Ž .Brain Research 821 1999 169–176

Research report

Functional activation of EphA5 receptor does not promote cell proliferationin the aberrant EphA5 expressing human glioblastoma U-118 MG cell line

Victoria Bruce a, Gianfranco Olivieri a, Oliver Eickelberg b,1, Guido C. Miescher a,)

a Neurobiology Laboratory, Department of Research, UniÕersity Hospital Basel, CH-4031 Basel, Switzerlandb Pneumology Laboratory, Department of Research, UniÕersity Hospital Basel, CH-4031 Basel, Switzerland

Accepted 29 December 1998

Abstract

Ž .Eph receptors are a subfamily of receptor tyrosine kinases RTKs , that are activated by ephrin ligands and appear to play importantroles in axon guidance and cell migration during development of the nervous system. Over-expression or constitutive activation of Ephreceptors has been linked with increased proliferation in various tumours. We have recently described lineage aberrant expression ofEphA5 in primary human astrocytomas, glioblastomas and in the human glioblastoma U-118 MG cell line. A role for EphA5 expressionin these tumours is not apparent, and we have investigated the cellular effects of EphA5 activation using the human glioblastoma U-118MG cell line as a model. Immunofluorescent staining demonstrated cell surface expression of EphA5. Activation of the EphA5 receptorusing an ephrin-A1 recombinant fusion protein resulted in tyrosine phosphorylation of EphA5 in a time-dependent manner. Exposure ofU-118 MG glioblastoma cells to ephrin-A1 did not result in significant spontaneous or FCS-stimulated cell proliferation, though amarginal decrease was observed. This is in converse to the effects of Eph activation in other tumour cell lines, and is the first study toinvestigate EphA5 in glioblastoma cell lines. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: Receptor tyrosine kinase; Eph receptor; Glioblastoma cell line; Cell proliferation

1. Introduction

Ž .Receptor tyrosine kinases RTKs are known to beimportant regulators of cellular growth controlling cell

w xproliferation, differentiation and migration 26 . These cellactions are particularly prominent during vertebrate devel-opment, and overexpression or mutations in RTKs can leadto uncontrolled growth, resulting in oncogenesis. A sub-family of RTKs known as Eph receptors, and their cognateephrin ligands, appear to be involved during development

Ž .of the central nervous system CNS in attaining thenormal segregation of structurally defined regions by guid-ing the pathfinding of cells and axons, for review see

w xPasquale 22 . Interestingly, aberrant expression androrconstitutive activation of Eph receptors, compared to thecorresponding normal tissue, has been found in tumour cell

) Corresponding author. Biochemistry Institute, University of Fribourg,Rue de Musee 5, CH-1700 Fribourg, Switzerland. Fax: q41-26-300-9735;´E-mail: guido.miescher@unifr.ch

1 Current address: Department of Pathology, Yale University School ofMedicine, 310 Cedar Street, New Haven, CT 06520-8023, USA.

lines from the CNS and primary epithelial tumours, therebyimplicating Eph involvement in uncontrolled growth and

w xtumourigenesis 2,7,12,25,28The Eph receptors represent the largest subfamily of

w xreceptor tyrosine kinases 30 . Their characteristic featuresinclude a conserved cysteine-rich region and two fi-bronectin type III repeats in their extracellular domain anda cytoplasmic tyrosine kinase domain that is closely re-

w xlated to src-like intracellular tyrosine kinases 16 . EphAreceptors are activated by membrane-bound glycosylphos-

Ž . Žphatidylinositol GPI -linked ligands termed ephrin-A lig-.ands and EphB members are activated by

Ž .transmembrane-bound ligands termed ephrin-B ligandsw x1 . In contrast to other subfamilies of RTKs, which aregenerally activated by soluble ligands, most of the knowninteractions of Eph RTKs involve membrane-bound lig-

w xands 5 . However, soluble ligands presented as ligandfusion proteins that consist of dimeric ligand fused to theFc portion of human IgG1 can activate Eph receptors, andactivation of several Eph receptors may be enhanced byclustering of the ligand fusion protein with anti-human

w xIgG1 antibody 5 .

0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0006-8993 99 01112-9

( )V. Bruce et al.rBrain Research 821 1999 169–176170

The role of Eph receptors and ephrin ligands has beenlargely studied during the development of the nervoussystem. Their expression is often mutually exclusive and itseems that Eph receptor and ephrin ligand interactionsmediate inhibitory signals, resulting in a repulsion in the

w xdirection of migratory cells 13,27,31,33 . Eph and ephrininteractions also result in inhibitory signalling in axonal

w xguidance, by regulating axon fasciculation 18,32 or axonw xgrowth cone collapse 6,9 .

Studies investigating Eph activation and effects on cellproliferation have revealed contrasting results. Stimulationof a chimeric receptor composed of an EGF extracellulardomain and either EphA3 or EphB1 intracellular domainexpressed in NIH 3T3 cells, resulted in tyrosine phospho-rylation but not focus formation or a significant effect on

w xDNA synthesis 3,15 . However, ephrin-A1 exposure of anadenocarcinoma cell line expressing EphA2, and amelanoma cell line found to aberrantly express EphA2,lead to an increase in cell proliferation, thus supporting anoncogenic effect of Eph activation in these tumour cell

w xlines 7,25 . In addition, a study investigating aberrantEphA4 expression in mammary tumours of transgenicmice overexpressing the Ha-ras oncogene correlated in-creased EphA4 expression with increased invasive andundifferentiated tumour properties: no elevated EphA4 ex-pression was found in well differentiated non-metastatic

w xmammary tumours 2 .We have previously reported EphA5 expression in neu-

rons of the developing CNS in the rat and in human adultw xCNS 19,29 . Normal adult glia do not have detectable

EphA5 protein, however we have identified lineage aber-rant EphA5 expression in several primary human gliomasirrespective of their histological grading, and in the human

w xglioblastoma U-118 MG cell line 19 . A possible onco-genic role of EphA5 is not known and has not previouslybeen assessed in a glioblastoma cell line. To investigate acell proliferative effect of EphA5 activation we used thehuman glioblastoma U-118 MG cell line, and show thatEphA5 is expressed at the cell surface of U-118 MG cells.Activation of the EphA5 receptor in a time-dependentmanner resulted in tyrosine phosphorylation of EphA5.Ephrin-A1-Fc stimulation of U-118 MG cells resulted in amarginal decrease in cell proliferation, suggesting aberrantexpression of EphA5 is not implicated in promotion of cellproliferation in this cell line.

2. Materials and methods

The human glioblastoma U-118 MG cell line was cul-tured in Minimum Essential Medium with Earle’s salts andL-glutamine, supplemented with 10% foetal calf serumŽ .FCS and 1% non-essential amino acids. No antibiotics orantimycotics were added to the culture medium at anytime. Cells were cultivated in a humidified airr5% CO2

chamber at 378C.

2.1. Immunocytochemistry

Cells were cultured to 80% confluence and transferredto 96-well plates for living cell staining. Briefly, trypsinisedcells were blocked by washing =2 with phosphate-buffered

Ž .saline solution PBS containing 1% bovine serum albuminand 0.05% NaN , prior to centrifugation at 400=g for 23

min. All incubations were performed for 30 min on iceusing the same buffer and were followed by the samewashing procedure. The cells were incubated with 5 mgrmlof our affinity purified polyclonal rabbit antibody specific

w xfor the extracellular fibronectin IIb domain of EphA5 29Žor 5 mgrml ephrin-A1-Fc composed of ephrin-A1 fused

.to the Fc portion of human IgG1 or as a control withŽpolyclonal ephrin-A1 antibody Santa Cruz Biotechnology,

.CA, US . Incubation with Dichlorotriazinyl Amino Fluo-Ž .rescin-labelled anti-rabbit IgG antibody 20 mgrml or

ŽPhycoerythrin-labelled anti-human IgG antibody 20.mgrml was performed in darkness. Incubation with the

fluorescent secondary antibodies alone at 20 mgrml servedas controls. Cells were resuspended in 200 ml buffer andcentrifuged onto precoated gelatine–chroamalaun glassslides at 35=g for 1 min using a Shandon cytospin funnelŽ .Instrumenten Gesellschaft, Zurich, Switzerland .

2.2. Immunoprecipitation and immunoblotting

Cells were cultured to 80% confluence in 20 cm culturedishes, and maintained for 24 h in medium containing0.1% FCS in order to prevent non-specific stimulationprior to exposure. Cells were exposed to 5 mgrml ofephrin-A1-Fc. Cells cultured throughout in 10% FCSmedium served as a negative control. In separate experi-ments, ephrin-A1-Fc was clustered by mixing with ratanti-human IgG1 antibody on a rotating wheel for 30 min

Ž . w xat room temperature RT as described previously 5 .After ligand exposure, cells were washed 2= with ice-cold

ŽPBS and lysed in 1 ml lysis buffer 50 mM Tris–HCl pH8.0 containing 150 mM NaCl, 10 mM EDTA, 1% NP 40,

.0.02% NaN freshly supplemented with protease in-3Ž .hibitors Completee, Boehringer Mannheim and 1 mMŽ .Na VO Sigma . The cell lysates were further ho-3 4

mogenised by syringing =20 through a 25 G needleŽ .Terumo, Belgium and rotation for 1 h at 48C. To removecell debris, total cell lysates were centrifuged at 10,000=gfor 15 min at 48C. To immunoprecipitate, each sample wassplit into equal parts. One part was incubated with 10 mgof our antibody against the extracellular domain of EphA5,and the other part incubated with a control non-specificantibody. Samples were placed on a rotating wheel for 12h at 48C. Protein A sepharose 6MB beads were added,samples incubated for 1 h at 48C and pelleted by centrifu-gation at 1250=g for 10 min. Immunoprecipitates were

Žwashed three times with buffer 50 mM Tris–HCl pH 8.0.containing 150 mM NaCl, 0.1% NP40, 0.02% NaN3

Žfreshly supplemented with protease inhibitors Completee,

( )V. Bruce et al.rBrain Research 821 1999 169–176 171

. Ž .Boehringer Mannheim and 1 mM Na VO Sigma , mixed3 4

with SDS-gel sample buffer containing 20% Dithiothreitol,Ž .boiled 5 min and electrophoresed in 7.5% SDS poly-

w xacrylamide gels 14 . Proteins were transferred onto nitro-Žcellulose membranes Schleicher and Schuell, Keene, NH,

.USA by semi-dry electro blotting. The blots were incu-Žbated in blocking solution 10 mM Tris–HCl pH 8.0

.containing 150 mM NaCl, 0.05% Tween 20 supplementedŽ .with 5% skim milk powder Fluka, Switzerland for 1 h at

RT. Blots were probed overnight at 48C with either 0.33mgrml 4G10 monoclonal anti-phosphotyrosine antibodyŽ .Upstate Biotechnology, NY, USA or 0.25 mgrml affinitypurified polyclonal rabbit antibody specific for the kinase

w xdomain of EphA5 29 . Following incubation for 2 h atroom temperature with either goat anti-rabbit or goat anti-mouse horseradish peroxidase-labelled secondary antibod-ies, diluted according to the manufacturer’s instructionsŽ .Sigma , the blots were processed for chemiluminescenceŽ .ECL, Amersham, UK . In order to reprobe blots with analternative antibody, membranes were incubated for 30

Žmin at RT in stripping buffer 0.1 M glycine pH 2.5 inŽ . .double distilled dd H O and re-blocked. Stripped blots2

were probed with secondary-labelled antibody alone toverify that no residual signal was left.

[3 ]2.3. H -thymidine incorporation assay

De novo synthesis of DNA was determined by thew3 x w xincorporation of H -thymidine, as described 4 . U-118

MG glioblastoma cells were cultured to 70% confluence in48-well plates and maintained for 24 h in medium contain-ing 0.1% FCS in order to prevent non-specific stimulationprior to exposure. At each independent FCS medium con-

Ž .centration 0.1, 1, 5 or 10% FCS cells were exposed to0.05, 0.5 or 5 mgrml of ephrin-A1-Fc, or as a positivecontrol to 1, 10 or 20 ngrml platelet-derived growth factorŽ .PDGF-BB . Additional controls were made with cell ex-posure to the same concentrations of an analogous Fcdomain alone or EphA5-Fc fusion protein. Cells culturedthroughout in 0.1, 1, 5 or 10% FCS medium served asnegative controls. Stimulations were repeated in quadrupli-cate. In separate experiments, cells were exposed toephrin-A1-Fc further clustered by mixing with rat anti-hu-

Ž .man IgG see immunoprecipitation and immunoblotting .w3 x5 h after ephrin-A1-Fc exposure 1 mCirml H -thymidine

w3 xwas added and 30 h after exposure incorporation of H -thymidine determined. The cells were washed =2 withice-cold PBS and fixed with 100% methanol for 5 min.The cells were washed =2 in dd H O, followed by2

hydrolysis of single stranded nucleotides by incubationwith 0.3 M Trichloroacetic acid for 1 min. The cells werewashed =2 with dd H O, lysed with 0.5 M NaOH and2

cell culture plates placed on a shaker for 30 min. Eachsample was pipetted into a scintillation vial, scintillation

Ž .liquid Quickszint 212, Zinsser Analytic, UK added andcounts determined using a Liquid Scintillation Analyzer

Ž .Canberra-Packard, Zurich, Switzerland . Statistical signif-icance was tested using the Mann Whitney U Test.

3. Results

3.1. Cell surface expression of EphA5 in the humanglioblastoma U-118 MG cell line

To investigate EphA5 expression on the cell surface weincubated living cells with either our affinity purifiedpolyclonal antibody specific for the extracellular domainof EphA5 or ephrin-A1-Fc, and visualised receptor expres-sion by incubation with Dichlorotriazinyl Amino Fluo-rescin-labelled anti-rabbit IgG antibody or Phycoerythrinfluorescent labelled secondary anti-human IgG1 antibody.The fluorescent staining shows strong endogenous expres-sion of EphA5 on the cell surface of the U-118 MG

Ž .glioblastoma cell line Fig. 1a and b . Incubation of cellswith either secondary antibody alone gave no fluorescent

Žsignal staining with Dichlorotriazinyl Amino Fluorescin-.labelled anti-rabbit IgG antibody shown in Fig. 1c . To

determine endogenous ephrin-A1 expression, cells wereincubated with ephrin-A1 antibody, however no staining

Ž .was observed data not shown .

3.2. Phosphorylation of EphA5

We investigated EphA5 phosphorylation in response toephrin-A1-Fc exposure, by immunoprecipitation of EphA5from ephrin-A1-Fc exposed cell lysates using our antibodyspecific for the extracellular domain of EphA5, and im-

Žmunoblotting with an anti-phosphotyrosine antibody Fig..2 . Exposure to ephrin-A1-Fc resulted in a clear time-de-

pendent phosphorylation of EphA5. No detectable phos-phorylated EphA5 could be immunoprecipitated with lessthan 10 min exposure to ephrin-A1-Fc. However, after 10min ephrin-A1-Fc exposure, a phosphorylated protein atapproximately 120 kD was seen, the same size as the

w xknown molecular weight of EphA5 10,17,19 . Phosphory-Žlation peaked at 15 min ephrin-A1-Fc exposure Fig. 2a,

.lanes 3 and 5 . Ephrin-A1-Fc exposure for more than 15Ž .min did not show any phospho-proteins data not shown .

Immunoprecipitation with the control non-specific anti-body did not result in any immunoprecipitated proteins.Western blot analysis with anti-phosphotyrosine antibodyof unexposed cell lysates demonstrated the presence of

Ž .constitutively phosphorylated proteins Fig. 2a, lane CT .However, no constitutive phosphorylation of EphA5 was

Ž .found Fig. 2a, lane 1 . The blot was stripped and reprobedwith antibody against the EphA5 catalytic domain whichdemonstrated the presence of unphosphorylated EphA5 inunstimulated cells and confirmed the identity of the im-

Žmunoprecipitated 120 kD phosphoprotein as EphA5 Fig..2b . Separate experiments were performed to determine

any enhancement of EphA5 phosphorylation by stimula-

( )V. Bruce et al.rBrain Research 821 1999 169–176172

( )V. Bruce et al.rBrain Research 821 1999 169–176 173

Fig. 2. Phosphorylation of EphA5 in ephrin-A1-Fc exposed U-118 MG glioblastoma cells. Constitutively phosphorylated proteins are shown in WesternŽ . Ž .analysis of unexposed cell lysates a, lane CT , however no constitutive phosphorylation of EphA5 was found in EphA5 immunoprecipitates a, lane 1 . At

Ž .10 min exposure to ephrin-A1-Fc, phosphorylated EphA5 could be immunoprecipitated a, lane 3 and phosphorylation peaked at 15 min ephrin-A1-FcŽ . Ž . Ž .exposure a, lane 5 . b Characterization of immunoprecipitates using specific anti-EphA5 antibody. The blot from a was stripped and reprobed with an

antibody specific for the EphA5 catalytic domain. Only lanes containing immunoprecipitates showed bands, at approximately 120 kD, the knownŽ . Ž .molecular weight of EphA5 indicated with arrow . Rabbit heavy chains H are present in all lanes. Phosphorylated proteins in the supernatants can be

Ž . Ž .observed at each time point of ephrin-A1-Fc exposure a, lanes 2, 4 and 6 , however they did not react with the anti-EphA5 antibody b, lanes 2, 4 and 6 .

tion with ephrin-A1-Fc further clustered by rat anti-humanŽ .IgG1, however no enhancement was seen data not shown .

3.3. Marginal decrease in glioblastoma cell proliferationrate upon EphA5 actiÕation

To investigate a possible mitotic effect of EphA5 acti-vation we investigated cell proliferation of U-118 MGglioblastoma cells with exposure to 0.05, 0.5 or 5 mgrmlephrin-A1-Fc. Cell proliferation was investigated by assay-

w3 xing de novo DNA synthesis by determination of H -thymidine incorporation. Four independent FCS mediumconcentrations were used to exclude possible FCS-specific

effects. U-118 MG glioblastoma cell exposure to ephrin-A1-Fc resulted in a dose-dependent decrease in cell prolif-

Ž .eration at all concentrations of FCS tested Fig. 3 , how-ever this was not statistically significant.

PDGF is known to be a growth factor for many celltypes and as the U-118 MG glioblastoma cell line haspreviously been shown to express alpha and beta forms of

w xthe PDGF receptor 20 , PDGF-BB was used as a positivecontrol for cell proliferation assay. Exposure to 1, 10 or 20ngrml PDGF-BB in 0.1% FCS medium resulted in arespective dose-dependent increase in cell proliferation:18 466"S.D. 671; 20 853"S.D. 755, p-0.05 and21 929"S.D. 1158, p-0.05; compared to 0.1% FCS

Fig. 1. Immunofluorescent localisation of EphA5 on the cell surface of the U-118 MG glioblastoma cell line. U-118 MG cells were incubated withŽ . Ž .antibody specific for the extracellular domain of EphA5 a or ephrin-A1-Fc b and fluorescence visualised by further incubation with the appropriate

Ž .fluorescent secondary antibody. As a control cells were incubated with fluorescent secondary antibody alone c . The fluorescent staining depicts clustersof speckles outlining the cell circumference, demonstrating expression of EphA5 on U-118 MG glioblastoma cell surface.

( )V. Bruce et al.rBrain Research 821 1999 169–176174

Fig. 3. Cell proliferation of U-118 MG glioblastoma cells following exposure to ephrin-A1-Fc or Fc alone. At each FCS medium concentration tested,exposure to ephrin-A1-Fc resulted in a dose-dependent marginal decrease in cell proliferation, compared to respective FCS only controls. Exposure to Fc

Ž .alone indicated by unfilled symbols at the same concentrations did not alter cell proliferation compared to FCS control. As seen, the optimal growthŽ .conditions for the U-118 MG glioblastoma cell line are between 5% and 10% FCS medium concentration. Each count per minute cpm reading was

assayed in quadruplicate, and values on graph indicate means. Error bars indicate S.E.M. Similar results were obtained in seven independent experiments;the results of a representative experiment are shown.

only control value of 18 495"S.D. 1469. Each value isthe mean of quadruplicate samples and similar results wereobtained in four independent experiments. Exposure to Fc

Ž .alone Fig. 3 or EphA5-Fc ligand fusion protein did notresult in any alteration in cell proliferation compared toFCS controls.

4. Discussion

Expression or activation of Eph receptors has beenw xlinked with glioma tumourigenisis 19,28 and epithelial

w xcell lines 7,12,25 , however whether Eph has a mitoticeffect is not clear. With fluorescent staining we showEphA5 expression on the cell surface of the U-118 MGhuman glioblastoma cell line. Exposure to ephrin-A1-Fclead to specific time-dependent tyrosine phosphorylationof EphA5. U-118 MG cells cultured in different FCSmedium concentrations and exposed to ephrin-A1-Fc,showed a marginal decrease in cell proliferation. We there-fore suggest that in the U-118 MG human glioblastomacell line, EphA5 is functional, but activation is not associ-ated with any significant effect on cell proliferation.

Clustering of the ligand fusion protein at a variety ofconcentrations and time points of incubations did notfurther increase EphA5 phosphorylation from that obtainedwith ephrin-A1-Fc alone. This is in agreement with Gale

w xand Yancopoulos 11 , who found that ephrin-A3-Fc andephrin-A5-Fc stimulation of EphA5 could be strongly en-hanced with ligand fusion protein clustering, whilst stimu-lation with ephrin-A1-Fc alone already gave a strongactivation signal and was not significantly enhanced withfurther ligand clustering.

RTKs of the EphA family have been shown to havecontrasting effects on cell proliferation in cell lines of

w xdifferent origin 3,7,15,25 Conceivably, the different EphRTKs are linked to different intracellular signalling path-ways. An alternative explanation could be that structurallyvery similar RTKs expressed in different cell types mayhave opposite effects due to cell-type specific intracellular

w xsignalling proteins 26 . Thus, EphA5 activation may acti-vate a signalling pathway in the U-118 MG glioblastomacells not linked to increased mitotic activity. This may alsobe the case in studies investigating the cellular effects ofEph activation in cells not endogenously expressing Eph

w xreceptors 3,15 . Interestingly, the same Eph receptor in the

( )V. Bruce et al.rBrain Research 821 1999 169–176 175

same cell type may be involved in different cell states. Forexample, in mammary epithelial cells of the mouse, upreg-ulation of EphA2 and EphB4 have been correlated toproliferation, and downregulation of both Eph receptors to

w xdifferentiation, during the oestrus cycle 2 .Expression of EphA5 in gliomas may be involved in the

extensive migration found in most glioma tumours, andmay regulate cell adhesion molecules known to be down-

w xregulated in gliomas 8,23 . Downregulation of cell adhe-sion molecules has also been found during neural crest cellmigration, where Eph receptor involvement has been im-

w xplicated 13,27 . In addition, involvement in the failure ofmany glioma tumours to metastatize out of the nervoussystem may also be postulated as ephrin-A1 is expressed

w xby activated endothelium during inflammation 21,24 .However, glioma EphA5 expression may be linked withthe undifferentiated stage of the glioblastoma. Supportingthis suggestion, we have previously reported membraneEphA5 expression in immature undifferentiated neuralrglial cells in the germinal layer of the cerebral cortex in

w xnewborn rats 29 .In conclusion, our results indicate that EphA5 activation

in the U-118 MG human glioblastoma cell line does notlead to any significant changes in cell proliferation. Itwould be interesting to determine whether Eph receptorsare activated in primary human gliomas. Further studiesare required to determine whether oncogenic Eph expres-sion or constitutive activation involves mutated Eph genes,and whether levels of EphA5 vary with glioma progres-sion.

Acknowledgements

We are extremely grateful to Daniel Maier for criticalcomment and discussion, and to N. Schaeren-Wiemers andM. Roth. The ephrin-A1-Fc fusion protein was donated byN. Gale, Regeneron Pharmaceuticals, Tarrytown, NY,USA. The U-118 MG line was obtained from B. Wester-mark, University of Uppsala, Sweden. This work was

Žsupported by the Swiss Nationalfonds grant number 31-.45953.95 . O.E. is currently a Feodor-Lynen Fellow of the

Alexander von Humboldt Association.

References

w x1 Eph Nomenclature Committee, Unified nomenclature for Eph familyŽ .receptors and their ligands, the ephrins, Cell 90 1997 403–404,

letter.w x2 A.C. Andres, H.H. Reid, G. Zurcher, R.J. Blaschke, D. Albrecht, A.

Ziemiecki, Expression of two novel eph-related receptor proteintyrosine kinases in mammary gland development and carcinogenesis,

Ž .Oncogene 9 1994 1461–1467, published erratum appears in Onco-Ž .gene 1994 Aug;9 8 :2431.

w x3 R. Brambilla, A. Schnapp, F. Casagranda, J.P. Labrador, A.D.Bergemann, J.G. Flanagan, E.B. Pasquale, R. Klein, Membrane-bound LERK2 ligand can signal through three different Eph-related

Ž .receptor tyrosine kinases, EMBO J. 14 1995 3116–3126.

w x4 C.N. Chesterman, T. Walker, B. Grego, K. Chamberlain, M.T.Hearn, F.J. Morgan, Comparison of platelet-derived growth factorprepared from release products of fresh platelets and from outdated

Ž .platelet concentrates, Biochem. Biophys. Res. Commun. 116 1983809–816.

w x5 S. Davis, N.W. Gale, T.H. Aldrich, P.C. Maisonpierre, V. Lhotak,T. Pawson, M. Goldfarb, G.D. Yancopoulos, Ligands for EPH-re-lated receptor tyrosine kinases that require membrane attachment or

Ž .clustering for activity, Science 266 1994 816–819.w x6 U. Drescher, C. Kremoser, C. Handwerker, J. Loschinger, M. Noda,

F. Bonhoeffer, In vitro guidance of retinal ganglion cell axons byRAGS, a 25 kDa tectal protein related to ligands for Eph receptor

Ž .tyrosine kinases, Cell 82 1995 359–370.w x7 D.J. Easty, B.A. Guthrie, K. Maung, C.J. Farr, R.A. Lindberg, R.J.

Toso, M. Herlyn, D.C. Bennett, Protein B61 as a new growth-factor-expression of B61 and up-regulation of its receptor epithelial-

Ž .cell kinase during melanoma progression, Cancer Res. 55 19952528–2532.

w x8 K. Edvardsen, P.H. Pedersen, R. Bjerkvig, G.G. Hermann, J.Zeuthen, O.D. Laerum, F.S. Walsh, E. Bock, Transfection ofglioma-cells with the neural-cell adhesion molecule NCAM-effect

Ž .on glioma-cell invasion and growth in vivo, Int. J. Cancer 58 1994116–122.

w x9 J. Frisen, P.A. Yates, T. McLaughlin, G.C. Friedman, D.D. O’Leary,Ž .M. Barbacid, Ephrin-A5 AL-1rRAGS is essential for proper reti-

nal axon guidance and topographic mapping in the mammalianŽ .visual system, Neuron 20 1998 235–243.

w x10 N.W. Gale, S.J. Holland, D.M. Valenzuela, A. Flenniken, L. Pan,T.E. Ryan, M. Henkemeyer, K. Strebhardt, H. Hirai, D.G. Wilkin-son, T. Pawson, S. Davis, G.D. Yancopoulos, Eph receptors andligands comprise two major specificity subclasses and are recipro-

Ž .cally compartmentalized during embryogenesis, Neuron 17 19969–19.

w x11 N.W. Gale, G.D. Yancopoulos, Ephrins and their receptors: a repul-Ž .sive topic?, Cell and Tissue Research 290 1997 227–241.

w x12 E. Kiyokawa, S. Takai, M. Tanaka, T. Iwase, M. Suzuki, Y.Y.Xiang, Y. Naito, K. Yamada, H. Sugimura, I. Kino, Overexpressionof ERK, an EPH family receptor protein tyrosine kinase, in various

Ž .human tumors, Cancer Res. 54 1994 3645–3650.w x13 C.E. Krull, R. Lansford, N.W. Gale, A. Collazo, C. Marcelle, G.D.

Yancopoulos, S.E. Fraser, M. Bronner-Fraser, Interactions of Eph-related receptors and ligands confer rostrocaudal pattern to trunk

Ž .neural crest migration, Curr. Biol. 7 1997 571–580.w x14 U.K. Laemmli, Cleavage of structural proteins during the assembly

Ž .of the head of bacteriophage T4, Nature 227 1970 680.w x15 V. Lhotak, T. Pawson, Biological and biochemical activities of a

chimeric epidermal growth factor-Elk receptor tyrosine kinase, Mol.Ž .Cell. Biol. 13 1993 7071–7079.

w x16 R.A. Lindberg, T. Hunter, cDNA cloning and characterization ofeck, an epithelial cell receptor protein–tyrosine kinase in the ephrelk

Ž .family of protein kinases, Mol. Cell. Biol. 10 1990 6316–6324.w x17 P.C. Maisonpierre, N.X. Barrezueta, G.D. Yancopoulos, Ehk-1 and

Ehk-2: two novel members of the Eph receptor-like tyrosine kinasefamily with distinctive structures and neuronal expression, OncogeneŽ .8 1993 3277–3288.

w x18 L. Meima, I.J. Kljavin, P. Moran, A. Shih, J.W. Winslow, I.W.Caras, AL-1-induced growth cone collapse of rat cortical neurons iscorrelated with REK7 expression and rearrangement of the actin

Ž .cytoskeleton, Eur. J. Neurosci. 9 1997 177–188.w x19 G.C. Miescher, V. Taylor, G. Olivieri, T. Mindermann, E. Schrock,¨

A.J. Steck, Extensive splice variation and localization of the EHK-1receptor tyrosine kinase in adult human brain and glial tumors, Brain

Ž .Res. Mol. Brain Res. 46 1997 17–24.w x20 M. Nister, L. Claesson-Welsh, A. Eriksson, C.H. Heldin, B. Wester-

mark, Differential expression of platelet-derived growth factor recep-Ž .tors in human malignant glioma cell lines, J. Biol. Chem. 266 1991

16755–16763.

( )V. Bruce et al.rBrain Research 821 1999 169–176176

w x21 A. Pandey, H. Shao, R.M. Marks, P.J. Polverini, V.M. Dixit, Role ofB61, the ligand for the Eck receptor tyrosine kinase, in TNF-alpha-

Ž .induced angiogenesis, Science 268 1995 567–569.w x22 E.B. Pasquale, The Eph family of receptors, Curr. Opin. Cell. Biol.

Ž .9 1997 608–615.w x23 P.-H. Pedersen, K. Marienhagen, S. Mork, Bjerkvig, Migratory

pattern of fetal rat brain cells and human glioma cells in the adult ratŽ .brain, Cancer Res. 53 1993 5158–5165.

w x24 G.J. Pilkington, The paradox of neoplastic glial cell invasion of theŽ .brain and apparent metastatic failure, Anticancer Res. 17 1997

4103–4105.w x25 I.M. Rosenberg, M. Goke, M. Kanai, H.-C. Reinecker, D.K. Podol-¨

sky, Epithelial cell kinase-B61: an autocrine loop modulating intesti-nal epithelial migration and barrier function, Am. J. Physiol. 273Ž .1997 G824–G832.

w x26 J. Schlessinger, A. Ullrich, Growth factor signaling by receptorŽ .tyrosine kinases, Neuron 9 1992 383–391.

w x27 A. Smith, V. Robinson, K. Patel, D.G. Wilkinson, The EphA4 andEphB1 receptor tyrosine kinases and ephrin-B2 ligand regulatetargeted migration of branchial neural crest cells, Curr. Biol. 7Ž .1997 561–570.

w x28 C. Soans, J.A. Holash, E.B. Pasquale, Characterization of the ex-pression of the Cek8 receptor-type tyrosine kinase during develop-

Ž .ment and in tumor cell lines, Oncogene 9 1994 3353–3361.w x29 V. Taylor, G.C. Miescher, S. Pfarr, P. Honegger, H. Breitschopf, H.

Lassmann, A.J. Steck, Expression and developmental regulation ofEhk-1, a neuronal Elk-like receptor tyrosine kinase in brain, Neuro-

Ž .science 63 1994 163–178.w x30 N.L. Tuzi, W.J. Gullick, Eph, the largest known family of putative

Ž .growth factor receptors, Br. J. Cancer 69 1994 417–421.w x31 H.U. Wang, D.J. Anderson, Eph family transmembrane ligands can

mediate repulsive guidance of trunk neural crest migration andŽ .motor axon outgrowth, Neuron 18 1997 383–396.

w x32 J.W. Winslow, P. Moran, J. Valverde, A. Shih, J.Q. Yuan, S.C.Wong, S.P. Tsai, A. Goddard, W.J. Henzel, F. Hefti et al., Cloningof AL-1, a ligand for an Eph-related tyrosine kinase receptor in-

Ž .volved in axon bundle formation, Neuron 14 1995 973–981.w x33 Q. Xu, G. Alldus, N. Holder, D.G. Wilkinson, Expression of trun-

cated Sek-1 receptor tyrosine kinase disrupts the segmental restric-tion of gene expression in the Xenopus and zebrafish hindbrain,

Ž .Development 121 1995 4005–4016.