Increased_expression_of_bioactive_chemokines_in_human_cerebromicrovascular_endothelial_cells_and

  • 格式:pdf
  • 大小:1.38 MB
  • 文档页数:13

Ž.Journal of Neuroimmunology1011999148–160 r locate r jneuroimIncreased expression of bioactive chemokines in human cerebromicrovascular endothelial cells and astrocytes subjected tosimulated ischemia in vitroWandong Zhang a,Catherine Smith a,Anthony Shapiro a,Robert Monette a,James Hutchison b,Danica Stanimirovic a,)a Institute for Biological Sciences,National Research Council of Canada,Montreal Road Campus,Building M-54,Ottawa,ON,Canada K1A0R6b Children’s Hospital of Eastern Ontario,Ottawa,ON,Canada K1H8L1Received30March1999;received in revised form7July1999;accepted7July1999AbstractLeukocyte infiltration into the brain has been implicated in the development of ischemic brain damage.In this study,simulated in vitroŽ. ischemia r reperfusion and IL-1b were found to up-regulate both the expression of intercellular adhesion molecule-1ICAM-1in culturedŽ.human cerebromicrovascular endothelial cells HCEC and the adhesion of allogenic neutrophils to HCEC.Both HCEC and human fetal Ž.astrocytes FHAS also responded to IL-1b and to in vitro ischemia r reperfusion by a pronounced up-regulation of IL-8and MCP-1 mRNA and by increased release of IL-8and MCP-1in cell culture media.FHAS were found to release30-times higher levels of MCP-1 than HCEC under both basal and ischemic conditions.However,100u r ml IL-1b induced greater stimulation of both IL-8and MCP-1Ž.Ž. secretion in HCEC50and20times above controls,respectively than in FHAS three and two times above controls,respectively.IL-8 was the principal neutrophil chemoattractant released from IL-1b-treated HCEC,since IL-8antibody completely inhibited neutrophil chemotaxis enticed by HCEC media.However,the IL-8antibody neutralized only50%of IL-1b-stimulated neutrophil chemoattractants released from FHAS,and40%–60%of ischemia-stimulated chemotactic activity released by either HCEC or FHAS.These results suggest that simulated in vitro ischemia,in addition to IL-8and MCP-1,stimulates secretion of other bioactive chemokines from HCEC and FHAS.q1999Elsevier Science B.V.All rights reserved.Keywords:Chemokines;Adhesion molecules;Cerebral endothelial cells;Astrocytes;Human;Hypoxia1.IntroductionBrain inflammation has been implicated in the pathogenic cascade leading to secondary ischemic brain Ž.damage Kim,1996;Feuerstein et al.,1997.Neutrophils are commonly the first inflammatory cells to infiltrate the ischemic brain,followed by mononuclear phagocytes Ž.Barone et al.,1991;Feuerstein et al.,1994.The recruit-ment of inflammatory cells into the brain is mediated byŽleukocyte r endothelial adhesion molecules Kishimoto and.Rothlein,1994;Kim,1996and by the creation of specificŽchemotactic gradients in the ischemic brain tissue Glabin-ski et al.,1995;Feuerstein et al.,1997;Ransohoff andCorresponding author.Tel.:q1-613-993-3730;fax:q1-613-941-4475;E-mail:danica.stanimirovic@nrc.ca.ŽTani,1998.The reduction of circulating neutrophils i.e., .Ž.neutropenia Matsuo et al.,1994,administration of anti-bodies against endothelial or leukocyte-expressed adhesionŽmolecules in animal models of cerebral ischemia Chopp .et al.,1994,and knock-out of the gene encoding theŽ.Žintercellular adhesion molecule-1ICAM-1Soriano et .al.,1996,have been shown to limit neutrophil infiltration into the brain and to reduce infarct size and brain swelling.Chemokines,a family of8–12kDa peptides,have been shown to entice selective leukocyte recruitment at periph-Ž. eral inflammation sites Rollins,1997;Baggiolini,1998. Chemokine selectivity to subpopulation of leukocytes is determined by the distribution of four cysteines in a highlyŽconserved N-terminal domain,such as that a CXC;proto-.typic member IL-8chemokines primarily attract neu-Ž.trophils,b CC;prototypic member MCP-1chemokinesŽattract both monocytes and lymphocytes,g C;lympho-()W.Zhang et al.r Journal of Neuroimmunology1011999148–160149.Žtactin chemokines draw lymphocytes,and d CX C;neu-3.Žrotactin both neutrophils and monocytes Baggiolini et al.,.1997;Bazan et al.,1997.Recent studies have shown that chemokines and their receptors are expressed in a varietyŽof brain cells,including astrocytes Hayashi et al.,1995;.Ž. Gourmala et al.,1997,microglia Ehrlich et al.,1998,Ž. macrophages Calvo et al.,1996;Gourmala et al.,1997,Ž.and neurons Horuk et al.,1997.Chemokines have been suggested to play a role in the pathophysiology of brain injury accompanying autoim-Ž.Žmune Ransohoff,1997,post-traumatic Fan et al.,1995;.Ransohoff and Tani,1998,and,more recently,post-Žischemic brain inflammation Kim et al.,1995;Matsumoto .Žet al.,1997.Increased levels of MCP-1Kim et al.,1995;.Wang et al.,1995,macrophage inflammatory protein-1Ž.Ž.MIP-1Kim et al.,1995;Takami et al.,1997,and IL-8Ž.Matsumoto et al.,1997have been detected in the is-chemic rat brain,and systemic administration of anti-IL-8 antibody has been shown to reduce cerebral edema,Ž.blood-brain barrier BBB permeability,and infarct size inŽ. experimental models of stroke Matsumoto et al.,1997.Since neutrophils and monocytes are the inflammatoryŽ.Žcells commonly found in ischemic brain lesion s Barone.et al.,1991;Feuerstein et al.,1994,two groups of chemokines are likely important in recruiting these cellsŽ.into the ischemic brain:a neutrophil chemokines,belong-Ž.ing to a and d families,and b monocyte chemokines, belonging to b family.A prototype chemokine from eachŽ.Žof these families,i.e.,IL-8a family,neurotactin d .Ž.family,and MCP-1b family were analyzed in this studyŽ. in both cultured human cerebral endothelial cells HCECŽ.and human fetal astrocytes FHAS subjected to simulated in vitro ischemia.The results show that both cell types respond to IL-1b and to in vitro ischemia by up-regulating the expression r secretion of IL-8and MCP-1,and other neutrophil chemoattractants,and are likely important par-ticipants in the process of inflammatory cell recruitment into the ischemic brain.2.Materials and methods2.1.MaterialsAll culture media were obtained from Gibco BRL Ž.Gaithersburg,MD.Fetal bovine serum was purchasedŽ.from Hyclone Logan,UT.Endothelial cell growth sup-plement,ITS e Premix,and human fibronectin were pur-Žchased from Collaborative Biomedical Products Bedford, .MA.Human serum and gelatin were obtained from Sigma Ž.St.Louis,MO.The mouse melanoma cell line,Cloud-Ž.man S91clone M-3cells was from the American TypeŽ.Culture Collection Rockville,MD.IL-1b,IL-8,antibodyŽ.to human ICAM-1clone CD54,and antibody to humanŽIL-8were obtained from Upstate Biotechnology Lake.Placid,NY.Colloidal gold-conjugated goat anti-mouse IgG and peroxidase-conjugated goat anti-mouse IgG anti-Žbodies were from Accurate Chemical and Scientific West-.Ž. bury,NY.Silver-enhancing solution IntenSE e M wasŽ.from Amersham Canada Oakville,ON,while the peroxi-XŽdase substrate,2,2-azinobis3-ethylbenzthiazoline-6-.Ž.sulfonic acid ImmunoPure ABTS was from PierceŽ.Ž. Rockford,IL.Diethyl pyrocarbonate DEPC,Tri reagent, Trizma,collagenase IV,and the Ficoll gradients, Histopaque-1077and-1119,were obtained from Sigma. Taq DNA polymerase was purchased from Promega Ž.Madison,WI.Superscript RT-II,deoxyribonucleitides,Ž.Oligo dT primer,phenol,and agarose were pur-12–18chased from Gibco BRL.Chloroform,2-propanol,for-mamide,and formaldehyde were from Fisher Scientific Ž.Nepean,ON.2.2.Cell culturesThe protocols used in these studies have been approved by the Human Research Ethics Committees of the National Research Council of Canada,Montreal Neurological Insti-Ž.tute McGill University,and the Children’s Hospital of Eastern Ontario.HCEC were isolated using previously described proto-Ž. cols Gerhart et al.,1988;Stanimirovic et al.,1996. Briefly,the samples of human temporal cortex removed surgically for the treatment of epilepsy were dissected, minced,homogenized and filtered sequentially throughŽ350and112-m m mesh nylon nets Nitex,Tetko,Elmsford,.NY,USA.The pellets collected from filtrates were cen-Ž.trifuged3,000=g;15min;48C in20%dextran and then filtered through a20-m m mesh nylon net.Microvessels and capillaries retained on the net were dissociated with1 mg r ml type IV collagenase for15min at378C and platedŽin growth media containing65%medium M199Earle’s salts,25mM Hepes,4.35g r l sodium bicarbonate and3.mM L-glutamine,10%fetal calf serum,5%human serum,Ž20%murine melanoma cell mouse melanoma,Cloudman.S91,clone M-3,melanin producing cells-conditioned me-dia,5m g r ml insulin,5m g r ml transferrin,5ng r ml selenium,and10m g r ml endothelial cell growth supple-ment.Endothelial cell colonies emerging from attached microvessels4–5days after seeding were isolated using Ž.cloning rings BELCO Glass,Vineland,NJ and2–3of these cloned colonies were pooled and further passaged. Purity of HCEC cultures generated by these procedures is routinely assessed by the immunocytochemical staining for Factor VIII-related antigen and the lack of staining for smooth muscle a-actin,and is estimated to be)95%. The morphological,phenotypic,biochemical and func-tional characteristics of these HCEC cultures have beenŽdescribed in detail previously Stanimirovic et al.,1996;.Muruganandam et al.,1997.()W.Zhang et al.r Journal of Neuroimmunology1011999148–160 150Ž.Cultures of fetal10–18weeks of gestation humanŽ.brain astrocytes FHAS were provided by Drs.J.AntelŽand W.Yong Montreal Neurological Institute,Montreal, .QC,Canada.These cultures were prepared using previ-Ž.ously described procedures Yong et al.,1992.Briefly, brain samples were trypsinized,homogenized and filtered through a130-m m mesh.Pellets were then resuspended in a medium containing95%Dulbecco’s modified Eagle’sŽ.Ž.medium DME4500mg r l glucose and5%fetal bovine serum,and plated onto poly-L-lysine coated tissue culture dishes.Contaminating neurons were completely eliminated after two rounds of trypsinization and passaging.More than95%of cells in these cultures were immunopositiveŽ.Žfor the glial fibrillary acidic protein GFAP Yong et al., .1992.2.3.Simulated inÕitro ischemiaThe term‘simulated in vitro ischemia’is used in this study to describe an in vitro model of combined severe hypoxia and glucose and nutrient deprivation that lacks a blood flow component of in vivo ischemia.The simulated in vitro ischemia was induced as described previously Ž.Stanimirovic et al.,1997a.HCEC and FHAS grown in 35-mm dishes were washed twice in phosphate-buffered Ž.saline PBS,and then subjected to a combination ofwŽ.glucose-free Krebs solution containing in mM119NaCl,x 4.7KCl,1.2KH PO,25NaHCO,2.5CaCl,1MgCl24322Ž.and severe hypoxia-2%oxygen in an anaerobic cham-Ž. ber Anaerobic System model1024,Forma Scientific equipped with a humidified,temperature-controlled incu-bator directly accessible within the chamber.The entire system was purged with95%N r5%CO atmosphere.22Ž.Recovery i.e.,simulated reperfusion was achieved byŽ.exposing cells to ambient air reoxygenation and replac-ing media with Krebs buffer containing5mM glucose. The culture media and cells were simultaneously harvested at the end of a4-h simulated in vitro ischemia and at4,16, and24h of recovery.Respective controls were maintained in a20%oxygen and Krebs buffer supplemented with 5-mM glucose for matching periods of time.In parallel experiments,both HCEC and FHAS were also exposed to various concentrations of the pro-in-Ž. flammatory cytokine,IL-1b50–100u r ml,and media and cells were harvested at4,8,20,and28h after IL-1b addition.2.4.RT-PCRTotal RNA was extracted from HCEC and FHAS grownŽ.in35-mm Petri dishes using Tri reagent500m l r dish and subsequently purified using the protocol provided by the manufacturer.The RNA pellets were resuspended in50m lof DEPC-treated dH O and incubated at558C for10min.2The quality of the RNA was confirmed for each sampleŽusing formaldehyde-agarose gel electrophoresis.RNA0.5 .Ž.m g was then mixed with0.5m g of oligo dT12–18 primers,heated for10min at708C,and then chilled on ice.Synthesis of single-stranded cDNA was performed byŽ.reverse transcription428C,1h in a reaction mixture Ž.final volume20m l containing4m l of5=first strand Ž.buffer Gibco BRL,2m l of0.1M DTT,1m l of10mMyŽdNTP,100units of RNase H reverse transcriptase Su-.perScript e II,Gibco BRL,and10m l of DEPC-treateddH0.The reverse transcriptase was inactivated by heating 2the reaction mixture at708C for15min.The20-m l of the resulting reaction mixture was diluted with20m l ofdH O,and4m l of the cDNA from this mixture was 2subsequently used in a50-m l PCR reaction.Specific primers were designed according to published sequences of the human chemokines IL-8,MCP-1andneurotactin,and housekeeping genes,b-actin and b-mi-2 croglobulin,and were synthesized using a PerSeptive Ž.Biosystem Framingham,MA synthesizer.Sequences of the primers used are shown in Table1.PCR amplifications were carried out in a final volumeŽ.of50m l containing1=reaction buffer Promega,1.5mM MgCl,0.2mM each of four dNTPs,0.4m M each of 2Žtwo pairs of the primers primers for an internal control gene and the primers for either IL-8,MCP-1or neuro-.tactin,2.5units Taq DNA polymerase,and4m l cDNA. The amplification reactions were performed in a PTC-200Table1Sense and anti-sense sequences of primers for the human chemokines,IL-8,MCP-1,and neurotactin,and for the housekeeping genes,b-actin and Ž.b2-microglobulin b2-M used in PCR reactions in this study.The RT-PCR reactions using the paired primers listed in the table generate a289-bp DNA fragment for IL-8,a257-bp fragment for MCP-1,a943-bp fragment for neurotactin,a504-bp fragment for b-actin,and a137-bp fragment for b2-microglobulin,respectivelyX XŽ.Ž.Gene5Primer sense primer3Primer anti-sense primerX X X XIL-85-ATGACTTCCAAGCTGGCCGTG-35-CTCCACAACCCTCTGCACCCA-3X X X XMCP-15-GCTCGCTCAGCCAGATGCAAT-35-TGGGTTGTGGAGTGAGTGTTC-3X X X XNeurotactin5-GATACCTGTAGCTTTGCTCATCC-35-TGGTAGGTGAACATGGCCACC-3X X X Xb-Actin5-GACTATGACTTAGTTGCGTTA-35-GCCTTCATACATCTCAAGTTG-3X X X Xb2-M5-CAGCAAGGACTGGTCTTTCTAC-35-GCTACCTGTGGAGCAACCTGC-3()W.Zhang et al.r Journal of Neuroimmunology1011999148–160151Ž. Peltier Thermal Cycler MJ Research,Watertown,MA in Ž.0.2-ml tubes Rose Scientific,Alberta,Canada.All ampli-fications were done using a denaturation step at948C for 30s,annealing step at558C for45s and polymerization step at728C for40s,and were carried out for45cycles.Ž.Aliquots of the PCR10m l and restriction digests were subjected to electrophoresis on a1.5%agarose in a Tris-borate buffer containing0.5m g r ml ethidium bromide,and then photographed.The internal control genes and chemokine genes were linearly amplified during the45 PCR cycle.The relative densities r volumes of the bands on film negatives were measured using a Computing Densito-Ž.Žmeter model300A Molecular Dynamics,Sunnyvale, .CA and analyzed using ImageQnaNT,version4.1soft-Ž.ware Molecular Dynamics.2.5.ImmunocytochemistrySurface expression of ICAM-1in HCEC was assessed by indirect immunocytochemistry.Briefly,HCEC grown on10m g r ml human fibronectin-coated glass coverslips were washed with PBS and incubated at room temperature with2m g r ml mouse hybridoma monoclonal antibody Ž.Ž.IgG to human ICAM-1clone CD54for40min in medium M199.Cells were then washed in PBS and incu-bated for1h with a1:50dilution of a4-nm colloidal gold particle-conjugated goat anti-mouse IgG antibody.Cover-slips were then fixed for30s with9.25%formaldehyde and45%acetone in PBS,washed,and incubated in a silverŽ.enhancing solution IntenSE e M for25min.Coverslips were washed and counterstained with Giemsa stain for15 min.Non-specific staining was determined by either the exclusion of the primary antibody,or by replacing primary anti-ICAM-1antibody with the isotype-matched antibody Ž.to GFAP Accurate Chemical and Scientific.2.6.ELISAŽ.An enzyme-linked immunosorbent assay ELISA was used to quantify the levels of ICAM-1expressed in HCEC and the levels of IL-8and MCP-1in culture media of HCEC and FHAS.For ICAM-1measurements,HCECŽ4 cultures grown in96-well microtiter plates2=10.cells r100m l per well at378C in5%CO were sequen-2tially incubated with a primary monoclonal anti-humanŽ.ICAM-1antibody2m g r ml for1h at378C,followed by 1:500diluted peroxidase-conjugated goat anti-mouse IgG in PBS for45min at378C.Non-specific binding sites wereŽ.blocked with2%bovine serum albumin BSA in PBS for 30min at378C.After each incubation,the plates were washed three times with PBS.Color was developed by the addition of100-m l of a1mg r ml solution of the horse-XŽradish peroxidase substrate,2,2-azinobis3-ethylbenz-.Žthiazoline-6-sulfonic acid diammonium salt ImmunoPure.ABTS.The reaction was stopped after5min by the addition of an equal volume of1%SDS to each well.TheŽ.optical density O.D.of the developed color was read atŽ405nm using a SpectraMAX Molecular Devices,Menlo .Park,CA microplate reader.The levels of IL-8and MCP-1released from variously treated HCEC and FHAS were quantified by a‘sandwich’ELISA using commercially obtained ELISA kits,Quan-Ž. tikine human IL-8kit R&D System,Minneapolis,MNŽand ID Elisa e MCP-1kit ID Labs Biotechnology,Lon-.Ž5. don,ON.Either HCEC or FHAS2=10cells r dish grown in35-mm dishes were used in these experiments. Aliquots of culture media were collected,centrifuged at 14,000rpm for5min at48C,and the ELISA assays were carried out as instructed by the manufacturers.2.7.Neutrophil adhesion and chemotaxisHuman neutrophils were isolated from fresh,EDTA-treated venous blood obtained from adult healthy volun-Žteers medical staff at Children’s Hospital of Eastern On-.tario.The neutrophil-containing band was separated onŽdiscontinuous polysucrose–sodium diatrizoate Histo-.paque-1077and-1119gradients as described by EnglishŽ.and Andersen1974.Contaminating erythrocytes were removed by brief hypotonic lysis in ice-cold0.15%NaCl. Microscopic examination and acetic acid–Cresyl violet staining indicated that more than98%of the remaining cells were neutrophils.The neutrophils were then labeledŽ. with10m M calcein-AM Molecular Probe,OR,USA in PBS for20min at378C.The cells were centrifuged at room temperature for5min at1400rpm,washed twice with PBS,and re-suspended in PBS at a final density of 2.5=106cells r ml.Confluent monolayers of control or treated HCEC grown in96-well microtiter plates were washed twice with PBSŽ4.and calcein-labeled neutrophils5=10cells r100ul were then added to each well for1h at378C.Non-adherent neutrophils were removed by two rapid washes with100 m l PBS and the intensity of the fluorescence remaining in each well was determined using a CytoFluor2350fluores-Ž.cence microplate reader Millipore,Bedford,MA equipped with485r22nm excitation and530r25nm emission filters.The numbers of neutrophils adhering to the HCEC monolayers were calculated using a standard curve gener-Žated from the fluorescence values of increasing1000–.100,000numbers of labeled neutrophils in suspension. Alternatively,variously treated HCEC monolayers were pre-labeled with the fluorescent dye,3,3X-dipropylthiadi-ŽŽ..Ž. carbocyanine iodide DiSC51m M,20min prior to3the addition of calcein-labeled neutrophils and fluorescentŽimages were generated using a Zeiss LSM410Carl Zeiss,.Thornwood,NY inverted laser scanning microscope Ž.LSM equipped with an Argon r Krypton ion laser and a Zeiss LD achroplan20=,0.4NA objective.The images()W.Zhang et al.r Journal of Neuroimmunology1011999148–160 152were generated using the488-nm and the568-nm excita-tion lines with a FT655dichroic mirror and a FT560 beam splitter positioned in the light path.Emitted fluores-Ž.cence of both calcein and DiSC5was collected simulta-3neously after being filtered through a510–525nm band-pass filter and a610-nm long-pass filter,respectively. Confocal apertures for each recorded wavelength were adjusted to a full width half maximum of20m m.Standard image processing was performed to enhance brightnessŽand contrast using ImageSpace software Molecular Dy-.namics.Chemotaxis of calcein-AM labeled neutrophils induced by media collected from variously treated HCEC and FHAS,or known chemoattractants,in the presence or absence of the anti-IL-8antibody,was assessed by a quantitative in vitro method using a ChemoTx a101-5Ž.Neuro Probe,Gaithersburg,MD assembly consisting of a 96-well plate and a polycarbonate filter membrane,asŽ.described by Junger et al.1993.Briefly,the wells of the plates were loaded with media collected from variously treated cells or solutions containing known concentrations of chemoattractants,framed filter was positioned on top, 50,000neutrophils suspended in20m l of matching non-conditioned media were applied on the top of each mem-brane r well,and the assembly was incubated for90min at 378C.The numbers of neutrophils transmigrated into wells of the96-well plate were quantified by measuring intensityŽ.of fluorescence excitation r emission:485r530nm in aŽ.CytoFluor2350reader Millipore,against the standard curve constructed with increasing numbers of labeled neu-trophils.3.Results3.1.Effects of simulated inÕitro ischemia on the expres-sion of ICAM-1in HCEC and the adhesion of allogenic neutrophils to HCECŽ.Simulated in vitro ischemia4h followed by a24-h recovery resulted in an up-regulation of ICAM-1in HCECŽ. as determined by immunocytochemistry Fig.1A and B Ž.and ELISA Fig.1G.The same duration of ischemia r re-covery effected a2–3-fold increase in a number of allo-genic neutrophils firmly adhering to HCEC monolayers asŽobserved r quantified by confocal microscopy Fig.1D and .Ž.E and fluorometry Fig.1H.The exposure of HCEC toŽ.IL-1b100u r ml for24h caused a dramatic up-regu-Ž.lation of ICAM-1Fig.1C and G,and an increase in theŽ. number of neutrophils3–4-fold above control adheringŽ.to IL-1b-pretreated HCEC Fig.1F and H.We have previously reported that both cytokines and in vitro-ischemia stimulate the transcription of ICAM-1in HCEC Ž.Stanimirovic et al.,1997a.In both cases,increased neu-trophil adhesion to HCEC was blocked by a combinationŽof anti-ICAM-1and anti-CD18antibodies Stanimirovic et .al.,1997a.3.2.Effects of simulated inÕitro ischemia and IL-1b on the expression r secretion of chemokines in HCEC and FHAS3.2.1.The expression of neurotactin in HCEC and FHASThe low levels of expression of the recently discovered,brain-enriched neutrophil chemoattractant with a CX C3Ž.motif Pan et al.,1997,neurotactin,was detected by RT-PCR in the whole human brain homogenate,as well as in HCEC,FHAS,and human umbilical vein endothelial Ž.Ž.cells HUVEC in culture data not shown.Neurotactin mRNA expression in HCEC and FHAS was not affected by either of the following stimuli:in vitro ischemia r re-Ž.covery,IL-1b50–100u r ml;4–24h,and phorbol ester Ž.Ž.20–40nM;6h data not shown,whereas a transient induction of neurotactin mRNA was detected in HUVECŽ. stimulated with100u r ml IL-1b for8h data not shown. Increased expression of fractalkine r neurotactin mRNA by IL-1b and TNF a has previously been demonstrated inŽ. peripheral human endothelial cells Bazan et al.,1997, and brain microglia in mice with experimental autoim-Ž.mune encephalomyelitis Pan et al.,1997.These experi-ments suggested that neurotactin is likely not critically involved in ischemia-induced HCEC r neutrophil interac-Ž.tion s in vitro.3.2.2.Effects of simulated inÕitro ischemia on the expres-sion and secretion of IL-8in HCEC and FHAS In vitro ischemia resulted in a transient up-regulation ofŽ.IL-8mRNA in HCEC at4h of recovery Fig.2A.IL-8 mRNA returned back to control levels at16and24h of Ž.recovery Fig.2A.The levels of immunoreactive IL-8 released into HCEC media increased as early as4h after Ž.ischemia Fig.2B and continued to accumulate in cultureŽ.media up to16h of recovery Fig.2B.In FHAS,IL-8mRNA was already up-regulated at the end of a4-h ischemia and throughout the period of post-Ž.ischemic recovery up to24h Fig.3A.The maximalŽ.up-regulation was seen at4-h post-ischemia Fig.3A.The increased mRNA expression was accompanied by up to 3-fold increases in the levels of immunoreactive IL-8in Ž.FHAS media Fig.3B.The expression of two housekeeping genes,b-actin andŽ.b-microglobulin Table1,used alternatively in various 2experiments,was not affected by ischemia r recovery pro-Žtocols in either HCEC or FHAS Fig.2A and,Figs..3–5A.3.2.3.Effects of simulated inÕitro ischemia on the expres-sion and secretion of MCP-1in HCEC and FHAS The MCP-1mRNA was up-regulated at4,16and24hŽ. after transient in vitro ischemia in both HCEC Fig.4A()W.Zhang et al.r Journal of Neuroimmunology1011999148–160153Ž.Fig.1.Effects of simulated in vitro ischemia r reperfusion or IL-1b on the expression of ICAM-1in HCEC A–C,G and the adhesion of homologous Ž.Ž.Ž. neutrophils to HCEC monolayers D–F,H.ICAM-1expression was determined by immunocytochemistry panels A–C and by ELISA G,andŽ.Ž.Ž.neutrophil adhesion was assessed by fluorescence confocal microscopy panels D–F and fluorometry H as described in Section2.A basal expression of ICAM-1in HCEC.Insert micrograph shows a non-specific staining in which primary anti-ICAM-1antibody was replaced with the isotype-matched Ž.Ž.anti-GFAP antibody;D basal level of neutrophil adhesion to control HCEC monolayer;B and E HCEC were exposed to4h ischemia r24h recovery;Ž.C and F HCEC were exposed to100u r ml IL-1b for24h.The images shown are representative of at least five independent experiments.Each bar in G and H represents the mean"S.D.of six replicate wells in one of at least three experiments yielding similar results.Asterisks indicate significant wŽ.xdifferences p-0.01;ANOVA followed by the Fisher’s protected least square difference PLSD multiple comparisons compared to control levels.a Ž.Ž.Indicates significant difference p-0.01;ANOVA compared to in vitro ischemia I4r R24.()W.Zhang et al.r Journal of Neuroimmunology 1011999148–160154Fig.2.Effects of simulated in vitro ischemia r reperfusion on IL-8mRNA Ž.expression in HCEC A and the secretion of immunoreactive IL-8in Ž.Ž5.HCEC media B .Confluent HCEC ;2=10cells r dish were exposed to 4-h simulated in vitro ischemia and the indicated periods of recovery as described in Section 2.IL-8mRNA expression was determined by a Ž.Ž.semi-quantitative RT-PCR A .Volumes of IL-8bands closed bars were Ž.Ž.expressed as a percentage of control gene b -actin bands open bars amplified in same PCR reactions.IL-8secretion in cell media of control Ž.Ž.Ž.open bars and ischemic cells closed bars was quantified by ELISA B in same experiments.Each bar represents the mean "S.D.of four sepa-rate gels and ELISA determinations,each performed in duplicate.Aster-w isks indicate a significant difference p -0.01;ANOVA followed by x Fisher’s PLSD multiple comparisons from corresponding control levels.Ž.and FHAS Fig.5A .MCP-1levels in HCEC media increased initially at 4h of recovery and reached a 3-fold increase above the corresponding controls at 16and 24h Ž.after ischemia Fig.4B .In FHAS,immunoreactive MCP-1levels were already elevated at the end of ischemia and were 2–2.5-fold higher than corresponding control levels Ž.at all times of recovery Fig.5B .Both basal and ischemia-stimulated levels of immunore-active MCP-1released into FHAS media were 20–30Žtimes higher than those released by HCEC i.e.,ng r ml vs..Ž.pg r ml Fig.4B and,Fig.5B .3.2.4.Effects of IL-1b on the expression and secretion of IL-8and MCP-1in HCEC and FHASIL-1b strongly up-regulated the expression of IL-8mRNA and MCP-1mRNA in both HCEC and FHAS at4,Fig.3.Effects of simulated in vitro ischemia r reperfusion on the expres-Ž.sion of IL-8mRNA in FHAS A and the secretion of immunoreactive Ž.Ž5.IL-8in FHAS media B .Confluent FHAS ;2=10cells r dish were exposed to 4-h simulated in vitro ischemia and the indicated periods of recovery as described in Section 2.IL-8mRNA expression was deter-Ž.Žmined by semi-quantitative RT-PCR A .Volumes of IL-8bands closed .Ž.bars were expressed as a percentage of control gene b -actin bands Ž.open bars amplified in same PCR reactions.IL-8secretion in cell media Ž.Ž.of control open bars and ischemic cells closed bars was quantified by Ž.ELISA B in same experiments.Each bar represents the mean "S.D.of four separate gels r experiments.Asterisks indicate a significant difference Žp -0.01;ANOVA followed by the Fisher’s PLSD multiple compar-.isons from corresponding control levels.。