essential aminoacidsNMDA, N-methyl-D-aspartic acid; NO, Nitric oxide; PBP-ERK1/2, Phosphorylated ERK1/2; PGE2, ProstaglandJNK1/2, Phosphorylated JNK1/2; pNA, P-nitroaniline;SDS-PAGE, Sodium dodecyl sulphate-polyacrylamideTumour necrosis factor-α; TREM2, Triggering receptorTRITC, Tetramethylrhodamine isothiocyanate; UCB, Un⁎ Corresponding author. Fax: +351 217946491.

E-mail address: [email protected] (D. Brites).Available online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.nbd.2010.08.010

a b s t r a c t
a r t i c l e i n f o
Article history:Received 11 May 2010Revised 26 July 2010Accepted 11 August 2010Available online 19 August 2010

Keywords:Microglial activationPhagocytic activityInflammatory signalling pathwaysMitogen activated protein kinasesNuclear factor-κBHyperbilirubinemiaCyclooxygenase-2Matrix metalloproteinases

Microglia constitute the brain's immunocompetent cells and are intricately implicated in numerousinflammatory processes included in neonatal brain injury. In addition, clearance of tissue debris by microgliais essential for tissue homeostasis and may have a neuroprotective outcome. Since unconjugated bilirubin(UCB) has been proven to induce astroglial immunological activation and neuronal cell death, we addressedthe question of whether microglia acquires a reactive phenotype when challenged by UCB and intended tocharacterize this response.In the present study we report that microglia primary cultures stimulated by UCB react by the acquisition of aphagocytic phenotype that shifted into an inflammatory response characterized by the secretion of the pro-inflammatory cytokines tumour necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6, upregulation ofcyclooxygenase (COX)-2 and increased matrix metalloproteinase (MMP)-2 and -9 activities. Furtherinvestigation upon upstream signalling pathways revealed that UCB led to the activation of mitogen-activatedprotein kinases (MAPKs) and nuclear factor (NF)-κB at an early time point, suggesting that these pathwaysmightunderlie both the phagocytic and the inflammatory phenotypes engaged by microglia.Curiously, the phagocytic and inflammatory phenotypes inUCB-activatedmicroglia seemto alternate along time,indicating that microglia reacts towards UCB insult firstly with a phagocytic response, in an attempt to constrainthe lesion extent and comprising aneuroprotectivemeasure. UponprolongedUCBexposureperiods, either a shifton global microglia reaction occurred or there could be two distinct sub-populations of microglial cells, onedirected at eliminating the damaged cells by phagocytosis, and another that engaged a more delayedinflammatory response.In conclusion, microglial cells are relevant partners to consider during bilirubin encephalopathy and themodulation of its activation might be a promising therapeutic target.

ntral nervous system; COX-2,dified Eagle's medium-Ham'smine tetraacetic acid; ELISA,llular signal regulated kinase 1/ate; GFAP, Glial fibrillary acidicn serum albumin; Iba1, Ionizedn G; IL-1β, Interleukin-1β; IL-6,/2, c-JunN-terminal kinase 1/2;rotein kinase; MMP, Matrix; NF-κB, Nuclear factor-κB;S, Phosphate-buffered saline;in E2; PI, Propidium iodide; P-P-p38, Phosphorylated p38;gel electrophoresis; TNF-α,

expressed on myeloid cells-2;conjugated bilirubin.

irect.com).

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

Hyperbilirubinemia is a common condition in the neonatal periodand results from a limited ability of the newborns to excrete an overproduced bilirubin (Dennery et al., 2001; Watson, 2009). Physiologicalto pathological transition is driven by the multifocal deposition ofunconjugated bilirubin (UCB) in selected regions of the brain leading toencephalopathy and kernicterus (Hansen, 2002; Porter and Dennis,2002). This event is directly correlated with death, as well as withimpairments of neural development and hearing (Oh et al., 2003).Additionally, moderate hyperbilirubinemia has been proven to beassociated with a significant increase in minor neurologic dysfunctionthroughout the first year of life (Soorani-Lunsing et al., 2001) and hasalso been related to the outcome of mental disorders such asschizophrenia (Miyaoka et al., 2000).

The cytotoxic effects of UCB in the central nervous system (CNS) havebeenbroadly studiedandcomprise several features suchas: perturbationof nerve cell and mitochondria membranes (Rodrigues et al., 2002b,c);inhibition of glutamate uptake prolonging its presence in the synapticcleft (Silva et al., 1999, 2002); N-methyl-D-aspartic acid (NMDA)-

http://dx.doi.org/10.1016/j.nbd.2010.08.010

mailto:[email protected]

http://dx.doi.org/10.1016/j.nbd.2010.08.010

http://www.sciencedirect.com/science/journal/09699961

664 S.L. Silva et al. / Neurobiology of Disease 40 (2010) 663–675

mediated excitotoxicity (Brito et al., 2010; Grojean et al., 2000, 2001;McDonald et al., 1998); and increase in intracellular calcium (Brito et al.,2004). All these eventsmay culminate in cell death by both necrosis andapoptosis (Rodrigues et al., 2002a; Silva et al., 2001) being neuronsmoresusceptible to death mechanisms than astrocytes (Falcão et al., 2006;Silva et al., 2002). Some of the injurious effects of UCB on astrocytes arean elevated glutamate secretion and the activation of inflammatorypathways that lead to cytokine release (Falcão et al., 2006; Fernandes etal., 2006, 2004). Furthermore, our groupwas thefirst todemonstrate thatUCB activates and damages microglial cells (Gordo et al., 2006). Indeedmicroglia showed to be the most reactive brain cells when compared toastrocytes and neurons, as they evidence increased UCB-induced celldeath, release of glutamate and cytokine production (Brites et al., 2009).

Microglial cells residewithin the CNSparenchyma (Streit, 2002) andengage several important roles in the developing brain (Cuadros andNavascues, 1998; Kim and de Vellis, 2005) as well as in pathologicalconditions (Block and Hong, 2005; Nakajima and Kohsaka, 2004). Inresponse to injury, microglia turn into an activated state and display acomplexity of phenotypic alterations that illustrate what is called“reactive microglia.” This activation entails several features such as:dramatic morphologic changes by the acquisition of an amoeboidphenotype (Kreutzberg, 1996); upregulation of intracellular enzymesand cell surface markers, release of pro-inflammatory mediators,oxygen radicals and proteases (Kim and de Vellis, 2005), antigenpresentation (Aloisi, 2001) and phagocytosis (Chew et al., 2006).

Moreover, the implication of microglia to neonatal pathologicconditions has been acknowledged (McRae et al., 1995; Vexler andYenari, 2009), since the production of inflammatorymediators by thesecells is a major contributor to hypoxic–ischemic injury in the neonatalbrain (Doverhag et al., 2010).

Microglial activation must not be viewed as an “on/off” process, butrather as a shift between activity states, altering between a surveyingand an effector status. In fact, microglia's activation process is anadaptive one but, depending on the circumstances in which it occurs,may have neuroprotective or neurotoxic outcomes (Hanisch andKettenmann, 2007). Indeed, microglial involvement in various neuro-degenerative disorders is notorious, namely in Parkinson's disease(Tansey et al., 2008), Alzheimer's disease (Kim and Joh, 2006), multiplesclerosis (Jack et al., 2005; Muzio et al., 2007), and human immunode-ficiency virus (HIV)-associated dementia (Gonzalez-Scarano andBaltuch, 1999), mostly owing to its inflammatory character. Yet,microglial phagocytic role in numerous neurodegenerative diseases,aswell as in acute brain injury, is essential for tissue debris removal andcontributes for a pro-regenerative environment (Neumann et al., 2009).Moreover, phagocytic clearance of debris may be considered aprotective measure as it constitutes an attempt to restrain furtherdetrimental inflammatory responses (Napoli and Neumann, 2009).

In this study we characterize the microglial response to UCBstimulation by evaluating both its phagocytic properties and theinflammatory mechanisms engaged upon activation. We observed, forthe first time, that UCB induces an increase in the phagocytic propertiesof microglia, followed by a shift into a rather inflammatory responsewith prolonged exposure time. Moreover, this inflammatory responsetriggered by UCB follows different temporal profiles of interleukin(IL)-1β, tumour necrosis factor (TNF)-α and IL-6 secretion. Remarkably,an increase in TNF-α and IL-1β release is observed prior to the secretionof IL-6. Our findings also point to mitogen-activated protein kinases(MAPKs) and nuclear factor (NF)-κB as probable signalling pathwaysinvolved in microglial reactivity to UCB as theymight be entailed eitherin inflammation (Hanisch et al., 2001) or phagocytosis (Sun et al., 2008;Tanaka et al., 2009). In fact, we demonstrate that MAPKs phosphory-lation is an essential step for NF-κB nuclear translocation. Furthermore,our results reveal the induction of cyclooxygenase (COX)-2 expressionand matrix metalloproteinase (MMP)-2 and -9 activity in a later phaseof microglial response to UCB, implicating these events in the overalldeleterious and inflammatory response.Weprovide evidence thatUCB-

induced cytokine secretion may participate in MMP activation andhypothesize that these events might be reciprocally regulated, furthercontributing to the complex network of microglia activation process.Taken together, these results strongly imply a multiple response ofmicroglia to UCB, suggesting that those cells are relevant partners toconsider during bilirubin encephalopathy.

Materials and methods

Chemicals

Dulbecco's modified Eagle's medium-Ham's F12 medium (DMEM-Ham's F-12), Opti-MEMmedium, fetal bovine serum(FBS), L-glutamine,sodium pyruvate and non-essential aminoacids (NEA) were purchasedfrom Biochrom AG (Berlin, Germany). Antibiotic antimycotic solution(20×), human serum albumin (HSA; Fraction V, fatty acid free), bovineserum albumin (BSA), Hoechst 33258 dye, biotinylated tomato lectin(Lycopersicon esculentum), avidin-fluorescein isothiocyanate (FITC),avidin-tetramethylrhodamine isothiocyanate (TRITC), fluorescentlatex beads 1 μm (2.5%), mouse anti-β-actin, FITC-labelled goat anti-rabbit IgG, rabbit anti-glial fibrillary acidic protein (GFAP), TRITC-labelled goat anti-rabbit IgG, Coomassie Brilliant Blue R-250 andpropidium iodide (PI) were from Sigma Chemical Co. (St. Louis, MO).UCBwas alsoobtained fromSigmaandpurifiedaccording to themethodof McDonagh and Assisi, 1972. Trypsin/Ethylenediamine tetraaceticacid (EDTA) solution (0.25% trypsin, 1 mMEDTA inHank's balanced saltsolution) and Alexa Fluor 594 chicken anti-goat IgG were purchasedfrom Invitrogen Corporation (Carlsbad, CA).

FuGENEHDTransfection Reagentwas acquired fromRocheMolecularBiochemicals (Mannheim, Germany); Dual Luciferase reporter assaysystemwas from Promega (Madison, WI, USA); and Caspase-3, -8 and -9substrates, Ac-DEVD-pNA, Ac-IETD-pNA and Ac-LEHD-pNA, respectively,were purchased from Calbiochem (San Diego, CA, USA). Concentratedsolutions (10 mM) of MAPK pathways inhibitors SB203580 (p38 MAPKinhibitor; Calbiochem, San Diego, CA, USA), and U0126 [Extracellularsignal regulated kinase (ERK1/2)-upstream inhibitor; Promega,Madison,WI, USA], were prepared in dimethylsulfoxide.

Recombinant rat IL-1β and DuoSet® ELISA kits were from R&DSystems, Inc. (Minneapolis, MN, USA). Nitrocellulose membrane,Hyperfilm ECL and Horseradish peroxidase-labelled goat anti-mouseIgG were obtained from Amersham Biosciences (Piscataway, NJ, USA).LumiGLO®, Cell lysis buffer®, rabbit anti-phosphorylated-p38 (P-p38)and rabbit anti-phosphorylated-ERK 1/2 (P-ERK1/2) were from CellSignalling (Beverly, MA, USA).

Mouse anti-phosphorylated-c-Jun N-terminal kinase (P-JNK1/2),rabbit anti-p65NF-κB subunit and horseradish peroxidase-labelled goatanti-rabbit IgG were from Santa Cruz Biotechnology (Santa Cruz, CA,USA). Goat anti-ionized calcium-binding adaptor molecule 1(Iba1) wasfrom Abcam (Cambridge, UK).

All other chemicals were of analytical grade and were purchasedfrom Merck (Darmstadt, Germany).

Primary culture of microglia

Animal care followed the recommendations of European Conven-tion for the Protection of Vertebrate Animals Used for Experimentaland other Scientific Purposes (Council Directive 86/609/EEC) andNational Law 1005/92 (rules for protection of experimental animals).All animal procedures were approved by the Institutional Animal Careand Use Committee. Every effort wasmade tominimize the number ofanimals used and their suffering.

Mixed glial cultures were prepared from 1- to 2-day-old Wistar ratsas previously described (McCarthy and de Vellis, 1980), with minormodifications (Gordo et al., 2006). Cells (4×105 cells/cm2) were platedon uncoated 12- or 6-well tissue culture plates (Corning Costar Corp.,Cambridge, MA) in culture medium (DMEM-Ham's F-12 medium

665S.L. Silva et al. / Neurobiology of Disease 40 (2010) 663–675

supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate,nonessential amino acids 1×, 10% FBS, and 1% antibiotic-antimycoticsolution) andmaintained at 37 °C in ahumidified atmosphere of 5%CO2.

Microglia were isolated as previously described by Saura et al., 2003.Briefly, after 21 days in culture, microglia were obtained by mildtrypsinization with a trypsin-EDTA solution diluted 1:3 in DMEM-F12for 20–45 min. The trypsinization resulted in detachment of an upperlayer of cells containing all the astrocytes, whereas the microgliaremained attached to the bottom of the well. The medium containingdetached cells was removed and replaced with initial mixed glial-conditioned medium. Twenty-four hours after trypsinization, theattached cells were subjected to the different treatments. The use of21-days-in-vitro cells intents to achieve themaximal yield and purity ofthe cultures. In fact, astrocyte contamination was less than 2%, asassessed by immunocytochemical staining with a primary antibodyagainst GFAP followed by a species-specific fluorescent-labelledsecondary antibody. Microglia were counterstained with a biotinylatedtomato lectin (Lycopersicon esculentum), using a 1:166 dilution in 1%Triton X-100® in phosphate-buffered saline (PBS) overnight at 4 °Cfollowed by 1 h incubation at room temperature with avidin-TRITC in a1:100 dilution in PBS; the nuclei were immunostained with Hoechst33258 dye. Thus, the high purity level of microglia cultures excludesinterference of contaminating astroglial cells.

Cell treatment

Microglial cells were incubated in the absence (control) or in thepresence of 50 μM UCB plus 100 μM HSA, from 5 min to 48 h, at 37 °C.A UCB stock solution (10 mM) was prepared in 0.1 M NaOH immedi-ately before use and the pH of the incubation medium was restored to7.4 by addition of equal amounts of 0.1 MHCl. All the experiments withUCB were performed under light protection to avoid photodegradation.

To study the role of MAPK pathways in microglial response to UCB,cells were pretreated for 20 min with 10 μM of the MAPK inhibitorsprior to UCB stimulation: SB203580, a selective inhibitor of p38 MAPKand U0126, a selective inhibitor of the MAPK kinases (MEK)1/2,upstream kinases in the ERK1/2 pathway.

The involvement of IL-1β in MMP activation was investigated bytreating microglia with 2 ng/mL recombinant rat IL-1β or vehiclealone, in the presence of 100 μMHSA, for 30 min and 1 h at 37 °C. Theselected concentration of cytokine was based on the maximal levelsobtained in our culture model upon UCB stimulation.

Measurement of cytokine release

Aliquots of the cell culture media were collected at the end of theincubations and, after removal of cellular debris by short centrifuga-tion, placed in a 96-well microplate and assessed in triplicate forTNF-α, IL-1β and IL-6 with specific DuoSet® ELISA Development kitsfrom R&D Systems, according to the manufacturer's instructions.Results were expressed as pg/mL.

Western blot assay

Western blot assay was carried out as usual in our lab (Fernandeset al., 2006). Briefly, total protein was extracted from primary microgliausing Cell lysis buffer®. Protein extracts were separated on a 10% sodiumdodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). Fol-lowing electrophoretic transfer onto a nitrocellulose membrane andblocking with 5% milk solution, the blots were incubated with primaryantibody overnight at 4 °C [anti-P-p38 MAPK (1:1000), anti-P-ERK1/2(1:1000), anti-P-JNK1/2 (1:200), anti-COX-2 (1:1000) or anti-β-actin(1:10000) in 5% (w/v) BSA] and with horseradish peroxidase-labelledsecondary antibody [anti-mouse (1:5000) or anti-rabbit (1:5000)] for1 h at room temperature. Protein bands were detected by LumiGLO®and visualized by autoradiography with Hyperfilm ECL.

Detection of NF-κB activation

To assay the transcriptional activity of NF-κB, reporter gene analysiswas applied. A reporter plasmid under control of NF-κB binding siteswas provided by Dr. Guy Haegeman (Flanders Interuniversity Institutefor Biotechnology and University of Gent, Belgium). NF-κB-dependentreporter plasmids, p(IL6κB)350hu.IL6P-luc+, contain three NF-κBbinding sites in the promoter region, while NF-κB-independentplasmids, p50hu.IL6P-luc+, do not (Vanden Berghe et al., 1998).These reporter genes were introduced into microglial cells usingFuGENE HD Transfection Reagent. After 24 h of transfection, cells weretreated with 50 μM UCB plus 100 μM HSA from 30 min to 4 h, at 37 °C.Luciferase assays were carried out using a Dual Luciferase ReporterAssay System (Promega), according to the instructions in manufac-turer's manual. Firefly and renilla luciferase activities were measuredusing a luminometer (Berthold Technologies, Wildbad, Germany).Firefly luciferase activity value was normalized to renilla luciferaseactivity value from pSV-Sport-Rluc plasmid. Readings of promoteractivities of NF-κB-independent plasmids, p50hu.IL6P-luc+ andp1168hIL6m NF-κB-luc (plasmid presenting a mutation in NF-κBbinding sites), were also performed. Results were presented as foldchange of the relative luciferase activity compared to the respectivecontrol.

For immunofluorescence detection of NF-κB nuclear translocation,cells were fixed for 20 min with freshly prepared 4% (w/v) parafor-maldehyde in PBS and a standard immunocytochemical technique wasperformed using a polyclonal rabbit anti-p65 NF-κB subunit antibody(1:200) as the primary antibody and a FITC-labelled goat anti-rabbitantibody (1:160) as the secondary antibody. To identify the totalnumber of cells, microglial nuclei were stained with Hoechst 33258 dyeas previously described. Fluorescence was visualized using a LeicaDFC490 camera adapted to anAxioskop®microscope (Zeiss). Pairs of U.V. and green-fluorescence images of ten random microscopic fields(original magnification: 400×) were acquired per sample. NF-κB-positive nuclei (identified by localization of the NF-κB p65 subunitstaining exclusively at the nucleus) and total cells were counted (N500cells per sample) to determine the percentage of NF-κB-positive nuclei.Results were expressed as fold change versus respective control.

Morphological analysis

For morphological analysis, cells were fixed as described aboveand a standard indirect immunocytochemical technique was carriedout using a primary antibody raised against Iba-1 (goat, 1:500) and asecondary Alexa Fluor 594 chicken anti-goat antibody (1:200).Fluorescent images were acquired using a Leica DFC490 cameraattached to an Axioskop® microscope (Zeiss).

Assessment of microglial phagocytic properties

After treatmentwith UCB, cells were incubatedwith 0.0025% (w/w)1 μm fluorescent latex beads for 75 min at 37 °C and fixed with freshlyprepared 4% (w/v) paraformaldehyde in PBS. Labelling with tomatolectin was performed followed by avidin-TRITC and the nuclei counter-stained with Hoechst 33258 dye. U.V., green and red-fluorescenceimages of fifteen random microscopic fields (original magnification:630×)were acquired per sample. The number of ingested beads per cellwas counted in approximately 250 cells per sample.

Gelatin zymography

Aliquots of culture supernatant were analyzed by SDS-PAGEzymography in 0.1% gelatine–10% acrylamide gels under non-reducingconditions. After electrophoresis, gels were washed for 1 h with 2.5%Triton X-100 (in 50 mMTris pH7.4; 5 mMCaCl2; 1 μMZnCl2) to removeSDS and renature the MMP species in the gel. Then the gels were

Fig. 1. UCB induces the release of TNF-α, IL-1β and IL-6 by microglia following differenttemporal profiles. Rat corticalmicroglial cellswere treatedwith50 μMUCB in thepresenceof 100 μMHSA for the indicated time periods. TNF-α, IL-1β and IL-6 concentrations in themedia were determined by ELISA and expressed as mean ± SEM cytokine release fromfour independent experiments performed in triplicate, after deduction of cytokine valuesin control assays. *pb0.05 and **pb0.01 vs. respective control.


incubated in the developing buffer (50 mM Tris pH7.4; 5 mM CaCl2;1 μM ZnCl2) overnight to induce gelatine lysis. For enzyme activityanalysis, the gels were stained with 0.5% Coomassie Brilliant Blue R-250and destained in 30% ethanol/10% acetic acid/H2O. Gelatinase activity,detected as a white band on a blue background, was quantified bycomputerized image analysis and normalizedwith total cellular protein.

Evaluation of microglial cell death

Necrotic-like cell deathwasassessedbymonitoring the cellularuptakeof the fluorescent dye propidium iodide [PI; 3,8-diamino-5-(3-(diethyl-methylamino)propyl)-6-phenyl phenanthridinium diiodide]. PI readilyenters and stains non-viable cells, but cannot cross the membrane ofviable cells. This dye binds to double-stranded DNA and emits redfluorescence (630 nm; absorbance 493 nm). Unpermeabilized adherentcells cultured on coverslips were incubated with a 75 μM PI solution for15 min in the absence of light. Subsequently, cells were fixedwith freshlyprepared 4% (w/v) paraformaldehyde in PBS and the nuclei immunos-tained with Hoechst 33258 dye.

Red-fluorescence and U.V. images of ten random microscopicfields (original magnification: 400×) were acquired per sample andthe percentage of PI positive cells was counted and expressed as foldversus respective control.

Activities of caspase-3, -8 and -9 were measured by a commercialcolorimetric method. Cells were harvested, washed with ice-cold PBSand lysed for 30 min on ice in the lysis buffer [50 mM HEPES (pH 7.4);100 mM NaCl; 0.1% (w/v) CHAPS; 1 mM dithiothreitol (DTT); 0.1 mMEDTA]. The activities of caspase-3, -8 and -9 were determined in celllysates by enzymatic cleavage of chromophore p-nitroanilide (pNA)from the substrate Ac-DEVD-pNA for caspase-3, Ac-IETD-pNA forcaspase-8 and Ac-LEHD-pNA for caspase-9, according tomanufacturer'sinstructions. The proteolytic reaction was carried out in protease assaybuffer [50 mM HEPES (pH 7.4); 100 mM NaCl; 0.1% (w/v) CHAPS;10 mM DTT; 0.1 mM EDTA; 10% (v/v) glicerol], containing 2 mMspecific substrate. Following incubation of the reactionmixtures for 1 to2 h at 37 °C, the formation of pNAwas measured in a microplate reader(PR2100, BioRadLaboratories, Inc.) atλ=405 nmwitha referencefilterat 620 nm. Readings were normalized to total protein contentdetermined using a protein assay kit (Bio-Rad, Hercules, CA, USA)according to the manufacturer's specification, and expressed as foldchange of respective control.

Statistical analysis

Results of, at least, three different experiments were expressed asmean ± S.E.M. Significant differences between two groups weredetermined by the two-tailed t-test performed on the basis of equaland unequal variance as appropriate. Comparison of more than twogroups was done by ANOVA using Instat 3.05 (GraphPad Software,San Diego, CA, USA) followed by multiple comparisons Bonferronipost-hoc correction. Statistical significance was considered for ap value less than 0.05.

Results

UCB triggers IL-1β, TNF-α and IL-6 secretion following differenttemporal profiles

Supporting evidence reports reactive microglia as one of the mainsources of pro-inflammatory cytokines in the brain (Hanisch, 2002),which are known to exert deleterious effects in nerve cells (Rothwell,1999). Previous results suggested that UCB is able to induce aninflammatory response by microglia (Gordo et al., 2006). Thus, weintended to further characterize those inflammatory events by theevaluation of the temporal secretion profile of IL-1β, TNF-α and IL-6. InFig. 1 it can be observed that UCB stimulates cytokine release in a time-

dependent manner but with different temporal profiles. In fact, TNF-αand IL-1β seem to be the first to be upregulated, as an increase in thosecytokine levels in culture supernatants can be observed as early as 2 hafter the addition of 50 μMUCB. However, while peak values for TNF-αare reached at 4 h of UCB exposure (changing from 365 pg/mL forcontrol conditions to 520 pg/mL for UCB, pb0.05) and decline graduallyalong time of exposure (although an additional increase is noticedat 24 h), themaximum release of IL-1βwas only achieved at 12 h, but ina much higher amount (ranging from 980 pg/mL for control conditionsto 1700 pg/mL for UCB, pb0.05). On the other hand, IL-6 secretion wasonly noticed from 2 h onwards, reaching peak levels at 8 h of UCBexposure (shifting from 1650 pg/mL for control conditions to 2050 pg/mL for UCB, pb0.01) and decreasing thereafter. These results seem toindicate that microglia responds to UCB stimulus with a ratherinflammatory profile which is manifested for prolonged incubationperiods. As earlier results on astrocytes demonstrated that UCB-inducedcytokine secretion involves MAPK and NF-κB activations (Fernandes etal., 2007, 2006) we sought to verify if the same inflammatory signallingpathways are maintained by microglia.

p38 and ERK1/2 phosphorylation is elicited by UCB in microglia at anearly time point

MAPKs have been reported by several studies to be involved in theproduction of inflammatory mediators by microglia (Bhat et al., 1998;Lee et al., 2000; Waetzig et al., 2005), but their involvement on UCBmicroglial stimulation is still not known. So, we assessed thephosphorylated (activated) forms of all three MAPKs (p38, ERK1/2and JNK1/2) in total cell lysates of UCB-exposed microglia, by westernblot, using specific antibodies.

As shown in Fig. 2, upon UCB stimulation, P-p38 and P-ERK1/2expression were significantly upregulated in a rapid but transientmanner. A 1.4-fold induction (pb0.05 vs. control) was observed forP-p38 as early as 15 min after UCB exposure, and this activation wassustained until 30 min of incubation (1.3-fold, pb0.05 vs. control),fading afterwards. A second activation peak was observed after 2 h ofexposure (1.4-fold, pb0.05), declining to control levels at 6 h. In regardto P-ERK1/2, a 1.2-fold increase (pb0.05 vs. control) was observed at15 min of exposure and, as for P-p38 a second peak, although smaller,was noticed at 1 h of exposure (1.1-fold, pb0.05) with values thatdiminish from then on. Aweaker increase in P-JNK1/2was perceived at15 and 30 min of UCB exposure; however, this effect was not


significantly different from the respective controls. Prolonging UCBincubation to longer periods did not modify the pattern of MAPKactivation (data not shown). These results indicate that MAPKsactivation is a rather early event on microglial activation, seeming toinvolve two activation peaks. Next we found interesting to check if thisactivation could be followed by the engagement of NF-κB nucleartranslocation.

NF-κB signalling pathway is triggered in UCB-activated microglia

NF-κB is described as a convergent point of signalling for microglialactivation by cytokines and other stressors (Glezer et al., 2007) and itsimplication in the inflammatory response induced by UCB in astrocyteshas already been established (Fernandes et al., 2006). Hence, weexamined the effect of UCB on NF-κB transactivation in microglial cellsby gene reporter assay (Fig. 3A). The results indicated that UCBmarkedlyinduced NF-κB activation at 15 and 30min of exposure (1.4-fold, pb0.01for both time points). It should be stated that readings of the promoteractivities of p50hu.IL6P-luc+ and p1168hIL6m NF-κB-luc plasmids(empty and mutated vectors, respectively) showed no change in thepresence or absence of UCB (data not shown), thus validating the assay.

To further confirm the activation of this signalling pathway weinvestigated NF-κB activation in microglia exposed to UCB at varioustime points by the immunochemical assessment of the cytoplasmic ornuclear localization of p65 NF-κB subunit. Interestingly we found NF-κBtranslocation into the nucleus to be significantly increased from 15 minto 2 h of exposure when compared to the respective controls (Fig. 3Band C), which is in line with our previous observations regarding NF-κBtranscriptional activation by UCB. Maximum levels of nuclear NF-κBwere observed at 30 min (2.2-fold, pb0.01), while from 4 h onwards

Fig. 2. MAPKs activation is elicited by UCB in microglia. Rat cortical microglia were exposedlysates were analysed by western blotting with antibodies specific for the phosphorylated fone experiment are shown. Similar results were obtained in three independent experimentswith respect to β-actin protein expression and expressed as mean ± SEM fold change com

NF-κB was mostly localized in the cellular cytoplasm. These resultsfollow the ones from MAPK activation, suggesting that, as previouslyverified in astrocytes, both events are connected.

UCB-induced NF-κB translocation depends on both ERK1/2 and p38

To assess whether MAPKs phosphorylation is an essential step forUCB-evoked NF-κB translocation, we investigated this event afterexposure of microglia to UCB alone or in combination with the MAPKinhibitors SB203580 (p38 selective inhibitor) and UO126 [(MEK)1/2selective inhibitor, upstream kinases in the ERK1/2 pathway]. The useof 30 min and 1 h incubation periods was based on previous resultsshowing that maximal translocation of NF-κB to the nucleus inUCB-challenged microglia occurs between 30 min and 4 h.

As expected, pre-incubation with SB203580 and U0126 complete-ly abrogated UCB-stimulated NF-κB nuclear translocation after 30 min(pb0.01) and 1 h (pb0.05) of UCB stimulation (Fig. 4), thus providingproof of concept that p38 and ERK1/2 phosphorylation are requiredfor the engagement of NF-κB pathway upon UCB exposure.

So far our results have described the engagement of earlyinflammatory signalling pathways that culminate in the achievementof an inflammatory phenotype characterized by the release of pro-inflammatory mediators. We next intended to characterize UCB-stimulated microglia at a morphological level, in order to verify theachievement of this activated state.

Microglia depict morphological changes upon UCB stimulation

Modification of microglial morphology is one of the hallmarks ofits activation profile and has been widely used to categorize different

to 50 μM UCB in the presence of 100 μM HSA for the indicated time periods. Total cellorms of the three MAPKs, P-p38, P-ERK1/2 and P-JNK1/2. (A) Representative results of. (B) The intensity of the bands was quantified by scanning densitometry, standardizedpared with control conditions. *pb0.05 vs. respective control.

image of Fig.�2

Fig. 3. UCB activates NF-κB signalling pathway in microglia. Rat cortical microglia were exposed to 50 μM UCB in the presence of 100 μM HSA for the indicated time periods.(A) Microglial cells were transiently transfected with κB-dependent luciferase plasmids and control plasmids. Relative luciferase activities were plotted as fold change of respectivecontrols. To further confirm NF-κB activation, cells were fixed and immunostained with an antibody against p65 NF-κB subunit. (B) Representative results of one experiment areshown. Scale bar, 20 μm. (C) The percentage of NF-κB-positive nuclei was calculated and expressed as fold change versus respective control. Results are mean ± SEM from threeindependent experiments performed in triplicate. *pb0.05 and **pb0.01 vs. respective control.


activation states (Chew et al., 2006; Kim and de Vellis, 2005; Lynch,2009; Raivich et al., 1999). For that reason, we evaluated themorphology and reactivity of UCB-stimulated microglia by immuno-cytochemistry after 4 and 24 h incubation periods. Our resultsindicate that, after a short UCB exposure, microglia shifted from anelongated morphology to a large and amoeboid shape (Fig. 5).This phenotype is characteristic of activated or reactive microglia(Nakajima and Kohsaka, 2004). In contrast, for longer exposureperiods, microglia display fragmented and condensed cytoplasm, afeature described by other authors (Fendrick et al., 2007; Hasegawa-Ishii et al., 2010) as cytorrhexis, indicative of microglia degenerationand senescence.

Interestingly, the inflammatory phenotype previously describedoccurs after prolonged UCB incubation periods. However, activationfeatureswere also observed for shorter stimulations. For thatmatter,wesought to evaluate the phagocytic properties of UCB-challengedmicroglia and, more importantly, to verify whether this reactivephenotype occurred prior, simultaneously or after the inflammatoryresponse triggered by UCB, in order to further characterize thechronologic events of UCB microglial activation.

UCB differently modulates microglial phagocytosis depending onexposure time

Phagocytosis is one of the main features of microglial activationdumping cell debris prior to cell regeneration, and can be involved inthe pathogenesis of several brain dysfunctions (Neumann et al., 2009).However, this microglial property was never investigated under UCBstimulation. As can be seen in Fig. 6, a sharp increase in the uptake offluorescent latex beads by UCB-stimulated microglia occurs from 2 hon, reaching a maximum peak after 4 h of exposure, (~50%, pb0.05).This is a strong evidence that UCB is able to induce microglialphagocytic properties in a rather short-term exposure. Unspecificentry of latex beads due to UCB-induced cell permeabilization wasexcluded by performing the phagocytosis assay in microglial cellsincubatedwith UCB for 4 h and simultaneously determining PI uptake.Dead cells did not appear to engulf particles (data not shown).Interestingly, prolonging UCB incubation until 8 and 12 h slightlyreverts this phagocytic ability although it approached control valuesupon 24 h of exposure. This may indicate that, after an initialphagocytic reaction, microglial cells shifted to the previously observed

image of Fig.�3

Fig. 4. Phosphorylation of p38 and ERK1/2 is essential for UCB-evoked NF-κB activation. Rat cortical microglia were exposed for 30 min and 1 h to 50 μMUCB alone or in combinationwith 10 μM of the p38 inhibitor SB203580 or the ERK1/2 inhibitor U0126. Cells were fixed and immunostained with an antibody against p65 NF-κB subunit. (A) Representativeresults of one experiment are shown. Scale bar, 20 μm. (B) The percentage of NF-κB-positive nuclei was calculated and expressed as fold change vs. respective control. Results aremean ± SEM from three independent experiments performed in triplicate. **pb0.01 vs. respective control, §pb0.05 and §§pb0.01 vs. UCB alone.


inflammatory response to UCB stimulus. To add on the characteriza-tion of microglial behaviour we advanced to the evaluation of othermarkers of its inflammatory response such as (MMPs activity andCOX-2 expression.

Release of active MMPs is enhanced upon UCB stimulation of microglia

MMPs are a family of proteases with many important roles in normaldevelopment although they may also participate in several neuronaldiseases such as multiple sclerosis, ischemia and neuroinflammationgiven their ability to degrade the basal lamina surrounding the blood–

brain barrier allowing infiltration of immune cells, and thus aggravatinginflammatory reactions in the CNS (Michaluk and Kaczmarek, 2007;Rosenberg, 2002). Since microglia have been reported to secrete activeMMPs further contributing to neuronal injury (Kauppinen and Swanson,2005; Woo et al., 2008), and given the fact that cytokines are reported tostimulateMMPs secretion and activity (Gottschall and Yu, 1995; Lin et al.,2009;Vincenti andBrinckerhoff, 2007),we intended to evaluate the levelsof active MMPs secreted by these cells in response to UCB and to verify ifthis activation could be ascribed to UCB-induced IL-1β.

Cell supernatants collected after UCB incubation were subjected togelatin zymography for the assessment of MMP-2 andMMP-9 activity

image of Fig.�4

Fig. 5. UCB-stimulated microglia show reactive morphological changes. Rat corticalmicrogliawere exposed to50 μMUCB in thepresenceof100 μMHSA for the indicated timeperiods. Cells were fixed and immunostained with an antibody against Iba1 reactivitymarker. Representative results of one experiment are shown. Scale bar, 20 μm.


levels. As can be seen in Fig. 7 there is a slight but significant increasein the activity of MMP-2 and MMP-9 (1.1-fold, pb0.05) after aprolonged exposure time (24 h) to UCB. Since this event occurs afterthe onset of cytokine secretion elicited by microglia, a possibleregulation of this event by inflammatory mediators could beoccurring. In fact, stimulation of microglia with 2 ng/mL of IL-1β(which correspond to maximal levels obtained in our culture modelupon UCB stimulation) increases significantly MMP-2 and MMP-9activity at both 12 (1.3-fold and 1.2-fold, pb0.05 and pb0.01,respectively) and 24 h of exposure (1.2-fold and 1.3-fold, pb0.01,respectively).

UCB-stimulated microglia evidence enhanced COX-2 expression

COX-2 is the enzyme responsible for the production of prostanoidssuch as prostaglandin E2 (PGE2), which is known to be involved in theinitiation and propagation of the immune response (de Oliveira et al.,2008). The expression of this enzyme can be induced by lipopolysac-charide (LPS) and cytokines (Levi et al., 1998). It has been proven tobe elevated in Alzheimer's disease (Yokota et al., 2003) and ischemia(Iadecola et al., 1999). As depicted in Fig. 8, a significant upregulationof COX-2 expression was only noticed after 12 and 24 h of UCBexposure (1.1-fold, pb0.01 and 1.2-fold, pb0.05, respectively), whencompared to the respective controls, again suggesting a later responseof microglia to UCB, in line with the previous results.

Finally, we were interested in examining how UCB interactionaffected microglial cell survival.

UCB reduces microglial viability leading to loss of membrane integrityand increased caspase activity

To evaluate the necrotic-like cell death we used the uptake of thefluorescent dye PI as an indicator of membrane integrity and celldamage since this polar substance can only enter dead or dying cells. Toaddress the possible involvement of the apoptotic pathways inmicroglial demise we determined the relative levels of caspase activity,since these proteases have been traditionally viewed as centralregulators of apoptosis (Fink and Cookson, 2005). As depicted inFig. 9, UCB stimulation only arouses increased PI uptake from4 to 12 hofexposure, reaching maximum significance at 8 h. Accordingly, theactivities of the initiator caspase-8 and -9 were found to be significantlyelevated in response to UCB from 2 to 12 h of exposure reachingmaximum the activities 6 h (Fig. 10), while effector caspase-3 was

significantly increased at a relatively later time points (from 4 to 12 h ofexposure). It is interesting to notice that cell death seems to occur on theonset of inflammatory response and when phagocytic activity declines,again suggesting a possible double response from microglia to UCB.

Discussion

In this paperwedescribe, for thefirst time, different activation statesof microglia in the presence of UCB, since these cells display bothphagocytic and inflammatory phenotypes. Indeed, this study is originalin depicting the increased phagocytic properties of microglia upon UCBstimulation. In addition, our group was also the first to implicatemicroglial cells in the inflammatory response elicited by UCB (Gordoet al., 2006). Thus, in this study we further investigated the activationprofile ofmicroglia under UCB stimulation, by the evaluation of some ofits characteristical features, and the signalling events involved in cellresponse. In fact, previous studies performed by our group have proventhat UCB induces immunological responses in astrocytes by theactivation of inflammatory pathways and secretion of glutamate(Fernandes et al., 2006, 2004), and also that UCB is neurotoxic (Falcãoet al., 2007; Silva et al., 2001). Accordingly, it is conceivable thatincreasedmicroglia reactivitymay further contribute to neuronal injuryduring hyperbilirubinemia.

Microglia contribute to both innate and adaptive immune responsesin the brain (Chew et al., 2006). As innate immune cells, they constitutethe first line of defence against invading microorganisms. The hallmarkindicators of such response are the production of pro-inflammatorycytokines, the upregulation of cell surface antigens and phagocytosis(Town et al., 2005). In addition, phagocytosis of debris bymicroglia canbe beneficial in several pathological conditions, such as multiplesclerosis (Takahashi et al., 2007) and Alzheimer's disease (Simard etal., 2006), as it restricts lesion extension and facilitates tissue recovery.

The fact that UCB may alter the function of various cells of theimmune system (both in vivo and in vitro) seems to be firmlyestablished and a wide range of immunosuppressive effects onperipheral immune cells are summarized by Vetvicka et al., 1991, suchas alterations on antigen expression, chemotaxis, bactericidal activity,proliferative response of T lymphocytes, or antibody production. On theother hand, an increase in phagocytosis of both peripheral bloodgranulocytes andmonocytes after UCB treatmentwas reported byMileret al., 1985. Thus, a rather contradictory immunosuppressive–immu-nostimulant status seems tobeobserveduponUCBchallenge thatmightbe explained by dose- or time-dependent effect (Vetvicka et al., 1991).

Our findings demonstrate that, in conditions that intend to mimic amild hyperbilirubinemia, enhancing of microglial phagocytic propertiesby UCB is an early, but transient, event that seems to be lost withincreased time of exposure. Thus, we may assume that phagocytosis isthe first response towards UCB insult and may constitute a neuropro-tective measure.

Various conditions have been shown to greatly modify microglialphagocytic activity, such as cytokines (Koenigsknecht-Talboo andLandreth, 2005) and LPS (Sun et al., 2008), among others. Interestingly,the study performed by von Zahn et al., 1997 reports an induction ofnearly two-fold increase in the uptake of uncoated latex particles byTNF-α-stimulated microglia, substantiating this cytokine as an auto-crine activator of microglial immune functions. Indeed, UCB-activatedmicroglia are reportedly one of the main sources of TNF-α, even whencompared to astrocytes (Brites et al., 2009). Similarly towhatwealreadyobserved for astrocytes (Fernandes et al., 2006), our results point toTNF-α as the first cytokine to be released by microglia upon UCBchallenge and, remarkably, its temporal profile of secretion is ratherparalleled by the observed phagocytic alterations. TNF-α secretionreaches a maximum level at 4 h of UCB exposure, when microglialphagocytic properties are significantly increased and a decline inUCB-induced TNF-α release is observed from this point on, coincidingwith the decreased phagocytosis elicited by UCB.

image of Fig.�5

Fig. 6.Microglial phagocytosis is differently modulated by UCB. Rat cortical microglia were exposed to 50 μM UCB in the presence of 100 μMHSA for the indicated time periods andincubated with 1 μm fluorescent latex beads as described in Materials and methods. (A) Representative results of one experiment are shown. Scale bar, 20 μm. (B) Results areexpressed as number of ingested beads per cell (±SEM) from three independent experiments. *pb0.05 vs. respective control.


Besides its role as microglial phagocytosis inducer, several experi-ments have also implicated TNF-α in demyelination (Akassoglou et al.,1998) and neuronal degeneration (Allan and Rothwell, 2003; Silva et al.,2006). This cytokine, along with IL-1β, participates in astrogliosis(Hanisch, 2002). IL-1β is involved in fever induction and edema,stimulation of COX-2, release of nitric oxide (NO) and free radicals

(Rothwell, 1999), also participating in the recruitment of circulatingleukocytes into the CNS due to its ability to upregulate the expression ofadhesionmolecules and chemokine synthesis (Lee and Benveniste, 1999;Sedgwick et al., 2000). IL-6 can have both pro- and anti-inflammatoryfunctions and is produced in the early phases of CNS insult (Raivich et al.,1999). Our results clearly imply microglia as an important player in the

image of Fig.�6

Fig. 7. MMP-2 and MMP-9 activities are enhanced upon UCB stimulation. Culturesupernatants from rat corticalmicroglial cells were harvested after incubationwith 50 μMUCB in the presence of 100 μMHSA, or with 2 ng/mL IL-1β, for the indicated time periodsand subjected to zymography as described in Materials and methods. (A) Representativegels of one experiment are reported.MMP-2 andMMP-9were identifiedby their apparentmolecular mass of 67 and 92 kDa, respectively. (B) The intensity of the bands wasquantified by scanning densitometry, standardized with respect to total protein contentand expressed as mean ± SEM fold change compared with control conditions. *pb0.05and **pb0.01 vs. respective control.

Fig. 8. COX-2 expression is upregulated by UCB in microglia. Rat cortical microglia wereexposed to 50 μM UCB in the presence of 100 μM HSA for the indicated time periods.Total cell lysates were analysed by western blotting. (A) Representative results of oneexperiment are shown. Similar results were obtained in three independent experi-ments. (B) The intensity of the bands was quantified by scanning densitometry,standardized with respect to β-actin protein and expressed asmean± SEM fold changecompared with control conditions. *pb0.05 vs. respective control.

Fig. 9. UCB induces microglial decreased viability and membrane disruption. Rat corticalmicrogliawere exposed to50 μMUCB in thepresenceof100 μMHSA for the indicated timeperiods and incubated with 75 μM PI as described in Materials and methods. Thepercentage of PI-positive cells was calculated and expressed as fold vs. respective control.Results are mean ± SEM from three independent experiments performed in triplicate*pb0.05 and **pb0.01 vs. respective control.


inflammatory response instigatedbyUCB, since theobservedearly releaseof TNF-α, previously discussed, is followed by a later but intense secretionof IL-6 and an even stronger induction of IL-1β.

Strikingly, UCB seems to induce amajor release of pro-inflammatorycytokines in a time period in which phagocytosis is already absent.Recent reports have, in fact, substantiated the existence of a non-phlogistic (non-inflammatory) phagocytic response from microglia(Neumann et al., 2009), triggered by apoptotic stimuli and potentiallymediated by phosphatidylserine receptors and triggering receptorexpressed on myeloid cells-2 (TREM2) (Hsieh et al., 2009; Takahashiet al., 2005). Additionally, IL-1β and PGE2 were shown to suppressmicroglial ability to phagocytise insoluble fibrillar β-amyloid deposits,suggesting that a pro-inflammatory milieu inhibits this type ofphagocytosis (Koenigsknecht-Talboo and Landreth, 2005).

Our findings also suggest a role for MMPs in UCB-induced microgliareactivity, since their activity is enhanced upon prolonged exposureperiods to this molecule. MMP-9 has been associated with glutamatedysfunction (Michaluk and Kaczmarek, 2007) and its release can be a

cause of microglia-induced neuron death (Kauppinen and Swanson,2005). In addition, inhibition of gelatinases (MMP-2 and -9) showedefficacy in reducing neural injury and dampening neuroinflammation(Leonardo et al., 2008). Thus, these proteases seem to activelyparticipate in inflammatory events and their activity is very tightlyregulated (Sternlicht andWerb, 2001). Cytokines are firmly establishedinducers of MMP expression and secretion (Gottschall and Yu, 1995; Itoet al., 1996), and the induction ofMMPs has been shown to bemediatedbyMAPKs, NF-κB and activator protein-1 signalling pathways (Lin et al.,2009; Shakibaei et al., 2007; Vincenti and Brinckerhoff, 2007;Woo et al.,2008). Interestingly, in our studymodel,MMPs enhanced activity occursat a later time of exposure, whenMAPKs andNF-κB activation aswell ascytokine secretion have already taken place, suggesting that theseevents might be involved in the activation of MMPs induced by UCB.Moreover, active MMPs may also participate in the regulation ofcytokine activity by promoting the secretion and activation of thesemolecules (Chauvet et al., 2001; Kim et al., 2005; Nuttall et al., 2007;

image of Fig.�7

image of Fig.�8

image of Fig.�9

Fig. 10. Microglial apoptotic cell death is elicited by UCB. Rat cortical microglia wereexposed to 50 μMUCB in the presence of 100 μMHSA for the indicated time periods. Theactivities of caspase-3, -8 and -9were determined in cell lysates by enzymatic cleavage ofchromophore pNA from specific substrates. Results are expressed as fold of respectivecontrol at each time point. Data are means ± SEM from at least three independentexperiments. *pb0.05 and **pb0.01 vs. respective control.


Woo et al., 2008) or, on the other hand, by negatively regulating theirbiological activities (Ito et al., 1996). As stated above, MMPs increasedactivity in UCB-stimulated microglia takes place after the peak ofcytokine secretion, suggesting a possible reciprocal regulation betweenpro-inflammatory cytokines andMMPs, since the lattermolecules couldbe involved in the termination of the inflammatory response by meansof the degradation of IL-1β. Our results further address this issue sinceIL-1β demonstrated to intensely elevate MMP-2 and -9 activationproviding proof of concept that IL-1β secretion produced upon UCBstimulation, is at least in part, responsible for MMP activation.Discrepancy between the activation levels observed for UCB or IL-1βalonemay be a result of UCB-multiple cytokine activation which resultsin a pleiotropic regulatory loop, absent in the second condition.

COX-2 expression can also be induced in microglia by severalinflammatory conditions. As can be seen in our results, UCB is able toinduce upregulation of COX-2 inmicroglial cells in a profile very similarto that observed for IL-6 and IL-1β, thus contributing to the overallinflammatory environment described so far and again pointing to aninflammatory response secondary to phagocytosis. Therefore, ourstudies suggest a dual role for microglia upon UCB stimulation, shiftingfrom a phagocytic and possibly neuroprotective phenotype towards aninflammatory and deleterious one. This is consistent with our findingsdemonstrating that microglia portray altered morphological featuresafter a prolonged UCB exposure, typical of an activated state.

As previously observed for astrocytes (Fernandes et al., 2006, 2004)we show here that UCB stimulation of microglial cells also involves theactivation of MAPKs and NF-κB. MAPKs can be activated by a variety ofdifferent stimuli (Roux and Blenis, 2004), and the engagement of thissignalling pathway can lead to the phosphorylation of severalsubstrates, including transcription factors such as NF-κB, which mayultimately lead to the enhanced transcription of genes encoding for pro-inflammatory cytokines (Koj, 1996). Activation of p38 and ERK1/2 areregarded as essential steps for cytokine induction since their involve-ment in TNF-α, IL-1β, IL-6, COX-2 and inductible nitric oxide synthase(iNOS) expression in microglia has been widely established (Bhat et al.,1998; Hanisch et al., 2001; Lee et al., 2000). Intriguingly, MAPKsactivation, particularly p38, seems tobe also involved in the induction ofmicroglia phagocytosis (Sun et al., 2008; Tanaka et al., 2009).

In this regard our data indicate a rapid activation of p38 and ERK1/2by UCB in microglial cells, which occurs prior to the production ofinflammatory mediators previously reported. In fact, MAPKs activationbyUCB inmicroglia is triggeredat amuchearlier stage than inastrocytes(Fernandes et al., 2006), reinforcing the greater responsiveness of theseglial cells during hyperbilirubinemia, but also suggesting that the earlyphagocytic response of microglial cells to UCBmay be under the controlof p38 and ERK1/2 activation. In this case, the latter activation peakobserved would engage the pro-inflammatory cascade that results inIL-1β and IL-6 enhanced secretion, aswell as the inductionof COX-2 andMMPs, a feature already observed for other immune cells (Gong et al.,2008; Hwang et al., 1997).

Moving downstream on the intracellular signalling pathways is theoriginal observation that, as in astrocytes (Fernandes et al., 2007, 2006),NF-κB activation is also present in microglia exposed to UCB.Interestingly, maximum activation of NF-κB takes place during andafter the earlyMAPKsphosphorylation andagainprior to theproductionof IL-1β, TNF-α and IL-6, postulating a possible involvement of NF-κB inboth phagocytic and inflammatory responses elicited by UCB inmicroglia. The observations that NF-κB nuclear translocation in UCB-stimulated microglia is completely abrogated when microglia arepretreated with p38 and ERK1/2 inhibitors provided an unequivocalproof of MAPKs involvement in NF-κB engagement in UCB-challengedmicroglia which had already been previously established by otherauthors in different disease models (Wilms et al., 2003).

It is worthwhile to mention that cell viability and membraneintegrity are compromised from 4 h onwards, indicating increased celldamage induced by UCB on microglial cells from this point onwards. Infact, UCB-induced apoptotic and necrotic microglial cell death havealready been established (Gordo et al., 2006). Our results furtherindicate that both the extrinsic and intrinsic apoptotic pathways aretriggered culminating in the activation of effector caspase-3 andconsequently causing cell death. However, cell death phenomenareach maximum peaks between 6 to 8 h but decrease for longerincubation periods. Together with the findings described above thesedata portray an interesting hypothesis for microglia response to UCBstimulus. So, it is conceivable that, either a shift on global microgliareaction occurs, or there are two distinct sub-populations of microglialcells displaying complementary activation features, one directed ateliminating the damaged cells by phagocytosis, that died afterengulfment of beads, and another engaging a more delayed inflamma-tory response. Actually, fragmentationof cytoplasm(cytorrhexis)whichis suggested in our 24 h morphological observations, has been pointedto be indicative of widespread microglial degeneration in amyotrophiclateral sclerosis models (Fendrick et al., 2007). Degenerative changes inmicroglia such as beading and clusters of fragmented twigs have alsobeen demonstrated in the aged brain (Hasegawa-Ishii et al., 2010).Which of the above mentioned hypotheses is the more valid demandsfurther elucidation and will clarify the multifaceted profile of microgliaactivationunderUCBstimulation. The complexnetworkofUCB-inducedevents in microglia, as well as the proposed interactions between them,is depicted in Fig. 11.

image of Fig.�10

Fig. 11. Schematic representation of time-dependent microglial activation induced by UCB. Upon UCB stimulation of microglial cells, MAPK and NK-κB signalling pathways areengaged, culminating in the generation of a phagocytic response followed by an inflammatory profile. Both phenotypes might alternate due to a reciprocal regulatory effect or to theexistence of two different subpopulations engaging both types of response, being the phagocytic subpopulation firstly extinguished and replaced by a rather inflammatorysubpopulation. This inflammatory profile is characterized by the increased release of pro-inflammatory cytokines TNF-α, IL-1β and IL-6, the upregulation of COX-2 and enhancedactivities of MMP-2 and MMP-9. Regulatory interactions between the UCB-induced events are portrayed in the figure.


In conclusion, our experiments evidence that phagocytosis isdifferently modulated by UCB depending on the time of exposure,prevailing at an early time point, which is followed by the release ofinflammatory cytokines, and activation of MMP-2 and -9, as well as ofCOX-2. Thus, microglial phagocytosis and inflammatory responsestand out as important events prompted by UCB. To what extent theactivation of microglia by UCB has a beneficial or detrimental outcomeis yet to be determined in future studies, where the influence of othernerve cells will be evaluated. Nevertheless, modulation of microglialactivation seems to be a promising target in neonatal bilirubinencephalopathy.

Acknowledgments

The authors thank Elsa Rodrigues for her expertise with genereporter assays. This work was supported by POCI/SAU-MMO/55955/2004 and PTDC/SAU-NEU/64385/2006 grants, from Fundação para aCiência e a Tecnologia (FCT), Lisbon, Portugal and FEDER (to D.B.). S.L.S.was recipient of a PhD fellowship (SFRH/BD/30326/2006) from FCT.

References

Akassoglou, K., et al., 1998. Oligodendrocyte apoptosis and primary demyelinationinduced by local TNF/p55TNF receptor signaling in the central nervous system oftransgenic mice: models for multiple sclerosis with primary oligodendrogliopathy.Am. J. Pathol. 153, 801–813.

Allan, S.M., Rothwell, N.J., 2003. Inflammation in central nervous system injury. Philos.Trans. R. Soc. Lond. B Biol. Sci. 358, 1669–1677.

Aloisi, F., 2001. Immune function of microglia. Glia 36, 165–179.Bhat, N.R., et al., 1998. Extracellular signal-regulated kinase and p38 subgroups of

mitogen-activated protein kinases regulate inducible nitric oxide synthase andtumor necrosis factor-alpha gene expression in endotoxin-stimulated primary glialcultures. J. Neurosci. 18, 1633–1641.

Block, M.L., Hong, J.S., 2005. Microglia and inflammation-mediated neurodegeneration:multiple triggers with a common mechanism. Prog. Neurobiol. 76, 77–98.

Brites, D., et al., 2009. Biological risks for neurological abnormalities associated withhyperbilirubinemia. J. Perinatol. 29 (Suppl 1), S8–S13.

Brito, M.A., et al., 2004. A link between hyperbilirubinemia, oxidative stress and injuryto neocortical synaptosomes. Brain Res. 1026, 33–43.

Brito, M.A., et al., 2010. N-methyl-D-aspartate receptor and neuronal nitric oxidesynthase activation mediate bilirubin-induced neurotoxicity. Mol. Med. Jun 30.[Epub ahead of print].

Chauvet, N., et al., 2001. Rat microglial cells secrete predominantly the precursor ofinterleukin-1beta in response to lipopolysaccharide. Eur. J. Neurosci. 14,609–617.

Chew, L.J., et al., 2006. Microglia and inflammation: impact on developmental braininjuries. Ment. Retard. Dev. Disabil. Res. Rev. 12, 105–112.

Cuadros, M.A., Navascues, J., 1998. The origin and differentiation of microglial cellsduring development. Prog. Neurobiol. 56, 173–189.

de Oliveira, A.C., et al., 2008. Regulation of prostaglandin E2 synthase expression inactivated primary rat microglia: evidence for uncoupled regulation of mPGES-1 andCOX-2. Glia 56, 844–855.

Dennery, P.A., et al., 2001. Neonatal hyperbilirubinemia. N. Engl. J. Med. 344, 581–590.Doverhag, C., et al., 2010. Galectin-3 contributes to neonatal hypoxic–ischemic brain

injury. Neurobiol. Dis. 38, 36–46.Falcão, A.S., et al., 2006. Bilirubin-induced immunostimulant effects and toxicity vary

with neural cell type and maturation state. Acta Neuropathol. 112, 95–105.Falcão, A.S., et al., 2007. Apoptosis and impairment of neurite network by short

exposure of immature rat cortical neurons to unconjugated bilirubin increase withcell differentiation and are additionally enhanced by an inflammatory stimulus.J. Neurosci. Res. 85, 1229–1239.

Fendrick, S.E., et al., 2007. Formation of multinucleated giant cells and microglialdegeneration in rats expressing a mutant Cu/Zn superoxide dismutase gene.J. Neuroinflammation. 4, 9.

Fernandes, A., et al., 2004. Cytokine production, glutamate release and cell death in ratcultured astrocytes treated with unconjugated bilirubin and LPS. J. Neuroimmunol.153, 64–75.

Fernandes, A., et al., 2006. Inflammatory signalling pathways involved in astroglialactivation by unconjugated bilirubin. J. Neurochem. 96, 1667–1679.

Fernandes, A., et al., 2007. MAPKs are key players in mediating cytokine release and celldeath induced by unconjugated bilirubin in cultured rat cortical astrocytes. Eur. J.Neurosci. 25, 1058–1068.

Fink, S.L., Cookson, B.T., 2005. Apoptosis, pyroptosis, and necrosis: mechanisticdescription of dead and dying eukaryotic cells. Infect. Immun. 73, 1907–1916.

Glezer, I., et al., 2007. Neuroprotective role of the innate immune system by microglia.Neuroscience 147, 867–883.

Gong, Y., et al., 2008. Triptolide inhibits COX-2 expression and PGE2 release bysuppressing the activity of NF-kappaB and JNK in LPS-treated microglia. J.Neurochem. 107, 779–788.

Gonzalez-Scarano, F., Baltuch, G., 1999. Microglia as mediators of inflammatory anddegenerative diseases. Annu. Rev. Neurosci. 22, 219–240.

Gordo, A.C., et al., 2006. Unconjugated bilirubin activates and damages microglia. J.Neurosci. Res. 84, 194–201.

Gottschall, P.E., Yu, X., 1995. Cytokines regulate gelatinase A and B (matrixmetalloproteinase2 and 9) activity in cultured rat astrocytes. J. Neurochem. 64, 1513–1520.

Grojean, S., et al., 2000. Bilirubin induces apoptosis via activation of NMDA receptors indeveloping rat brain neurons. Exp. Neurol. 166, 334–341.

Grojean, S., et al., 2001. Bilirubin exerts additional toxic effects in hypoxic culturedneurons from the developing rat brain by the recruitment of glutamateneurotoxicity. Pediatr. Res. 49, 507–513.

Hanisch, U.K., 2002. Microglia as a source and target of cytokines. Glia 40, 140–155.

image of Fig.�11


Hanisch, U.K., Kettenmann, H., 2007. Microglia: active sensor and versatile effector cellsin the normal and pathologic brain. Nat. Neurosci. 10, 1387–1394.

Hanisch, U.K., et al., 2001. The protein tyrosine kinase inhibitor AG126 prevents themassive microglial cytokine induction by pneumococcal cell walls. Eur. J. Immunol.31, 2104–2115.

Hansen, T.W., 2002. Mechanisms of bilirubin toxicity: clinical implications. Clin.Perinatol. 29, 765–778 viii.

Hasegawa-Ishii, S., et al., 2010. Morphological impairments in microglia precede age-related neuronal degeneration in senescence-accelerated mice. Neuropathology,May 19. [Epub ahead of print].

Hsieh, C.L., et al., 2009. A role for TREM2 ligands in the phagocytosis of apoptoticneuronal cells by microglia. J. Neurochem. 109, 1144–1156.

Hwang, D., et al., 1997. Expression of mitogen-inducible cyclooxygenase induced bylipopolysaccharide: mediation through both mitogen-activated protein kinase andNF-kappaB signaling pathways in macrophages. Biochem. Pharmacol. 54, 87–96.

Iadecola, C., et al., 1999. Cyclooxygenase-2 immunoreactivity in the human brainfollowing cerebral ischemia. Acta Neuropathol. 98, 9–14.

Ito, A., et al., 1996. Degradation of interleukin 1beta by matrix metalloproteinases.J. Biol. Chem. 271, 14657–14660.

Jack, C., et al., 2005. Microglia and multiple sclerosis. J. Neurosci. Res. 81, 363–373.Kauppinen, T.M., Swanson, R.A., 2005. Poly(ADP-ribose) polymerase-1 promotes

microglial activation, proliferation, and matrix metalloproteinase-9-mediatedneuron death. J. Immunol. 174, 2288–2296.

Kim, S.U., de Vellis, J., 2005. Microglia in health and disease. J. Neurosci. Res. 81, 302–313.Kim, Y.S., Joh, T.H., 2006. Microglia, major player in the brain inflammation: their roles

in the pathogenesis of Parkinson's disease. Exp. Mol. Med. 38, 333–347.Kim, Y.S., et al., 2005. Matrix metalloproteinase-3: a novel signaling proteinase from

apoptotic neuronal cells that activates microglia. J. Neurosci. 25, 3701–3711.Koenigsknecht-Talboo, J., Landreth, G.E., 2005. Microglial phagocytosis induced by

fibrillar beta-amyloid and IgGs are differentially regulated by proinflammatorycytokines. J. Neurosci. 25, 8240–8249.

Koj, A., 1996. Initiation of acute phase response and synthesis of cytokines. Biochim.Biophys. Acta 1317, 84–94.

Kreutzberg, G.W., 1996. Microglia: a sensor for pathological events in the CNS. TrendsNeurosci. 19, 312–318.

Lee, S.J., Benveniste, E.N., 1999. Adhesion molecule expression and regulation on cells ofthe central nervous system. J. Neuroimmunol. 98, 77–88.

Lee, Y.B., et al., 2000. p38 map kinase regulates TNF-alpha production in humanastrocytes and microglia by multiple mechanisms. Cytokine 12, 874–880.

Leonardo, C.C., et al., 2008. Delayed administration of a matrix metalloproteinaseinhibitor limits progressive brain injury after hypoxia-ischemia in the neonatal rat.J. Neuroinflammation 5, 34.

Levi, G., et al., 1998. Regulation of prostanoid synthesis in microglial cells and effects ofprostaglandin E2 on microglial functions. Biochimie 80, 899–904.

Lin, C.C., et al., 2009. IL-1 beta promotes A549 cell migration via MAPKs/AP-1- and NF-kappaB-dependent matrix metalloproteinase-9 expression. Cell. Signal. 21, 1652–1662.

Lynch, M.A., 2009. The multifaceted profile of activated microglia. Mol. Neurobiol. 40,139–156.

McCarthy, K.D., de Vellis, J., 1980. Preparation of separate astroglial and oligodendroglialcell cultures from rat cerebral tissue. J. Cell Biol. 85, 890–902.

McDonagh, A.F., Assisi, F., 1972. The ready isomerization of bilirubin IX-α in aqueoussolution. Biochem. J. 129, 797–800.

McDonald, J.W., et al., 1998. Role of glutamate receptor-mediated excitotoxicity inbilirubin-induced brain injury in the Gunn rat model. Exp. Neurol. 150, 21–29.

McRae, A., et al., 1995. Microglia activation after neonatal hypoxic–ischemia. Brain Res.Dev. Brain Res. 84, 245–252.

Michaluk, P., Kaczmarek, L., 2007. Matrix metalloproteinase-9 in glutamate-dependentadult brain function and dysfunction. Cell Death Differ. 14, 1255–1258.

Miler, I., et al., 1985. The effect of bilirubin on the phagocytic activity of mouse peripheralgranulocytes and monocytes in vivo. Folia Microbiol. (Praha) 30, 267–271.

Miyaoka, T., et al., 2000. Schizophrenia-associated idiopathic unconjugated hyperbilir-ubinemia (Gilbert's syndrome). J. Clin. Psychiatry 61, 868–871.

Muzio, L., et al., 2007. Multifaceted aspects of inflammation in multiple sclerosis: therole of microglia. J. Neuroimmunol. 191, 39–44.

Nakajima, K., Kohsaka, S., 2004. Microglia: neuroprotective and neurotrophic cells in thecentral nervous system. Curr. Drug Targets Cardiovasc. Haematol. Disord. 4, 65–84.

Napoli, I., Neumann, H., 2009. Microglial clearance function in health and disease.Neuroscience 158, 1030–1038.

Neumann, H., et al., 2009. Debris clearance by microglia: an essential link betweendegeneration and regeneration. Brain 132, 288–295.

Nuttall, R.K., et al., 2007. Metalloproteinases are enriched in microglia compared withleukocytes and they regulate cytokine levels in activated microglia. Glia 55, 516–526.

Oh,W., et al., 2003. Association between peak serum bilirubin and neurodevelopmentaloutcomes in extremely low birth weight infants. Pediatrics 112, 773–779.

Porter, M.L., Dennis, B.L., 2002. Hyperbilirubinemia in the term newborn. Am. Fam.Physician 65, 599–606.

Raivich, G., et al., 1999. Neuroglial activation repertoire in the injured brain: gradedresponse, molecular mechanisms and cues to physiological function. Brain Res.Brain Res. Rev. 30, 77–105.

Rodrigues, C.M.P., et al., 2002a. Bilirubin induces apoptosis via the mitochondrialpathway in developing rat brain neurons. Hepatology 35, 1186–1195.

Rodrigues, C.M.P., et al., 2002b. Bilirubin directly disrupts membrane lipid polarity andfluidity, protein order, and redox status in rat mitochondria. J. Hepatol. 36,335–341.

Rodrigues, C.M.P., et al., 2002c. Perturbation of membrane dynamics in nerve cells as anearly event during bilirubin-induced apoptosis. J. Lipid Res. 43, 885–894.

Rosenberg, G.A., 2002. Matrix metalloproteinases in neuroinflammation. Glia 39,279–291.

Rothwell, N.J., 1999. Annual review prize lecture cytokines—killers in the brain?J. Physiol. 514 (Pt 1), 3–17.

Roux, P.P., Blenis, J., 2004. ERK and p38 MAPK-activated protein kinases: a family ofprotein kinases with diverse biological functions. Microbiol. Mol. Biol. Rev. 68,320–344.

Saura, J., et al., 2003. High-yield isolation of murine microglia by mild trypsinization.Glia 44, 183–189.

Sedgwick, J.D., et al., 2000. Tumor necrosis factor: a master-regulator of leukocytemovement. Immunol. Today 21, 110–113.

Shakibaei, M., et al., 2007. Suppression of NF-kappaB activation by curcumin leads toinhibition of expression of cyclo-oxygenase-2 and matrix metalloproteinase-9 inhuman articular chondrocytes: Implications for the treatment of osteoarthritis.Biochem. Pharmacol. 73, 1434–1445.

Silva, R.F., et al., 1999. Inhibition of glutamate uptake by unconjugated bilirubin incultured cortical rat astrocytes: role of concentration and pH. Biochem. Biophys.Res. Commun. 265, 67–72.

Silva, R.F.M., et al., 2001. Bilirubin-induced apoptosis in cultured rat neural cells isaggravated by chenodeoxycholic acid but prevented by ursodeoxycholic acid.J. Hepatol. 34, 402–408.

Silva, R.F.M., et al., 2002. Rat cultured neuronal and glial cells respond differently totoxicity of unconjugated bilirubin. Pediatr. Res. 51, 535–541.

Silva, R.F.M., et al., 2006. Dissociated primary nerve cell cultures as models forassessment of neurotoxicity. Toxicol. Lett. 163, 1–9.

Simard, A.R., et al., 2006. Bone marrow-derived microglia play a critical role inrestricting senile plaque formation in Alzheimer's disease. Neuron 49, 489–502.

Soorani-Lunsing, I., et al., 2001. Are moderate degrees of hyperbilirubinemia in healthyterm neonates really safe for the brain? Pediatr. Res. 50, 701–705.

Sternlicht, M.D., Werb, Z., 2001. How matrix metalloproteinases regulate cell behavior.Annu. Rev. Cell Dev. Biol. 17, 463–516.

Streit, W.J., 2002. Microglia as neuroprotective, immunocompetent cells of the CNS. Glia40, 133–139.

Sun, H.N., et al., 2008. Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase-dependent activation of phosphoinositide 3-kinase and p38 mitogen-activatedprotein kinase signal pathways is required for lipopolysaccharide-inducedmicroglial phagocytosis. Biol. Pharm. Bull. 31, 1711–1715.

Takahashi, K., et al., 2005. Clearance of apoptotic neurons without inflammation bymicroglial triggering receptor expressed onmyeloid cells-2. J. Exp. Med. 201, 647–657.

Takahashi, K., et al., 2007. TREM2-transduced myeloid precursors mediate nervoustissue debris clearance and facilitate recovery in an animal model of multiplesclerosis. PLoS Med. 4, e124.

Tanaka, T., et al., 2009. Engulfment of axon debris by microglia requires p38 MAPKactivity. J. Biol. Chem. 284, 21626–21636.

Tansey, M.G., et al., 2008. Neuroinflammation in Parkinson's disease: is there sufficientevidence for mechanism-based interventional therapy? Front. Biosci. 13, 709–717.

Town, T., et al., 2005. The microglial "activation" continuum: from innate to adaptiveresponses. J. Neuroinflammation 2, 24.

Vanden Berghe, W., et al., 1998. p38 and extracellular signal-regulated kinase mitogen-activated protein kinase pathways are required for nuclear factor-kappaB p65transactivation mediated by tumor necrosis factor. J. Biol. Chem. 273, 3285–3290.

Vetvicka, V., et al., 1991. The immunosuppressive effects of bilirubin. Folia Microbiol.(Praha) 36, 112–119.

Vexler, Z.S., Yenari, M.A., 2009. Does inflammation after stroke affect the developingbrain differently than adult brain? Dev. Neurosci. 31, 378–393.

Vincenti, M.P., Brinckerhoff, C.E., 2007. Signal transduction and cell-type specificregulation of matrix metalloproteinase gene expression: can MMPs be good foryou? J. Cell. Physiol. 213, 355–364.

von Zahn, J., et al., 1997. Microglial phagocytosis is modulated by pro- and anti-inflammatory cytokines. Neuroreport 8, 3851–3856.

Waetzig, V., et al., 2005. c-Jun N-terminal kinases (JNKs) mediate pro-inflammatoryactions of microglia. Glia 50, 235–246.

Watson, R.L., 2009. Hyperbilirubinemia. Crit. Care Nurs. Clin. N. Am. 21, 97–120 vii.Wilms, H., et al., 2003. Activation of microglia by human neuromelanin is NF-kappaB

dependent and involves p38 mitogen-activated protein kinase: implications forParkinson's disease. FASEB J. 17, 500–502.

Woo, M.S., et al., 2008. Inhibition of MMP-3 or -9 suppresses lipopolysaccharide-induced expression of proinflammatory cytokines and iNOS in microglia. J.Neurochem. 106, 770–780.

Yokota, O., et al., 2003. Cyclooxygenase-2 in the hippocampus is up-regulated inAlzheimer's disease but not in variant Alzheimer's disease with cotton woolplaques in humans. Neurosci. Lett. 343, 175–179.

Documents

Features of bilirubin-induced reactive microglia: From phagocytosis to inflammation