Activation of Cultured Human Endothelial Cells Byrecombinant Lymphotoxin:Comparison with Tumor Necrosisfactor and Interleukin 1 Species'

Home , Lymphotoxin

0022-1767/87/13810-3319802.00/0 vol. 138.3319-3324. 10. May 15. 1987 THEJOURNAL OF IMMUNOUJCY No. Prlnted In U.S. A. Copyright IB 1987 by The American As80clatlonof Immunologists

ACTIVATION OF CULTURED HUMAN ENDOTHELIAL CELLS BYRECOMBINANT LYMPHOTOXIN:COMPARISON WITH TUMOR NECROSISFACTOR AND INTERLEUKIN 1 SPECIES'

JORDAN S. POBER,'"LYNNE A. LAPIERRE," ALANH. STOLPEN,' TOMMYA. BROCK," TIMOTHY A. SPRINGER,'WALTER FIERS,* MICHAEL P. BEVILACQUA," DONNA L. MENDRICK," AND MICHAEL A. GIMBRONE JR" From the +Departmentof Pathology, Brigham and Women's Hospital and Harvard Medical School;the 'Department of Pharmacology, Harvard Medical School;the 'Laboratory of Membrane Immunochemistry, Dana-Farber Cancer Institute, Boston, MA 021 15; and the 'Laboratory of Molecular Bfology, State Universftyof Ghent, Ghent, Belgfum

Recombinant human lymphotoxin (LT) was com- pretreatment with IL la or IL 10 only partially in- pared with recombinant human tumor necrosis fac- hibited induction of H4/18 binding by phorbolester, tor (TNF) for direct actions on cultured human en- and phorbol ester pretreatment consistently, albeit dothelial cells (HEC). At equivalent half-maximal partially, inhibited induction by IL 1 species. These concentrations (based on L929 cytotoxicity units) studies suggest that activated T cells through the LT and TNF each caused rapid and transientinduc- secretion of LT can in turn activate the local endo- tion (peak 4 to 6 hr) of an antigen associated with thelial lining so as topromote homingand extrava- leukocyte adhesion (detected by monoclonal anti- sation of inflammatory cells.Furthermore, these LT body H4/18),a rapid but sustainedincreased expres- actions can beaugmented or complemented by sion (plateau 24 hr) of a lymphocyte adhesion struc- other locally produced mediators such as IFN-r or ture (ICAM-1). a gradual (plateau 4 to 6 days) in- IL 1. crease in expression of HLA-A,B antigens, and gradual (4 to 6 days) conversion of HEC culture morphology from epithelioid to fibroblastoid, an effect Lymphotoxin (LT)3was initially defined as a protein enhanced by immune interferon (IFN-7). Induction factor generatedin response to T cell activation that could of H4/18 binding by maximal concentrations of LT mediate cytolysis of tumor cells and other targets (1). Its or TNF could not be augmented by addition of the spectrum of tumoricidal activity is strikingly similar to other cytokine, and 24 hr pretreatment with LT or that of the activated macrophageproduct tumor necrosis TNF produced hyporesponsiveness to both media- factor (TNF)(2). Subsequently, proteins with bothLT and tors for reinduction. H4/18 binding can be tran- TNF activities have been purified to homogeneity, se- siently induced by tumor-promoting phorbolesters. quenced, and cloned (3-6). Detailed structural analysis Pretreatment with either LT or TNF also fully inhibited induction of H4/18 binding by phorbol ester, has revealed regions of homology between LT and TNF. whereas phorbol ester pretreatment only variably Furthermore, thepurified proteins bind competitively to and partially inhibited reinduction by LT or TNF. receptorson certain tumor cells (7). Interestingly the These actions of LT on endothelium shared with genes for LT and TNF appear to be arranged in tandem TNF may serve in vivo to promote lymphocyte and within the major histocompatibility complex (MHC) (8). inflammatory leukocyte adhesion and transendo- although the significance of this finding is unknown. thelia1 migration. Recombinanthuman interleukin The recent availability of recombinant TNF has re- 1 species (IL la and IL 10) shared many of the vealed that this cytokine is a potent pleiotropic inflam- actions of LT and TNF and were indistinguishable matory mediator (9).Remarkably, manyof the actionsof from each other. However, IL 1 species could be TNF are similar or identical to those observed with nat- distinguished from LT/TNF by their relative inabil- ural or recombinant interleukin 1(IL 1)species. We have ity to enhance HLA-A,B expression, by their ability been studyingthe actions of TNF and IL 1 upon vascular to augment H4/18 binding caused by maximally ef- endothelial cells. For example, by using cultured human fective concentrations of LT or TNF, and by their endothelial cells (HEC), we and others have found that inability to inhibit reinduction of H4/18 binding by TNF and IL 1 cause transient cell surface expression of LT or TNF. In contrast tothe actions of LT or TNF, tissuefactor procoagulant activity(10-12) andtran- siently render endothelial cells markedly more adhesive Received for publication December 19. 1986. for inflammatory leukocytes and related cell lines (13- Accepted for publication February 6. 1987. 16).IL 1 has been reported to render HEC more adhesive The costs of publication of this article were defrayed in part by the for lymphocytes, although this effect appears to persist payment of page charges. This article must therefore be hereby marked aduertfsement in accordance with 18 U.S.C. Section 1734 solely to indi- longer than the adhesiveness for other leukocytes (17). cate this fact. Changes in HEC adhesiveness can be correlated with ' This work was supported by NIH grants HL-36003and 22602. Jordan S. Pober is an Established Investigator of the American Heart Association. changes in surface antigen expression.Specifically, both Research on TNF is supported by Biogen. TNF and IL 1 induce de novo synthesis and transient To whom correspondence should be addressed at theDepartment of Pathology. Brigham and Women's Hospital. 75 Francis Street. Boston. Abbreviations used: LT, lyrnphotoxin; HEC. human endothelial cells: MA 021 15. TNF. tumor necrosis factor. 3319 3320 L~PHOTO~NCAN ACTIVATE HU~NEN~T~LIUM expression (peak 4 to 6 hr) of an endothelial surface daunorubicin) and protein synthesis (cycloheximide and emetine) antigen associated with monokine-stimulated leukocyte inhibitors (all from Sigma) were added 15 to 30 min before the addition of cytokines or mediators at the concentrations indicated adhesion (16) recognized bymurine monoclonal antibody (found in other experiments to inhibit l9jmethionine incorporation H4/18 (18), and causea rapid but sustained increase in into trichloroacetic acid precipitable proteinby 95% or greater with- expression (above basal levels) of a lymphocyte adhesion out inducing cell detachment or lossof trypan blue excluslon.elther in the presence or absence of mediators). structure (called intercellular adhesion molecule or Morphology experiments were performed on early subcultured ICAM-I), recognizedby murine monoclonal antibody cells (first or second passage level) in microtiter wells. Cytokines RRl/l (19). TNF has been found to induce a gradual (4 were introduced at visual confluence. Cell monolayer organization was assessed by phase contrast microscopy after 2.5% glutaralde- to 6 day) increase in expression of class I MHC antigens hyde fixation, methanol dehydration. and air drying. Morphology (ZO),the recognition structures for cytolytic T lympho- was rated sem~q~tita~velyby two independent observers by using cytes. Coincident with the enhanced expression of class the following scale: 0 = cells are uniformly polygonal and non- I antigens, TNF causes a rearrangement of cultured HEC overlapping ("epithelfoid"); 1 = most cells are polygonal: 2 = most cells are elongated: and 3 = cells are thin and hIghly elongated, whorl monolayers. Treated cultures appear fibroblastoid, ex- patterns are evident."dendritic" forms may be present. and no hibit numerous intercellular gaps, and lose basement clusters of polygonal cells are found ~fibroblastoid").These changes membrane components (21). (IL 1 effects on HEC mor- were illustrated (21). Surface antigen expression was quantitated on serially subcul- phology have also been reported (221, but thesehave not tured cells (passage levels 3 to 12) either by radioimmunobinding been studied with recombinant IL 1 species free of poten- assay to replicate monolayersin microtiter wells, orby fluorescence tial TNF contamination. Furthermore, it cannot be as- flow cytometry on trypsin-versene suspended cell populations (18. 20). Monoclonal antibodies used in these studies wen H4/18 (IS), sessed from these reports whether IL 1-induced morpho- RR1/1 (anti-ICAM-1) (29), W6/32 (anti-HLA-A,B)(30). E1/1.2 (anti- logical changes are similar to those induced by TNF). mesenchymal cell)(19). and isotype matched nonbinding controls. Collectively, these and other studieshave suggested that In some radio~mmunobindingexperiments, test antibody binding is was normalized to E1/1.2 binding, but any experiment revealing vascular endothelium an important target cell for the significant variation in El/1.2 binding, indicativeof changes in cell proin~ammato~actions of TNF, as well as for those of number, was excluded. IL 1. It is not presently known whether LT will share with TNF its proin~ammato~,as well as tumoricidal RESULTS activities. In the present study we have compared the actions of recombinant human LT on cultured HEC with Human recombinant LT, like TNF, induces expression those of recombinant TNF and of recombinant IL 1 spe- of specific H4/18 binding, (Fig. l),increases expression cies. We find that in most assays of antigenic and mor- of ICAM-1 (measured by specific RRl/l binding: Fig. 2) phologic modulation, LT appears both qualitatively and and increases expression of HLA-A,B antigens (measured quantitatively similar to TNF in its spectrum of activities. by specific W6/32 binding, Fig.3 and Table I). The max- Furthermore, in some of the assays used, LT and TNF, imal increases produced by LT occasionally varies from although indistinguishable from eachother, may be clearly differentiated from IL la and IL 18.

MATERIALS AND METHODS HEC were isolated from umbilical veins (23) and were serially subcultured (24)using conditions and reagents described elsewhere (18). Cytokines and other mediators were added to the culture medium for the times and at the concentrations indicated in each experiment. The cytokines used in these studies were recombinant LT (expressed in CHO cells and partially purified to final concentration of 3.2 x 10' U/ml in phosphate-buffered saline, speclfic activity 3 X lo* U/mg protein), recombinant TNF (expressed in E. coli and purified to homogeneity, specific activity 2.5 x 10' U/mg) (25). recombinant IL la (Genzyme, Boston, MA) (26). recombinant IL. 119 (Cistron, Pinebrook. NJ) (27), and recombinant Immune interferon (IFN-y, expressed in CWO cells with final concentration of 2.5 X 1Os U/ml) (28).In some experiments, ~finity-purifiednatural IL 1 prep- arations (from Genzyme or Cistron) were used instead of recombinant IL 1 molecules. TNF and LT units were determined in L929 cytotoxicity assays (performed at Biogent courtesy of Drs. Jan Tav- ernier and Jose Van der Heyden): IL 1 units arebased on information provfded by the suppliers for thymocyte co-stimulating activity: and IFN--, activity was established in standard viral protection assays (28).Note that the two cytokines used inthis study that are naturally glycosylated (i.e.. LT and 1FN-7) were produced in mammalian cell expression systems and are thusglycosylated ln recombinant form as well. Other mediators used in these studies include the calcium ionophores A23 187 and ionomycfn (~~bf~hem,San Diego, CA), the tumor promoting phorbol esters 4-phorbol 12-myristate 13-acetate (PMA). and mezerein (Sigma ChemicalCo., St. Louis. MO). as well as the less active 4-0-methyl derivative of PMA (4-0-methyl €"A, also from Sigma). Stock solutions of these compounds were made in dimethyl sulfoxide (Sigma), and equivalent carrier concentrations were added tocontrol cultures in the reIevant experiments. For short term (less than 24 hr) induction experiments, cytokines or mediators oy ' ' I I wereadded in 20% fetal bovine serum in medium 199 wlthout 12.5 25 37.5 SO additional growth factors: no dffferences were notedwhether growth factors were included in the medium, whether serum was replaced CYTOKINE CONCENTRAW (Ulml) with albumin, or whether stimulation was conducted in protein free Ffgure2. Induction of H4/18 blnding by cytoklnes assessed by ra- medium. Finally, in some experiments, RNA (actinomycin D and dfofmmunoassay at 4 hr. Panels A and B arc aeparate expcrlments. LYMPHOTOXIN CANACTIVATE HUMAN ENDOTHELIUM 332 1 TABLE I1 Morphological rearrangement in HEC cultures at 96 hr Cytoklne Morphological AppearanceMorphological Cytoklne Concentrations IWW IL la IL 1s TNF LT 1" I* 2 0 2 3 2 3 1 3 3 2 3 10 2 3 3 3 3 3 3 3 3 30 3 Numbers represent thymocyte costimulating activity for IL 1 species or L929 cytotoxicity units X 10" for TNF and LT. 'Scale: 0,fully epithelioid: 1, mostly epithelioid: 2,mostly fibroblastoid: 3, fully fibroblastoid. Criteria are described in Matertals andMethods. CMOK*(E CONCENTRATKN (Ulrnl) In replicate cultures in thisexperiment, treatment with IFNq at 200 U/ Figure 2. Enhancement of ICAM-1 expression by cytokines assessed ml scored 1: co-addition of 1FN-y (200 U/ml) to any othercytokine at any by radioimmunoassay at 4 hr. concentration scored 3, indicating potential synergy.

TABLE 111 Induction ofH4/18 binding: specficity of mediators Speclfic H4/18 E&dlng (mean ""_" ---LT d Mediator (Conc) cpm/well f SD X 4 Expt. 1. None 0.8 * 0.2 IL 1 (5 U/ml)" 4.6 * 0.4 PMA (5 X lo-' M) 6.7 * 0.4 4-0-methyl PMA (5 x M) 0.8 i 0.5

Expt. 2. None 1.6 i 0.6 TNF (100 U/ml) 15.1 * 1.2 A23187 (5 X 10" M) 1.5 i 1.3 Ionomycin (1 x 1 0-e M) 2.0 i 1.2 " In this experiment. natural IL 1 was used. Binding was assessed at 4 OLb 160 ' 200 ' 3;)O ' 4hO hr in both experiments. CYTWINE CONCENTRATKN (Ulrnl) Figure 3. Enhancement of HLA-A.B antigen expression by cytokines assessed by radioimmunoassay at 4 hr.

TABLE I HLA-AB Antigen modulation SpeClflC Cytokine (conc) W6/32 Blndlng None 1 .o" IL la (5 U/ml) 1.1 IL 18 (5 U/ml) 1.6 TNF (20U/ml) 3.1 LT 125 U/mll 2.5 a Ratlo at 72 hr of treated tocontrol model specific W6/32 binding cell determined by FACS analysis. In this experiment, some IL la treated cells appeared to show diminished HLA-A.B expression. although the model fluorescence did not decrease. those obtained with TNF. but no consistent pattern was found. However, the concentrations of LT and TNF that induce half-maximal stimulation are consistently com- Figure4. Induction of H4/18 binding byPMA. Upper panel shows parable (10 to 50 U/ml) when standardized by cytotoxic time course in response to 100 nM PMA: lower panel shows expression at 4 hr as a function of PMA concentration. Each panel isderived from activity in L929 cells (Figs. 1 to 3). Like TNF, LT causes a separate experiment. the organization of HEC monolayers in endothelial cultures to change from epithelioid to fibroblastoid; cells in treated cultures areelongated, overlapping, and arranged TABLE 1V in whorls (Table11). Both TNFand LT synergize with IFN- Inductton of H4f18 binding: phorbol ester tachyphylaxis y in producing morphological rearrangement (Table11). Percent Inhlbltlonof To investigate the mechanismsof LT and TNF actions Pretreatment (24 hr) Restlmulatlon (4 hr)' on HEC, we used induction of specific H4/18 binding as PMA Mezerein an assayto test potential second messengers. We found PMA ( 100 nM) 90 94 no effect at 4 to 6 hr of incubation of ionophores that Mezerein (300 nM) 100 91 elevate intracellularcalcium (A23187 or ionomycin; Ta- Induced sDecific H4/18 binding to me&iator pretreated HEC ble 111). In contrast, we observed transient induction of "Percent Inhibition = 1 - H4/18 binding with the tumor promoting phorbol ester Induced sDecific H4/18 binding x 100 to sham pretreated HEC PMA (Table I11 and Fig. 4). but not with the less active In this experiment, H4/18 binding to mediator-pretreated cultures in derivative 4-0-methyl PMA (Table 111). Other tumor pro- the absenceof restimulation (Le., at 28 hr) was less than20% of binding moting phorbol esters such as phorbol 12,13-dibutyrate at 4 hr afterstimulation of sham-pretreated cultures. 3322 LYMPHOTOXIN CAN ACTIVATE HUMAN ENDOTHELIUM (not shown) or mezerein, (Table IV) were also effective. or TNF. Finally, pretreatment with PMA can produce Similar to induction by IL 1 species (18) or TNF (shown partialhyporesponsiveness for reinduction of H4/18 for comparison in Table V), induction of H4/18 binding binding by IL 1 species,but does not significantly inhibit by PMA is inhibited by RNA or protein synthesis inhibi- reinduction by TNF (Table VII). Interestingly, PMA pre- tors. Interestingly,HEC treated withPMA or mezerein for treatment does in some experiments(Table VII) partially 24 hr become refractoryto restimulation with either inhibit reinductionby LT. but experimental variability in tumor-promoting phorbol ester (TableIV). this result cautions against interpreting thiseffect as a To compare LT (and TNF) actions with those of IL 1, difference between LT and TNF. On the other hand, we next investigated the effects of recombinant IL la and pretreatment with TNF or LT markedly and consistently IL 18 in thevarious endothelialcell assays. Recombinant blocks reinduction by PMA, whereas pretreatment with IL 1a and IL 18 each induce expressionof H4/ 18 binding IL 1 species only partially inhibits reinduction by PMA (Fig. 1). IL la and IL 18 also mimic LT/TNF effects on (Table VII). endothelial cell morphology, and show similar synergy with IFN-7 (Table 11). Half maximaleffects with IL 1 DISCUSSION species for both actions are seen between 1 and 5 thy- The observations presented inthis report demonstrate mocyte co-stimulatory U/ml. None of the assayswe have that LT shares with TNF a number of direct endothelial used distinguishIL 1 a from IL 18. In contrast, IL 1 species cell actions. Specifically, LT causes enhanced expression can be differentiated from LT and TNF in several ways. of endothelial cell receptors for various leukocyte recog- First, IL 1 species are consistently less effective at in- nition structures (e.g., the antigen recognized by H4/18, creasing the expression of HLA-A,B antigens (Table I), ICAM-1, and HLA-A,B). The LT altered endothelial cell often showing no effectat all. Despite the lack of effect surface may serve to facilitateleukocyte and lymphocyte on HLA-A,B expression, IL 1 speciesat the sameconcen- adhesion at a site of cell-mediated immunity. In addition, trations areconsistently active in otherassays performed LT, like TNF, can cause morphological reorganization of on replicate cells in the same experiment (e.g., ICAM-1 HEC monolayers, a change that may serve to facilitate enhancement, shape change, etc.). Second, co-addition subsequent transendothelial trafficking of lymphocytes of IL 1 species to HEC treated with maximally-effective and other inflammatory leukocytes. We have hypothe- concentrations sf LT or TNF leads to small but measur- sized that these in vitro changes represent the same able augmentation of H4/18 binding, whereas the com- process that results in the in situ specialization of the bination of LT with TNF is no betterthan either cytokine high endothelialvenule (21). Individual experiments alone [Table VI). Similarly, the combination of maximally show some variations of the relative magnitude of the effective concentrations of IL la and IL 18 is no better maximal responsesto recombinant LT and TNF, al- that either cytokine alone, whereas each can be aug- though no consistent pattern has been noted. However, mented by LT or TNF. Third, although all four cytokinesthe concentrations ofLT and TNF that produce half produce tachyphylaxis after 24 hrof treatment for rein- maximal changes on HEC are consistently comparable duction of H4/18 binding by the same cytokine, LT and when assessed by bioassay units determined from mu- TNF produce cross-tachyphylaxis to each other but not rine L929 cytotoxicity. Overall, we find no evidence that to IL 1CY or IL 18 (Table VII). Similarly, IL la and 18 these recombinant mediators can be distinguished by produce cross-tachyphylaxis to each otherbut not to LT their HEC actions. It is unlikely that these recombinant mediators are cross-contaminated or contain otherHEC- TABLE V active mediators(e.g., endotoxin), aconclusion reinforced Induction ofH4/18 blndlng: metabolic lnhlbltlon? by our ability to selectively neutralize both TNF and LT TNF PMA with separate polyclonal sera purchased from Endogen (100 U/ml) (50 nM) (Boston, MA) (data not shown). Furthermore, a natural Cycloheximide (20pg/ml) 77% 83% LT preparation (kindly provided by Dr.Berish Rubin, New Emetine (5pM) 89% 97% Actinomycin D (5 &ml)90% 97% York Blood Center), produced similar effects on H4/18 Daunorubicin (5 dml)89% 97% binding and HLA-A,B antigen expression (J.S. Pober and a Percent Inhlbitlon J. L. Strominger. unpublished), confirming that recombinant LT expressed in CHO cells behaves like a human = (1 - ) x 100 Induction of specific H4/18 binding without inhibitor lymphocyte-derived product. assessed at 4 hours. The second major point of this report is that although LT and TNF appear to be similar to IL la and IL 18 in TABLE VI some endothelial cell assays, these two pairs of mediators Induction of H4/18 blndlng: combined addltionof cytoklnes can be distinguishedby other assays.We interpret these Percent ChangP Second Cytoklne observations as consistent with the hypothesis that LT Flrst Cytokine IL la IL 1s TNF LT and TNF share one receptor (7) but that IL la and 1s - share a second (31, 32). As we have argued previously IL 1(Y (25U/ml) (-22) +33 +17 IL 18 (25U/ml) (-29) - +2 1 +17 (19),it is unlikely that tachyphylaxis for reinduction of TNF (500U/ml) +54 +54 - +2 H4/18 binding results simply from receptor disappear- +17 +18 - LT (500U/ml) (-12) ance, because sustained treatment withcytokine is nec- *Percent Change essary to induce and maintainfibroblastoid morphology /Swcific H4/18 binding after two cytokines\ and other gradual changes. . - binding after first cytokine Binding after first cytokine Our PMA experiments areof interest withregard to the mechanism of action of LT, TNF, and IL 1 species. The in which specific H4/18 binding is normalized to percent of specific El/ 1.2 binding in each treatment group. factthat PMA andother phorbol esterscan induce LYMPHOTOXIN CAN ACTIVATE HUMAN ENDOTHELIUM 3323 TABLE VI1 Induction of H4/18bindlng: tachyphylaxis to cytokines" Percent Inhibltlon of Restimulation (4 hrIb Pretreatment (24 hr) IL lu IL la TNF LT PMA - IL la ( 10 U/ml) 68 71 (-21) 24 22 IL 18 ( 10 U/ml) 92 97 14 (-51 44 TNF ( 100 U/ml) 93 36 7827 82 LT ( 100 U/ml) 29 24 71 88 86 PMA (100 nM) 70 68 (-6) -62 90 "Each horizontal row is representative of three to six separate internally controlled experiments: comparisons of absolute values between experiments may be misleading. Italicized values indicate significant (Le., greater than 2/3) inhibition. The inhibition of reinduction by LT after PMA treatment hasbeen variable among six separate experiments. and is therefore underlined. Defined as in Table IV. In each experiment,H4/18 binding to mediator-pretreated cultures without restimulation was less than30% of binding at 4 hr afterstimulation of sham-pretreated cultures.

3. Aggarwal. B. B., B. Moffat, and R. N. Harkins. 1984. Human lym- expression of H4/18 binding suggests potential involve- photoxin. Production by a lymphoblastoid cell line. purification, and ment of a protein kinase C-dependent process (33).The initial characterization. J. BLol. Chem. 259:686. fact thatLT and TNF are active incells that have become 4. Aggarwal. B. B.. W. J. Kohr, P. E. Hass. B. Moffat, S. A. Spencer, W. J. Henzel, T. S. Bringman. G. E. Nedwin, D. V. Goeddel, and R. completely unresponsive toPMA (i.e., after PMA pretreat- N. Harkins. 1985. Human tumor necrosis factor. Production, purifi- ment) suggests that a pathway independent of protein cation and characterization. J. Biol. Chem. 260:2345. kinase C must exist. The mechanism by which LT or 5. Gray, P. W..B. B. Aggarwal, C. V. Benton. T. S. Bringman, W. J. Henzel, J. A. Jarrett, D. W.Leung, B. Moffat, P. Ng, L. P. Svedersky. TNF pretreatment inhibitsPMA induction of H4/18 bind- M. A. Palladino, and G. E. Nedwin. 1984. Cloning and expression of ing is not clear. It is interestingthat HEC that have been cDNA for human lymphotoxin. a lymphokine with tumour necrosis pretreated with TNF for 24 hr and become completely activity. Nature 312721. 6. Pennica. D.. G. E. Nedwin. J. S. Hayflick, P. H. Seeburg. R. Derynck, refractory to stimulationof H4/18 binding expressionby M. A. Palladino, W. J. Kohr, B. B. Aggarwal, and D. V. Goeddel. PMA, still can be induced to secrete von Willebrand factor 1984. Human tumor necrosis factor: precursor structure expression by PMA (B. M. Ewenstein and J. S. Pober, unpublished and homology to lymphotoxin. Nature 312724. 7. Aggarwal. B. B.. T. E. Eessalu. and P. E. Hass. 1985. Characteriza- observations): i.e., not all PMA responses are lost after tion of receptors for human tumour necrosis factor and their regu- TNF pretreatment. Perhaps LT or TNF pretreatment un- lation by y-interferon. Nature 318:665. couples protein kinase C from induction of the H4/18 8. Spies, T., C. C. Morton, S. A. Nedospasov. W. Fiers, D. pious. and J. L. Strominger. 1986. Genes for tumor necrosis factor and lym- bindingprotein, or perhaps LT or TNF inactivates a photoxin are linked tothe humanmajor histocompatibility complex. topologically or biochemically distinct subset of protein Proc. Natl. Acad. Scl. USA 83:8699. 9. Beutler, B.. and A. Cerami. 1986. Cachectin and tumour necrosis kinase C molecules; more than oneprotein kinase C factor as two sides of the samebiological coin. Nature 320:584. species have recently beenidentified (34).Alternatively, 10. Bevilacqua, M. P.. J. S. Pober. G. R. Majeau. R. S. Cotran, and M. PMA may activate expression of H4/18 binding protein A. Gimbrone Jr. 1984. Interleukin 1 (IL-1)induces biosynthesis and cell surface expression of procoagulant activity in human vascular (or von Willebrand factor release) through a mechanism endothelial cells. J. Exp. Med. 160:618. independent of a protein kinase C. 11. Bevilacqua, M. P.. J. S. Pober. G. R. Majeau, W. Fiers, R. S. Cotran. The key implication of our studies is that T lympho- and M. A. Gimbrone Jr. 1986. Recombinant tumor necrosis factor induces procoagulant activity in cultured human vascular endothe- cytes, activated by antigen, can secrete a factor(LT) that lium: characterization and comparison with the actions of interleu- has many of the same endothelial directed actions pre- kin 1. Proc. Natl. Acad. Scl. USA 83:4533. viously described for monocyte-derived inflammatory 12. Nawroth. P. P.. and D. M. Stem. 1986. Modulation of endothelial cell hemostatic properties by tumor necrosis factor. J. Exp. Med. mediators (IL 1 and TNF). We propose that antigen pres- 163:740. entation by endothelium to circulating T helpercells (35) 13. Bevilacqua, M. P., J. S. Pober. M. E. Wheeler, R. S. Cotran. and M. would trigger LT release, leading tosubsequent local A. Gimbrone Jr. 1985. Interleukin 1 acts on cultured human vascular endothelium to increase the adhesion of polymorphonuclear activation of the vascular lining. The LT-induced endo- leukocytes. monocytes and related leukocyte cell lines. J. Clln. In- thelial changeswould then serve in turnto amplify a cell- vest. 76:2003. 14. Gamble, J. R.. J. M. Harlan, S. J. Klebanoff, and M. A. Vadas. 1985. mediated immune response by enhancing local adhesion Stimulation of the adherence of neutrophils to umbilical vein endo- of lymphocytes and other leukocytes to the endothelial thelium by human recombinant tumor necrosis factor. Proc. Natl. cell surface and facilitating transendothelial migration. Acad. Scf. USA 82:8667. 15. Schleimer, R. P., and B. K. Rutledge. 1986. Cultured human vas- Our studies also suggestthat specific LT effects could be cular endothelial cells acquire adhesiveness for neutrophils after augmented by coincident generation of other cytokines stimulationwith interleukin 1, endotoxin and tumor-promoting (e.g., IL 1 or IFN-7). The reciprocal activation of T cells phorbol diesters. J. Imrnunol. 136:649. 16. Bevilacqua. M. P.. M. E. Wheeler, J. S. Pober, W. Fiers, D. L. and lymphocytes may be critical in the development of Mendrick. R. S. Cotran. and M. A. Gimbrone Jr. 1987. Endothelial- cell mediated responses in vascularized tissues. dependent mechanisms of leukocyte adhesion: regulation by interleukin 1 and tumor necrosis factor. In Leukocyte Emigration and Its Sequelae. H. Z. Movat, ed. S. Karger. New York. Acknowledgments. We thank Dr. Ramzi Cotran forhis 17. Cavender, D. E., D. 0. Haskard, B. Joseph, and M. Ziff. 1986. encouragement and support.We wish to thank Drs. Jan Interleukin 1 increases the binding of human B and T lymphocytes Tavernier, Jose Van der Heyden, and Rao Mowa for to endothelial cell monolayers. J. Immunol. 136:203. 18. Pober. J. S.. M. P. Bevilacqua. D. L. Mendrick, L. A. Lapierre, W. preparation of recombinant LT. We thank Mr. Paul Cohill Fiers, andM. A. Gimbrone Jr. 1986. Two distinct monokines, inter- and Ms. Patricia Riendeau forexcellent technical assist- leukin 1 and tumor necrosis factor, each independently induce bio- ance andMs. Donna Hickey for assistance inpreparation synthesis and transientexpression of the same antigen on the surface of cultured human vascularendothelial cells. J. Immunol. of the manuscript. 136: 1680. 19. Pober. J. S., M. A. Gimbrone Jr., L. A. Lapierre, D. L. Mendrick. W. REFERENCES Fiers, R. Rothelein, and T. A. Springer. 1986. Overlapping patterns of activation of human endothelial cells by interleukin 1, tumor 1. Ruddle, N. 1985. Lymphotoxin redux. Immunol. Today 6: 156. necrosis factor and immuneinterferon. J. Immunol. 137:1893. 2. Old, L. J. 1985. Tumor necrosls factor (TNF).Science 230:630. 20. Collins. T., L. A. Lapierre, W. Fiers. J. L. Strominger, and J. S. 3324 LYMPHOTOXIN CAN ACTIVATE HUMAN ENDOTHELIUM Pober. 1986. Recombinant tumor necrosis factor increases mRNA Sci. USA 81:7907. Levels and surface expression of HLA-A,B antigens in vascular en- 28. Scahill. S. J., R. Devos, J. Van der Heyden, and W. Fiers. 1983. dothelial cells and dermal fibroblasts fnuftro. Proc. Natl. Acad. Scf. Expression and characterization of the product of a human immune USA 83:446. interferon cDNA gene in Chinese hamster ovary cells. Proc. Natl. 21. Stolpen, A. H.. E. C. Guinan. W. Fiers, and J. S. Pober. 1986. Acad. Scf.USA 80:4654. Recombinant tumor necrosis factor and immune interferon act sin- 29. Dustin, M. L.. R. Rothlein. A. K. Bhan. C. A. Dinarello, and T. A. ply and In combination to reorganize human vascular endothelialcell Springer. 1986. Induction by IL 1 and interferon-y, tissue distribu- monolayers. Am. J. Pathol. 123~16. tion, biochemistry and function of a natural adherence molecule 22. Montesano. R.. L. Orci, and P. Vassalli. 1985. Human endothelial (ICAM-1).J. Imrnunol. 137:295. cell cultures: phenotypic modulation by leukocyte interleukins. J. 30. Barnstable. C. J., W. F. mer,G. Brown, G. Galfre, C. Milstein, Cell Physfol. 122:424. A. F. Williams, and A. Ziegler. 1978. Production of monoclonal 23. Gimbrone, M. A., Jr. 1976. Culture of vascular endothelium. Prog. antibodies to group A erythrocytes, HLA and other human cell sur- Hernost. Thrornb. 3:1. face antigens-new tools for genetic analysis. Cell 14:9. 24. Thornton, S. C., S. N. Mueller, and E. M. Levine. 1983. Human 31. Kilian, P. L., K. L. Kaffka, A. S. Stem, D. Woehle, W. R. Benjamin, endothelial cells: use of heparin In cloning and long-term serial T. M. Dechiara, U. Gubler, J. J. Farrar, S. B. Mizel, and P. T. cultivation. Science 222:623. Lomedico. 1986.Interleukin la and interleukin 18 bind to the same 25. Marmenout. A., L. Fransen, J. Tavernier. J. Van der Heyden, J. receptor on Tcells. J. Irnrnunol. 136:4509. Tizard, E. Kawashima, A. Shaw. M. J. Johnson, D. Simon, R. 32. Dower, S. K., S. R. Kronheim, T. P. Hopp, M. Cantrell. M. Deeley, Muller, M. R. Ruysschaert, A. Van met. and W. Fiers. 1985. S. Gillis, C. S. HeMey, and D. L. Urdal. 1986. The cell surface Molecular cloning and expression of human tumor necrosis factor receptor for interleukin-la and interleukin-lp areidentical. Nature and comparison with mouse tumor necrosis factor.Eur. J. Bfochern. 324:266. 152:512. 33. Nishizuka, Y. 1984.The role of protein kinase C in cell surface signal 26. March, C. J., B. Mosley, A. Larsen, D. P. Cerretti. 0. Braedt, V. transduction and tumourpromotion. Nature 308:693. Price, S. Gillis, C. S. Henney. S. R. Kronheim. K. Grabstein. P. J. 34. Knopf. J. L.. M.-H. Lee, L. A. Sultzman. R. W. Kriz, C. R. Lolmis. Conlon, T. P. Hopp, and D. Cosman. 1985. Cloning, sequence and R. M. Hewick, and R. M. Bell. 1986. Cloning and expression of expression of two distincthuman interleukin-1complementary multiple protein kinase C cDNAs. Cell 46:491. DNAs. Nature 315:641. 35. Pober, J. S., T. Collins, M. A. Gimbrone Jr.. P. Libby, and C. S. 27. Auron. P. E., A. C. Webb, L. J. Rosenwasser, S. F. Mucci, A. Rich, Reiss. 1986. Inducible expression of class I1 major histocompatibility S. M. Wolff, and C. A. Dinarello. 1984. Nucleotide sequence of complex antigens and theimmunogenicity of vascular endothelium. human monocyte interleukin 1 precursor cDNA. Proc. Natl. Acad. Transplantation41:141.