lipopolysaccharide affects golli expression and promotes proliferation of oligodendrocyte...

Lipopolysaccharide Affects GolliExpression and Promotes

Proliferation ofOligodendrocyte Progenitors

RADMILA FILIPOVIC AND NADA ZECEVIC*Department of Neuroscience, University of Connecticut Health Center, Farmington, Connecticut

KEY WORDS activated microglia; remyelination; mouse slice culture; demyelinat-ing diseases

ABSTRACT Proliferation of oligodendrocyte progenitor cells (OPCs) is important forinitial myelination as well as for remyelination in demyelinating diseases. Previously,we showed that numerous OPCs and activated microglia, are present around multiplesclerosis lesions, and that they accumulate Golli proteins. Golli proteins, present in bothneuronal and immune cells, might have a role in the immune processes, as well as indevelopment of neurons and oligodendrocytes. We hypothesize that Golli proteins,generated by microglia in response to inflammation, promote proliferation of OPCs. Totest this hypothesis, we induced inflammation in neonatal mouse brain slice culture withbacterial endotoxin lipopolysaccharide (LPS). Treated slices showed an increase in thenumber of OPCs. Several results support the notion that this effect of LPS is conveyedthrough activation of microglia and upregulation of Golli proteins. First, LPS-treatedbrain slices have increased expression of Golli proteins observed by immunofluorescenceand Western blot analysis. Second, Golli proteins were demonstrated only in the condi-tioned medium from LPS-treated microglial cell cultures (LPS-MCM), and were absentin either the conditioned media from LPS-treated astrocytes or the control media. Third,proliferation of purified OPCs was promoted with LPS-MCM or Golli proteins, but notwith LPS alone. Taken together, these results demonstrate that microglia and/or mi-croglia secreted factors, are necessary for the LPS-promoted proliferation of OPCs andsuggest possible involvement of Golli proteins as one of mediators in this process.© 2004 Wiley-Liss, Inc.

INTRODUCTION

Multiple sclerosis (MS) is an autoimmune andchronic inflammatory disease of the CNS, where oli-godendrocytes, myelin sheaths, and axons are de-stroyed (e.g., Trapp et al., 1998; Noseworthy et al.,2000; Steinman, 2001; Hemmer et al., 2002; Frank-lin, 2002). In addition to regions of myelin loss, calleddemyelinated plaques, pathophysiological studieshave shown “shadow plaques” that represent areas ofpartial remyelination (Raine et al., 1981; Nosewor-thy et al., 2000; Steinman, 2001; Hemmer et al.,2002; Franklin, 2002). Remyelination is never com-plete, however, which prompted efforts to under-stand factors that influence this process. In princi-

ple, remyelination could be expected if residentoligodendrocyte progenitor cells (OPCs) could be in-duced to proliferate and differentiate into myelin-forming mature oligodendrocytes (e.g., Keirsteadand Blakemore, 1999; Franklin, 2002). In our previ-ous study, we demonstrated that OPCs around MS

Grant sponsor: National MS Society; Grant number: RG-3083-B3/1.

*Correspondence to: Nada Zecevic, Department of Neuroscience, University ofConnecticut Health Center, 263 Farmington Ave, Farmington, CT 06030-3401.E-mail: [email protected]

Received 15 April 2004; Accepted 18 August 2004

DOI 10.1002/glia.20125

Published online 15 November 2004 in Wiley InterScience (www.interscience.wiley.com).

GLIA 49:457–466 (2005)

© 2004 Wiley-Liss, Inc.

lesions intermingled with activated microglia, resi-dent immune cells of the brain, and both cell typesshowed elevated levels of Golli proteins (Filipovic etal., 2002).

Gene expressed in oligodendrocyte lineage (Golli)proteins are transcribed from the Golli-MBP (myelinbasic protein) gene (Campagnoni et al., 1993). Ini-tially, Golli proteins were found in oligodendrocytes,but later they were described to be broadly distrib-uted in the central and peripheral nervous systems,as well as in the immune system (Campagnoni et al.,1993; Fritz and Kalvakolanu, 1995; Pribyl et al.,1996a,b; Landry et al., 1998; Givogri et al., 2000; Liuet al., 2001; Filipovic et al., 2002; Tosic et al., 2002).The role of Golli proteins is not well understood, butresults from various sources have described theirinvolvement in differentiation of neurons and oligo-dendrocytes, perhaps by modulating signaling path-ways (Reyes and Campagnoni, 2002; Fernandes etal., 2004). Although not detected in myelin, theiroverexpression in oligodendrocyte cell cultures in-duces formation of myelin sheaths (Campagnoni andSkoff, 2001). Localization of Golli proteins in boththe nervous and immune systems, suggests theiradditional role in autoimmune diseases (Voskuhl,1998; Clark et al., 1999).

In the present study, we explore further the relation-ship between inflammation, activation of microglia,and OPC proliferation in neonatal mouse brain. In-flammation was induced by the bacterial endotoxinlipopolysaccharide (LPS), which is commonly used toinduce inflammatory response in the brain and to ac-tivate microglia (Lee et al., 1993; Pang et al., 2000).Since both microglia and astrocytes secrete numerouscytokines, it is possible that balance between theseglial cells and the level of their activation might regu-late immune reactions in neuroinflammation (Xiao andLink, 1999). Activated microglia secrete pro-inflamma-tory cytokines that further damage neuronal tissue(Brosnan and Raine, 1996; Rothwell et al., 1997; Panget al., 2000; Duncan et al., 2002; Kong et al., 2002; Caiet al., 2003). However, inflammation could also stimu-late microglia and/or astrocytes to secrete factors thatpromote OPCs proliferation (Arnett et al., 2001; Konget al., 2002; Ransohoff et al., 2002), as reflected in thefinding that remyelination was more efficient when itwas linked to an inflammatory response, while macro-phage depletion impaired it (Kotter et al., 2001). Fur-thermore, activation of microglia and their MHC-IIexpression is an important factor in remyelination(Arnett et al., 2003).

In the present study, we observed that a single treat-ment of LPS on neonatal mouse brain slices and indissociated brain cell cultures promotes OPCs prolifer-ation through activation of microglia and secretion ofGolli proteins. These results are consistent with theidea that moderate inflammation could be beneficial torepair processes in the CNS.

MATERIALS AND METHODSMouse Brain Tissue

Microglia and oligodendrocyte progenitor cells werederived from 1-day-old (P1) ICR mice (Harlan, India-napolis, IN), whereas organotypic slice cultures wereobtained from P1–P3 pups. The animals used in thisresearch were acquired and cared for in accordancewith NIH Guide for the Care and Use of LaboratoryAnimals.

Organotypic Slice Culture

Mouse pups (P1–P3) (n � 5) were anesthetized bycooling and decapitated. The brains were quickly re-moved and put in cold, sterile Hanks’ balanced saltsolution (HBSS) (Sigma, St. Louis, MO) containing0.75% D-glucose saturated with a 95% O2–5% CO2 mix-ture. Brains were embedded in 3% low-melted agarose(Invitrogen, Carlsbad, CA) and cut in the frontal planeinto 400-�m-thick slices, using a Vibroslice (modelVSLM1; World Precision Instruments, Sarasota, FL).Several slices cut from each brain, starting at the fron-tal pole of the lateral ventricle, were immediatelytransferred onto 30-mm membrane inserts (0.4-�m po-rosity; Millipore, Bedford, MA). Inserts were placed in6-well plates containing 1 ml of culture medium, con-sisting of Dulbecco’s modified Eagle’s medium (DMEM,HEPES modification, Sigma) with 0.24% D-glucose,2 mM L-glutamine, N2 (Invitrogen), 5% fetal bovineserum (FBS; Sigma), and antibiotic-antimycotic (In-vitrogen). Using the interface technique (Stoppini etal., 1991), slices were kept in a humidified incubator at37°C with 5% CO2 for 1–5 days. Slices were treated for24 h with 5 �g/ml LPS (Sigma), or medium alone.During the last 24 h in culture 50 �M BrdU (Sigma)was added to each well. After that, slices were fixed in4% paraformaldehyde, cryoprotected with 30% sucrose,mounted on glass slides, and frozen at �70°C. Sliceswere resectioned into 20-�m-thick sections on a cryo-stat and processed for immunofluorescence.

Immunofluorescence

Mouse mixed glial cell cultures (n � 5) or oligoden-drocyte progenitor cell cultures (n � 5) were fixed in 4%paraformaldehyde for 20 min at 4°C and blocked with1% bovine serum albumin (BSA; Sigma), 5% normalgoat serum (Vector, Burlingame, CA), and 0.5%Tween-20 in phosphate-buffered saline (PBS) for30 min, and then incubated overnight at room temper-ature with only anti-Golli antibody, or anti-Golli anti-body with anti-RIP antibody (Developmental StudiesHybridoma Bank, Iowa City, IA). The anti-Golli anti-body was generated from recombinant peptide corre-sponding to 133 amino acids specific for the Golli pro-tein sequence (1:100, a gift from Dr. Campagnoni, LosAngeles, CA). Anti-RIP antibody recognizes early and

458 FILIPOVIC AND ZECEVIC

late oligodendrocyte progenitors (Friedman et al.,1989). Some brain sections from mouse slice cultures,or cell cultures were first blocked with blocking solu-tion, and then incubated overnight with anti-Golli, andmicroglia markers, tomato lectin (TL) or Ricinus com-munis agglutinin (RCA-1) (both at 1:200; Vector). Theother brain sections from slice cultures were incu-bated overnight with anti-Golli, or anti-platelet-de-rived growth factor (PDGF)R-� (1:500, a gift from Stall-cup, La Jolla, CA). PDGFR-� antibody labels earlyoligodendrocyte progenitors (Pringle and Richardson,1993). After washing with PBS, cells or brain sectionswere incubated for 2 h with the appropriate secondaryantibodies, either anti-rabbit IgG conjugated with flu-orescein (1:200, Vector), anti-mouse- IgG Alexa 555(1:400, Molecular Probes, Eugene, OR) or avidin con-jugated with rhodamine (1:200, Vector). For BrdU la-beling, slice culture sections and cell cultures weredenaturated in 2 N HCl for 10 min at room tempera-ture, washed with PBS and rinsed with 0.1 M sodiumborate (pH 8.5) for 10 min. Sections were incubatedfirst with blocking solution (3% NGS/PBS) for 1 h andthen incubated with mouse monoclonal anti-BrdU an-tibody (1:100, BD Biosciences, Palo Alto, CA). To testthe specificity of the reaction, cell cultures or brainsections were treated in the same way, except thatprimary antibodies were omitted.

SDS-PAGE and Western Blotting Analysis

Slices from mouse brains (n � 3) were rinsed twicewith PBS and lysed in PBS containing a mixture ofprotease inhibitors and detergents (1 mM PMSF, 1 �g/ml leupeptin, 2 �g/ml pepstatin, 0.5 mM EDTA, 1%Triton X-100). Tissue lysates were diluted in samplebuffer (62.5 mM Tris-HCl, pH 6.8, 2% sodium dodecylsulfate SDS, 10% glycerol, 5% �-mercaptoethanol,0.01% bromophenol blue) and boiled. Conditioned me-dia from LPS-treated microglia and astrocytes (n � 3),as well as from untreated cells were collected. Beforeanalysis of conditioned media, LPS was removed fromthe medium with polymixin B beads (Sigma, 1-h incu-bation at room temperature). Beads were spun, andconditioned medium was collected and boiled in thesame sample buffer as described above. Equal amountsof proteins (20 �g) were separated by sodium do-decyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (12% polyacrylamide gels) (Mini Protean III cellapparatus, Bio-Rad, Hercules, CA) and transferredelectrophoretically to nitrocellulose membranes (Bio-Rad). After blocking for 2 h (PBS with 5% (w/v) dry-skim milk and 0.1% Tween-20), membranes were incu-bated overnight with a polyclonal primary antibody toGolli (1:500) followed by biotinylated goat anti-rabbitsecondary antibody (1:500, Vector) and streptavidin-conjugated horseradish peroxidase (Vector). Mem-branes were rinsed and developed using an ECL plusWestern blotting detection system and Hyperfilm(Amersham, Little Chalfont, UK). Chemiluminescence

intensity was quantified using software Optimas Ver-sion 5.2 (Bioscan, Botwell, WA).

Dissociated Cell Cultures

Mixed glial cell cultures were generated from brainsof P-1 mice (n � 5). After careful removal of the me-ninges and pial blood vessels, brain tissue was mechan-ically minced and centrifuged (10 min/1,200 rpm). Thepellet was resuspended in HBSS (Sigma) containing25 mM HEPES (Sigma). Following trypsinization(0.025% trypsin, 20 min, 37°C), 100 �g/ml DNase I(Sigma) and 3 mM MgSO4 were added. After centrifu-gation, the pellet was resuspended in DMEM, contain-ing 3.7g/L NaHCO3, 10% FBS, 2 mM L-glutamine(Invitrogen) and antibiotics-antimycotics (2,000 U pen-icillin, 2,000 �g/ml streptomycin sulfate, 0.5 �g/ml am-photericin B) (Invitrogen). Cells were plated in poly-L-lysine-coated 75-mm2 flasks (BD Falcon, FranklinLakes, NJ) at a density of 5 � 107 cells/T75 flask andkept in a humidified atmosphere enriched with 5% CO2at 37°C. For BrdU studies, cells were plated in 8-wellchamber slides at a density of 5 � 104 cells/well; after10 days in culture they were treated with increasingconcentrations of LPS (1 �g/ml, 2.5 �g/ml, 5 �g/ml or10 �g/ml) for 24 h. Mixed glial cell cultures were main-tained for 2 weeks in medium that was changed everythird day. Cells were then shaken on an orbit shakerfor 2 h at 200 rpm at 4°C. The pellet containing micro-glia was resuspended in DMEM containing 10 ng/mlM-CSF (Peprotech, Rocky Hill, NJ), 10% FBS and2 mM L-glutamine, and plated in chamber slides at adensity of 4 � 104 cells/well (BD Falcon). Microglialcells were left for 2 days to allow the cells to startextending processes. Astrocytes that were still at-tached to the bottom of T75 flasks were trypsinized andreseeded in chamber slides in the same medium inwhich mixed glial cells were grown. Both astrocytesand microglia were treated with 5 �g/ml LPS (for 24 h)and then conditioned medium was collected. LPS wasremoved from the medium with polymixin B beads, asdescribed above; the conditioned medium was thenused for further experiments.

Purified oligodendrocyte progenitor cell (OPCs) cul-tures (n � 5) were obtained from mixed glial cell cul-tures prepared from neonatal mouse or rat telenceph-alon, as described previously (Bansal et al., 1999). Inbrief, mixed primary cultures were plated in DMEMcontaining 10% FCS for 10 days and then shaken at200 rpm (30 min at 37°C) in the presence of 5 mML-leucine methyl ester (L-LME, Sigma), to avoid con-tamination with microglia (Thiele et al., 1983). Afterremoval of microglia, new medium was added, and cellswere shaken overnight at 340 rpm. The cells dislodgedafter the “shake” were passed through a nylon mesh(40 �m) and further enriched by differential adhesionon tissue culture dishes to eliminate astrocytes andmicroglia (Bansal and Pfeiffer, 1992). Cells were tritu-rated by passage through a flamed Pasteur pipette in a

459POSITIVE ROLE OF INFLAMMATION

DMEM containing 4% DNase and 10% FBS. Cells werethen plated at the density of 12 � 104 cells/cm2 into50-�g/ml poly-D,L-ornithine-coated (Sigma) four-wellplates (Nunc) in an area of �1 cm2 in the middle of thewell. OPCs were maintained in a serum-free definedmedium, (mN2, Bansal and Pfeiffer, 1992) containingDMEM with 4.5 mg/ml D-glucose, 50 �g/ml humantransferrin, 5 �g/ml bovine pancreatic insulin, 15 nM3,3,5-triiodo-L-tryonine, 30 nM sodium selenium,10 nM D-biotin, 10 nM hydrocortisone, 0.11 mg/ml so-dium pyruvate, penicillin-streptomycin (10 IU/ml and100 mg/ml, respectively), and 0.1% BSA (all ingredi-ents from Sigma). After 2 days, cell cultures weretreated for 24 h by one of the following: 1 �g/ml, 2.5 �g/ml, 5 �g/ml or 10 �g/ml LPS, conditioned media dilutedin mN2 media, in ratio 1:1, conditioned media fromLPS-treated astrocytes (LPS-ACM), microglia (LPS-MCM), untreated microglia (MCM), untreated astro-cytes (ACM), or media alone (controls). Additionally,some OPC cultures were treated with Golli protein(20-kD isoform) (gift from Dr. Campagnoni, Los Ange-les, CA) in different concentrations (50 �g/ml, 25 �g/mlor 5 �g/ml). To study proliferation of OPCs, we added50 �M BrdU (Sigma) to each well for 24 h and identi-fied cells that were in the S phase of the cell cycle byimmunofluorescence using anti-BrdU antibody.

Cell Counting

Oligodendrocytes in mixed or purified OPC cultures(n � 5), which were double labeled with antibodies toGolli and BrdU, or PDGFR-� and BrdU were countedunder a microscope equipped with a digital camera,and expressed as a percentage of the total number ofGolli� or PDGFR-�� cells. In addition to anti-PDGFR-� antibody, we used anti-Golli antibody to la-bel OPCs, from early to late stages (Givogri et al.,2000). Ten adjacent view fields were examined at �20magnification in each culture. Means and standarderrors of the mean (SEM) were calculated, and statis-tical significances between treatments and controlswere estimated using Student’s t-test.

RESULTSLPS Induced Proliferation of Early OPCs in the

Subventricular Zone

To investigate how LPS affects OPCs and to specifyGolli proteins distribution in living brain tissue, wetreated slice cultures of neonatal mouse brain with asingle administration of 5 �g/ml LPS for 24 h (Fig. 1).Slice cultures were examined by immunofluorescenceafter 1 (1 div) and 5 days (5 div) in vitro. The mostprominent change in Golli expression was observedaround the lateral ventricles, in the cortical subven-tricular zone (SVZ). After 1 div, Golli immunoreactivityin the SVZ, which was moderate in control slices(Fig. 1B), increased upon LPS treatment (Fig. 1C).

This coincided with the higher number of PDGFR-��

early OPCs in the SVZ region (Fig. 1D,E). The mor-phology and distribution of Golli� cells resembled thatof PDGFR-�� cells viewed on consecutive sections, sug-gesting that both antibodies labeled early OPCs(Fig. 1B–E). However, Golli� cells were more numer-ous than PDGFR-�� cells, consistent with the notionthat anti-Golli antibody labels additional cell types,such as neurons and microglia. Incorporation of BrdUindicated that, after LPS administration, a greaternumber of both Golli� and PFGFR�� cells were divid-ing (cf. insets in Fig. 1B–E). The effect of LPS, however,was transient, since 5 days after a single LPS applica-tion in vitro, the expression of both Golli proteins andPDGFR-� (not shown) was similar to control slices.

Western blot analysis of slice cultures lysates re-vealed a similar transient increase in Golli expressionafter LPS treatment (Fig. 1H). One day after LPStreatment, the expression of the 20-kD Golli isoformincreased 106 9.7%, while the 29-kD isoform in-creased 28.4 2.1%, compared with controls. Five daysafter LPS treatment, both the 20-kD and the 29-kDisoforms were decreased (82.6 4.0% and 36.6 2.4%respectively), compared with blots from control slices(Fig. 1I).

Because microglia and astrocytes are the main celltypes that respond to LPS, we tested the expression ofGolli proteins in these two cell types. Consistent withprevious studies (Givogri et al., 2002), Golli proteinswere not found in glial fibrillary acidic protein (GFAP)-labeled astrocytes (data not shown), but they werepresent in activated microglia (Filipovic et al., 2002).Resting (ramified) microglia, present under controlconditions in slice culture, exhibited very low expres-sion of Golli proteins (Fig. 1F). This changed after LPStreatment, when microglia became activated, obtainedtypical ameboid morphology characterized by large cellbody and short processes, and increased expression ofGolli proteins (Fig. 1G, inset).

In agreement with results in slice culture, dissoci-ated microglial cell culture exhibited very low expres-sion of Golli proteins under control conditions (Fig. 2A),whereas Golli expression increased following LPStreatment for 24 h (Fig. 2B). This was accompaniedwith morphological transformation of RCA-1� micro-glial cells from resting (ramified) to activated (ame-boid) forms (cf. Fig. 2A and 2B). Ramified microglia hadsmall cell body and long bipolar processes, while ame-boid had larger cell body and short processes.

LPS Influenced OPCs Proliferation inDissociated Cell Culture

The sharp increase of the number of early OPCs inorganotypic slice cultures after LPS treatment led us totest the effect of LPS on proliferation of OPCs in dis-sociated mixed cell cultures. Although Golli antibodycould label both microglia and oligodendrocytes inmixed glial cell cultures, most proliferating cells had


the morphology of oligodendrocyte progenitor cells. Inaddition, Golli� cells were co-labeled with RIP anti-body, a known OPCs marker (Fig. 2C). In cultureswithout any treatment, 29.9 2.2% of Golli� cells wereproliferating, as seen by BrdU immunolabeling(Fig. 2 E,G). After applying LPS in different concentra-tions (1 �g/ml, 2.5 �g/ml, 5 �g/ml, and 10 �g/ml) tomixed cell cultures, proliferation of Golli� OPCs wasnot significantly higher after concentrations of 1 �g/ml(32.6 1.5%, P 0.5, n � 5) or 2.5 �g/ml (35.6 2.8%,P 0.5, n � 5), while 5 �g/ml LPS increased OPCsproliferation for 70% when normalized to control val-ues (51 1.9% vs. 29.9 2.2%; P 0.01, n � 5)(Fig. 2F,G). However, still higher concentration of LPS(10 �g/ml), decreased proliferation of Golli� OPCs(22.4 1.9 vs. 29.9 2.2%, P 0.01, n � 5). Similar toslice cultures, the positive effect of LPS in dissociatedmixed cell cultures was transient, since 5 days after the

LPS treatment (5 �g/ml), proliferation of Golli� OPCswas not different than in controls (25.3 3.7% vs.28.2 1.9, P 0.5, n � 5). Moreover, the morphologyof the Golli� cells changed according to experimentalconditions. In control cultures, Golli� cells were highlybranched with many processes, consistent with themorphology of oligodendrocyte progenitors (Fig. 2D,E).After LPS treatment, however, Golli� cells were bipo-lar, or had fewer and shorter processes (Fig. 2F) con-sistent with immature stage of OPCs development.

Direct vs. Indirect Influence of LPS on OPCs

To determine whether the increased proliferation ofGolli� OPCs was the result of direct action of LPS onOPCs, or of indirect action through factors secretedfrom either astrocytes or microglia, we established pu-

Fig. 1. Golli expression in cortical SVZ in mice slice culture. A:Drawing of frontal level of the mouse brain from which slices were cut.Boxed area represents the site where pictures were taken. In controlslice cultures, very weak expression of Golli proteins is observed in theSVZ (marked by dotted line). B: Cells labeled with Golli� (green) andBrdU� (red) are shown in insets. C: One day after LPS treatment(1 div) Golli expression is upregulated in SVZ. More dividing Golli�cells are shown by BrdU staining (yellow, inset). D: PDGFR-� cells inthe SVZ of the control slide. In the inset, dividing PDGFR�� cells areshown. E: In LPS-treated slices after 1 div more PDGFR-�� cells arepresent and more are proliferating (inset). F: Golli expression is weak

in tomato lectin (TL�) microglial cells in the SVZ of control slices, andis induced in microglia 1 div after LPS treatment (yellow color, insetshows higher magnification). Western blot analysis of slice culturetissue shows transient upregulation of Golli proteins 1 div after LPStreatment (G), but not after 5 div (H). Histogram shows the percent-age of change in chemiluminescence density of Golli isoforms (20 kDand 29 kD) after LPS treatment, compared with controls. Values aremeans SEM from thee different experiments (I). SVZ, subventricu-lar zone; lv, lateral ventricle; Cx, cortex; cc, corpus callosum; BG,basal ganglia; m, medial side. Scale bars � 100 �m in B–G; 25 �m inB for inserts in B,C,D,E.


rified OPCs cell cultures. In these cultures, under con-trol conditions, Golli proteins were expressed by OPCswith both immature and more mature morphology(Fig. 3A). Consistent with a previous result, only cellswith a round cell body and small number of shortprocesses, characteristic of immature OPCs, were pro-liferating as judged by their BrdU expression (Fig. 3C).After LPS treatment, cells reacted by changing theirmorphology, with most cells displaying only few pro-cesses branching from the cell body (Fig. 3B,D), but thepositive effect on cell proliferation was not demon-strated. Thus, after 1 �g/ml LPS treatment, prolifera-

tion of Golli� OPCs was not significantly different thanin controls (30.5 5.3% vs. 38.5 0.7%, P 0.5, n �5), while higher concentrations of LPS (2.5 �g/ml, 5 �g/ml and 10 �g/ml) reduced the number of proliferatingOPCs (27.4 4.4%, 26.9 3.3 or 19.6 4.6 respec-tively, P 0.01, n � 5) (Fig. 3E). Five days after LPStreatment OPCs proliferation was similar to controls(25.3 3.4 vs. 23.2 1.5, P 0.5, n � 5), confirmingtransient effect of LPS. These results suggested thatthe presence of other glial cells and/or glial-derivedfactors is necessary for the LPS to promote prolifera-tion of OPCs. To test this possibility further, we treated

Fig. 2. LPS induces upregulation of Golli�proteins in microglial cell cultures, and increasedproliferation of oligodendrocyte progenitor cells(OPCs) in mixed glial cell cultures. In mousemicroglial cell culture (RCA-1�, red), a low levelof Golli proteins (green) is observed in controlconditions (A), while expression of Golli proteinsis upregulated after LPS treatment (B, yellow,arrows). C: In mixed control glial cell cultures,Golli and RIP are co-localized in OPCs (yellow).D: Typical morphology of a Golli� oligodendro-cyte progenitor cell. E: Golli� cells in control cul-tures rarely incorporate BrdU (in red). After LPStreatment, Golli� OPCs have fewer processesand more cells incorporate BrdU (arrows, F). His-togram shows that 1 day after LPS treatment,the largest effect on BrdU incorporation in Golli�cells is seen in cultures treated with 5 �g/ml ofLPS (P** 0.01, n � 5). Lower concentrations ofLPS (1 �g/ml and 2.5 �g/ml) do not induce sig-nificant increase in OPC proliferation, whereas10 �g/ml LPS induce a decrease in BrdU incor-poration in Golli� OPCs. Five days after LPStreatment (5 �g/ml) proliferation of these cells issimilar to control cultures. (G). Scale bars �100 �m in A for A,B,E,F; 50 �m in C for C,D.


purified OPC cultures with conditioned media fromeither LPS-treated microglia (LPS-MCM), LPS-treatedastrocytes (LPS-ACM), untreated microglia (MCM), orastrocytes (ACM). The number of proliferating OPCs(Golli� cells) increased significantly only in the pres-ence of LPS-MCM (56 3.1% vs. 38.5 0.7% in con-trols, P 0.01, n � 5), as judged by their incorporationof BrdU. This was not the case when LPS-ACM wasadded (44.6 4.2%, P 0.5, n � 5; Fig. 3F), or whenconditioned media from untreated microglia or astro-cytes were added to purified OPC cultures (not shown).These results suggested that factors derived from acti-vated microglia, and not astrocytes, were potent pro-moters of OPCs proliferation.

Next we tested the possibility that Golli proteinswere secreted in LPS-MCM conditioned media after1 day of treatment. Indeed, Western blots analysis(n � 3) of conditioned media from untreated micro-

glia or astrocytes, and from LPS-treated microglia orastrocytes, demonstrated Golli proteins only in LPS-MCM (Fig. 3H). Of the two Golli isoforms (29 kD and20 kD), expression of the protein with lower molecu-lar weight was more prominent. This result led us toexamine the direct effect of the 20-kD isoform of Golliprotein on the OPC cultures (Fig. 3G). When purifiedOPC cultures (n � 3) were treated with the 20-kDGolli protein (24 h, 50 �g/ml), increased proliferationof PDGFR-�� OPCs was observed compared withcontrols (47.8 5% vs. 32.2 0.8%, P 0.01, n � 3).When lower concentrations of Golli protein (25 �g/mland 5 �g/ml) were added to the cultures, prolifera-tion of OPCs was similar to that of controls (35.8 1.6% and 31.4 0.8%, respectively, P 0.5, n � 3).This result suggests that the 20-kD isoform of Golliproteins was promoting OPC proliferation in a dose-dependent manner, and that high concentration was

Fig. 3. Effect of LPS and LPS-conditioned media on purified ol-igodendrocyte progenitor cells(OPC) cultures. A: In purifiedOPC cultures, under control con-ditions, Golli proteins are presentin both bipolar (arrows) and mul-tipolar cells. B: After LPS treat-ment, most OPCs retract pro-cesses (arrows). BrdU labeling ofGolli� cells in control conditions(C), and after LPS treatment (D).Dose and time response for LPS-induced proliferation of OPCs.Pure OPC cultures were treat-ed with different doses of LPS(1, 2.5, 5, 10 �g/ml). Histogramshows, that compared with con-trols, less Golli� cells incorporateBrdU after LPS treatment (P** 0.01, n � 5). E: Five days afterLPS (5 �g/ml) treatment, prolif-eration of OPCs is similar as incontrol cultures. F: Addition ofconditioned medium from LPS-treated microglia (LPS-MCM) in-creased Golli� cells proliferation(P** 0.01, n � 5), while condi-tioned medium from LPS-treatedastrocytes (LPS-ACM) has lesseffect on proliferation (P 0.5,n � 5). G: When purified rodentOPC cultures were treated with50 �g/ml Golli protein, signifi-cant increase in proliferation ofPDGFR-�� OPCs was observed.Western blot analysis of condi-tioned media collected after 5 �g/ml LPS treatment for 1 day fromLPS-MCM, LPS-ACM, and un-treated microglia and astrocytes(MCM and ACM) shows the pres-ence of both Golli isoforms only inLPS-MCM. Typical result ob-tained from three different exper-iments (H). Scale bars � 50 �m inA for A,B; 25 �m in C for C,D.


needed to produce this effect, which may be the caseduring inflammation.

Taken together, these results strongly suggest thatLPS treatment activates microglia and induces secre-tion of Golli proteins, which could be one of the factorsimportant for additional OPC proliferation (Fig. 4).

DISCUSSION

We have demonstrated in he present study that LPS-induced inflammation has a positive effect on prolifer-ation of OPCs and that this process is, at least partly,mediated through secretion of Golli proteins from acti-vated microglia. Our major findings are: (1) LPS treat-ment for 1 day promotes proliferation of OPCs in thecortical SVZ of the neonatal mouse brain and in mixedglial cell cultures; (2) the same effect is achieved whenpurified OPC cultures are treated by conditioned mediafrom LPS-activated microglia; (3) in the absence ofeither microglia or microglia-derived factors, LPS doesnot promote proliferation of OPCs, suggesting that itacts, in part, through this cell type; (4) Golli proteins,in particular the 20-kD isoform, are one of the media-tors secreted from activated microglia that promoteOPC proliferation; and (5) the effect of LPS on OPCproliferation is transient, as 5 days after the treatment,proliferation is similar to that of untreated controls.

Inflammatory Conditions Increase Expressionof Golli Proteins in Activated Microglia

LPS is a bacterial endotoxin commonly used to acti-vate microglia and induce microglia to secrete pro-inflammatory cytokines (Lee et al., 1993; Pang et al.,2000; Lehnardt et al., 2002). In our study, LPS treat-ment resulted in transformation of immune cells of thebrain, microglia, from ramified (resting) to ameboid(activated) form. In addition, upregulation of Golli pro-teins followed microglia activation. These results are inaccord with our previous studies, which demonstratedthat Golli proteins were present in human fetal micro-glia (Tosic et al., 2002) and were upregulated in acti-vated microglia, oligodendrocyte progenitors, and de-myelinated axons around MS lesions (Filipovic et al.,

2002). In the region around lesions, Golli� microglialcells had ameboid morphology, short processes, andexpressed HLA-DR antigen, all of which are character-istics of activated microglia (Filipovic et al., 2002). Sev-eral studies demonstrated that in addition to the CNS,Golli mRNAs and proteins were found in immune sys-tem as well (Campagnoni et al., 1993; Fritz and Kal-vakolanu, 1995; Pribyl et al., 1996a; Givogri et al.,2000; Feng et al., 2000). In addition, Golli mRNAs wereincreased in lymph nodes of mice with experimentalautoimmune encephalomyelitis (EAE) (Voskuhl, 1998),and Golli proteins were able to induce mild paralysis inrats (Clark et al., 1999). Furthermore, the relapse ratefollowing EAE induction is reduced in Golli gene defi-cient mice (Voskuhl et al., 2003). These findings to-gether suggest that Golli proteins might have a role inimmune reactions and pathology of demyelination. Re-sults from the present study shed somewhat differentlight and suggest that accumulation of Golli proteins inactivated microglia, observed around MS lesions (Fili-povic et al., 2002), could be beneficial for OPC prolifer-ation observed in “shadow plaques” (Raine et al., 1981,Noseworthy et al., 2000; Steinman, 2001; Hemmer etal., 2002; Franklin, 2002).

LPS-Activated Microglia Influence Proliferationof Early OPCs

Results in mouse slice culture show that LPS treat-ment, besides activation of resident microglia, affectsother cell types present in the CNS. Specifically, earlyOPCs localized in the SVZ responded by proliferationto LPS stimulus after only 24 h. The resulting increasein OPC proliferation might be either through direct cellcontacts (in mixed cultures) or through secreted factors(when conditioned media from LPS-treated microgliawas used), or both. In contrast, other reports of ratstudies have shown that administration of LPS maycause death of OPCs in the presence of microglia (Panget al., 2000; Lehnardt et al., 2002). It is possible thatdifferent experimental protocols can explain this dis-crepancy. We used short (24-h) exposure and high con-centrations (5 �g/ml) of LPS, whereas in these studiesdifferent conditions (concentrations and exposure time)were applied. Our rationale was to use a single appli-

Fig. 4. Schematic representa-tion of possible mechanism bywhich LPS treatment promotesproliferation of oligodendrocyteprogenitor cells (OPCs). In thismodel, the inflammatory agent,LPS first transforms microgliafrom a resting (1) to an activatedform (2). This activation resultsin accumulation (2), and subse-quent release of Golli proteins(3), which are involved in the in-crease of OPC proliferation (4).


cation of LPS to mimic the acute phase in MS, whereboth inflammation and remyelination are present(Franklin, 2002; Owens, 2003). We have shown that5 �g/ml of LPS after 1 day induced significant increasein proliferation of OPCs. However, the positive effect ofLPS was transient, since 5 days after single LPS ap-plication in vitro, the number of proliferating OPCswas similar to controls, as shown in both slice and cellcultures. Dose response experiments in dissociated cellcultures have shown that LPS in concentration of 5 �g/ml induced proliferation of OPCs, while higher concen-trations (10 �g/ml) decreased the number of proliferat-ing cells, suggesting concentration-dependent effect ofinflammation. We further tested the possibility thatGolli proteins are secreted in the media from LPS-treated microglial cells (LPS-MCM). Indeed, both iso-forms of Golli proteins were only present in LPS-MCM,and were absent in other conditioned or control media.In addition, in slice and dissociated microglial cell cul-tures the 20-kD isoform of Golli proteins was elevatedafter LPS treatment, in accord with Western blot anal-ysis of brain homogenates from MS patients, whichdemonstrated an increase of the same Golli isoform(our unpublished data). Furthermore, the 20-kD Golliisoform was able to induce an increase in cell prolifer-ation when added to purified OPC cell culture.

Thus, we propose that upon the onset of LPS- in-duced inflammation, Golli proteins are upregulated inactivated microglia and secreted outside the cells, pos-sibly via multivesicular bodies that fuse with cell mem-brane (Boes et al., 2004).

Our results are compatible with the hypothesis thatremyelination appears to be more efficient when it islinked to an inflammatory response, while macrophagedepletion impairs it (Kotter et al., 2001; Arnett et al.,2003). In MS, activated microglia are thought to con-tribute to demyelination by secreting pro-inflammatorycytokines, such as tumor necrosis factor-� (TNF-�) andinterleukin-1� (IL-1�), which induce destruction ofneurons and oligodendrocytes (Colton and Gilbert,1987; Merill, 1991; Banati et al., 1993; Dickson et al.,1993; Brosnan and Raine, 1996; Rothwell et al., 1997;Cai et al., 2003). Other reports, however, argue thatthese cytokines also enhance cell proliferation and dif-ferentiation during development (Giulian et al., 1988),and may promote cell repair and remyelination (Masonet al., 2001; Arnett et al., 2001, 2003; Ransohoff et al.,2002). Similar to this, it was demonstrated that acti-vated microglia protect oligodendrocytes by secretinginsulin-growth factor 2 (IGF-2) (Nicholas et al., 2002).This dual function of microglia has been often dis-cussed and prompted conclusions about microglial het-erogeneity in vivo (e.g., Andjelkovic et al., 1998; Car-son, 2002). In addition to microglia, astrocytes mayproduce remyelination-associated growth factors suchas PDGF, fibroblast growth factor (FGF-2), IGF-1, andtransforming growth factor-�1 (TGF-�1) that induceproliferation of OPCs (Hicks and Franklin, 1999). How-ever, in our study, secreted factors from astrocytes

were less efficient in promoting OPC proliferation thanmicroglial.

In conclusion, results presented in this paper sup-port the idea that microglia activated during onset ofinflammation are induced to secrete mediators, such asGolli proteins, which may have a positive role on pro-liferation of OPCs (Fig. 4). Since increased prolifera-tion of OPCs would be beneficial for remyelination inMS, further studies are needed to specify the level ofinflammation necessary to balance microglia-derivedfactors required for sustained OPC proliferation.

ACKNOWLEDGMENTS

The authors thank Drs. A. Campagnoni and W.B.Stallcup for providing Golli and PDGFR-� antibodies.We are grateful to Drs. S. Pfeiffer, R. Bansal, S.Hewett, M. Tosic and S. Sivaramakrishnan for theirvaluable comments on the early version of manuscript,S. Winkler for her help with oligodendrocytes progen-itor cell cultures, and B. Howard for editing the manu-script.

REFERENCES

Andjelkovic A, Nikolic B, Pachter JS, Zecevic N. 1998. Macrophages/microglial cells in human central nervous system during develop-ment: an immunohistochemical study. Brain Res 814:13–25.

Arnett HA, Mason J, Marino M, Suzuki K, Matsushima GK, TingJP-Y. 2001. TNF� promotes proliferation of oligodendrocyte progen-itors and remyelination. Nat Neurosci 4:1116–1121.

Arnett HA, Wang Y, Matsushima GK, Suzuki K, Ting JP. 2003.Functional genomic analysis of remyelination reveals importance ofinflammation in oligodendrocyte regeneration. J Neurosci 23:9824–9832.

Banati RB, Gehrman P, Schubert GW, Kreutzberg GW. 1993. Cyto-toxicity of microglia. Glia 7:111–118.

Bansal R, Pfeiffer SE. 1992. Novel stage in the oligodendrocyte lin-eage defined by reactivity of progenitors with R-mAb prior to O1antigalactocerebroside. J Neurosci Res 32:309–316.

Bansal R, Winkler S, Bheddah S. 1999. Negative regulation of oligo-dendrocyte differentiation by galactosphingolipids. J Neurosci 19:7913–7924.

Boes M, Cuvillier A, Ploegh H. 2004. Membrane specializations andendosome maturation in dendritic cells and B cells. Trends Cell Biol14:175–183.

Brosnan CF, Raine CS. 1996. Mechanisms of immune injury in mul-tiple sclerosis. Brain Pathol 6:243–257.

Cai Z, Pang Y, Lin S, Rhodes PG. 2003. Differential roles of tumornecrosis factor-alpha and interleukin-beta in lipopolysaccharide-induced brain injury in neonatal rat. Brain Res 975:37–47.

Campagnoni AT, Skoff RP. 2001. The pathology of myelin mutantsreveal novel biological functions of the MBP and PLP genes. BrainPathol 11:74–91.

Campagnoni AT, Pribyl TM, Campagnoni CW, Kampf K, Amur-Umarjee S, Landry CF, Handley VW, Newman SL, Garbay B,Kitamura K. 1993. Structure and developmental regulation of Golli-mbp, a 105-kilobase gene that encompasses the myelin basic pro-tein gene and is expressed in cells in the oligodendrocyte lineage inthe brain. J Biol Chem 268:4930–4938.

Carson MJ. 2002. Microglia as liaisons between the immune andcentral nervous systems: functional implications for multiple scle-rosis. Glia 40:218–231.

Clark L, Otvos L Jr, Stein PL, Zhang XM, Skorupa AF, Lesh GE,McMorris FA, Heber-Katz E. 1999. Golli-induced paralysis: a studyin anergy and disease. J Immunol 162:4300–4310.

Colton CA, Gilbert DL. 1987. Production of superoxide anions by aCNS macrophage, the microglia. FEBS Lett 223:284–288.


Dickson DW, Lee SC, Mattiace LA, Yen SH, Brosnan C. 1993. Micro-glia and cytokines in neurological disease, with special reference toAIDS and Alzheimer’s disease. Glia 7:75–83.

Duncan JR, Cock ML, Scheerinck JP, Westcott KT, McLean C, Har-ding R, Rees SM. 2002. White matter injury after repeated endo-toxin exposure in the preterm ovine fetus. Pediatr Res 52:941–949.

Feng J-M, Givogri MI, Bongarzone ER, Campagnoni C, Jacobs E,Handley VW, Schonmann V, Campagnoni AT. 2000. Thymocytesexpress the golli products of the Myelin Basic Protein gene andlevels of expression are age stage dependent. J Immunol 165:5443–5450.

Fernandes AO, Campagnoni CW, Kampf K, Feng JM, Handley VW,Schonmann V, Bongarzone ER, Reyes S, Campagnoni AT. 2004.Identification of a protein that interacts with the golli-myelin basicprotein and with nuclear LIM interactor in the nervous system.J Neurosci Res 75:461–471.

Filipovic R, Rakic S, Zecevic N. 2002. Expression of Golli proteins inadult human brain and multiple sclerosis lesions. J Neuroimmunol127:1–12.

Franklin RJM. 2002. Why does remyelination fail in multiple sclero-sis? Nat Rev:3(a) 705–714.

Friedman B, Hockfield S, Black JA, Woodruff KA, Waxman SG. 1989.In situ demonstration of mature oligodendrocytes and their pro-cesses: an immunocytochemical study with a new monoclonal anti-body, Rip. Glia 2:380–390.

Fritz RB, Kalvakolanu I. 1995. Thymic expression of the golli-myelinbasic protein gene in the SJL/J mouse. J Neuroimmunol 57:93–99.

Givogri MI, Bongarzone ER, Compagnoni AT. 2000. New insight onthe biology of myelin basic protein gene: the neural-immune con-nection. J Neurosci Res 59:153–159.

Giulian D, Young DG, Woodward J, Brown DC, Lachmann LB. 1988.Interleukin-1 is an astroglial growth factor in the developing brain.J Neurosci 8:709–714.

Hemmer B, Archelos JJ, Hartung H-P. 2002. New concepts in theimmunopathogenesis of multiple sclerosis. Nat Rev 3:291–301.

Hicks GL, Franklin RJM. 1999. Distinctive patterns of PDGF-A,FGF-2, IGF-1 and TGF-�1 gene expression during remyelination ofexperimentally-induced spinal cord demyelination. Mol Cell Neuro-sci 14:153–168.

Keirstead HS, Blakemore WF. 1999. The role of oligodendrocytes andoligodendrocyte progenitors in CNS remyelination. Adv Exp MedBiol 468:183–197.

Kong GY, Kristensson K, Bentivoglio M. 2002. Reaction of mousebrain oligodendrocytes and their precursors, astrocytes and micro-glia, to proinflammatory mediators circulating in the cerebral fluid.Glia 37:191–205.

Kotter MR, Setzu A, Sim FJ, van Rooijen N, Franklin RJM. 2001.Macrophage depletion impairs oligodendrocyte remyelination fol-lowing lysolecithin-induced demyelination. Glia 35:204–212.

Landry CF, Pribyl TM, Ellison JA, Givogri MI, Kampf K, CampagnoniCW, Campagnoni AT. 1998. Embryonic expression of the MyelinBasic Protein gene: identification of a promoter region that targetstransgene expression to pioneer neurons. J Neurosci 18:7315–7327.

Lee SC, Liu W, Dickson DW, Branch DW, Hammond E, Mitchel MD.1993. Cytokine production by human fetal microglia and astrocytes.Differential induction by lipopolysaccharide and IL-1�. J Immunol150:2659–2667.

Lehnardt S, Lachance C, Patrizi S, Lefebre S, Follet PL, Jensen FE,Rosenberg PA, Volpe JJ, Vartanian T. 2002. The Toll-Like ReceptorTLR4 is necessary for lipopolysaccharide-induced oligodendrocyteinjury in the CNS. J Neurosci 22:2478–2486.

Liu H, MacKenzie-Graham AJ, Palaszynski K, Liva S, Voskuhl RR.2001. “Classic” myelin basic proteins are expressed in lymphoidtissue macrophages. J Neuroimmunol 116:83–93.

Mason JL, Suzuki, K, Chaplin D, Matsushima GK. 2001. Interleukin1 promotes repair of the CNS. J Neurosci 21:7046–7053.

Merill JE. 1991. Effects of interleukin-1 and tumor necrosis factor onastrocytes, microglia, oligodendrocytes and glial precursors in vitro.Dev Neurosci 13:130–137.

Nicholas RS, Stevens S, Wing MG, Compston DA. 2002. Microglia-derived IGF-2 prevents TNF� induced death of mature oligoden-drocytes in vitro. J Neuroimmunol 124:36–44.

Noseworthy JH, Lucchinetti C, Rodriguez M, Weinshenker BG. 2000.Multiple sclerosis. N Engl J Med 343:938–952.

Owens T. 2003. The enigma of multiple sclerosis: inflammation andneurodegeneration cause heterogeneous dysfunction and damage.Curr Opin Neurol 16:259–265.

Pang Y, Cai Z, Rhodes PG. 2000. Effects of lipopolysaccharide onoligodendrocyte progenitor cells are mediated by astrocytes andmicroglia. J Neurosci Res 62:510–520.

Pribyl TM, Campagnoni CM, Kampf K, Handley VW, CampagnoniAT. 1996a. The major myelin protein genes are expressed in thehuman thymus. J Neurosci Res 45:812–819.

Pribyl TM, Campagnoni CM, Kampf K, Ellison JA, Landry CF,Kashima T, McMahon J, Campagnoni AT. 1996b. Expression ofmyelin basic protein gene locus in neurons and oligodendrocytes inthe human fetal central nervous system. J Comp Neurol 374:342–353.

Pringle NP, Richardson WD. 1993. A singularity of PDGF alpha-receptor expression in the dorsoventral axis of the neural tube maydefine the origin of the oligodendrocyte lineage. Development 117:525–533.

Raine CS, Scheinberg L, Waltz JM. 1981. Multiple sclerosis: oligoden-drocyte survival and proliferation in an active established lesion.Lab Invest 45:534–546.

Ransohoff RM, Howe CL, Rodriguez M. 2002. Growth factor treat-ment of demyelinating disease: at last, a leap into the light. TrendsImmunol 23:512–516.

Reyes SD, Campagnoni AT. 2002. Two separate domains in the gollimyelin basic proteins are responsible for nuclear targeting andprocess extension in transfected cells. J Neurosci Res 69:587–596.

Rothwell N, Allan S, Toulmond S. 1997. Perspective series: cytokinesand the brain. J Clin Invest 100:2648–2652.

Steinman L. 2001. Multiple sclerosis: a two-stage disease. Nat Immu-nol 2:762–764.

Stoppini LA, Buchs D, Muller A. 1991. A simple method for organo-typic cultures of nervous tissue. J Neurosci Methods 37:173–182.

Thiele DL, Kurosaka M, Lipsky PE. 1983. Phenotype of accessory cellnecessary for mitogen-stimulated T and B responses in humanperipheral blood: delineation by its sensitivity to the lysosomotricagent, L-leucine methyl ester. J Immunol 131:2282–2290.

Tosic M, Rakic S, Matthieu JM, Zecevic N. 2002. Identification of Golliand Myelin basic proteins in human brain during early develop-ment. Glia 37:219–228.

Trapp BD, Peterson J, Ranshoff RM, Rudick RA, Mork S, Bo L. 1998.Axon transaction in the lesions of multiple sclerosis. N Engl J Med338:278–285.

Voskuhl R. 1998. Myelin protein expression in lymphoid tissue: im-plications for peripheral tolerance. Immunol Rev 164:81–94.

Voskuhl RR, Pribyl TM, Kampf K, Handley V, Liu HB, CampagnoniC, Soldan SS, Messing A, Campagnoni AT. 2003. Experimentalautoimmune encephalomyelitis relapses are reduced in heterozy-gous golli MBP knockout mice. J Neuroimmunol 139:44 –50.

Xiao B-G, Link H. 1999. Is there a balance between microglia andastrocytes in regulating Th1/Th2 cell responses and neuropatholo-gies? Trends Immunol Today 20:477–479.


lipopolysaccharide affects golli expression and promotes proliferation of oligodendrocyte...

Documents