JOURNAL OF CEI,T,UT,AR PHYSTO1,OGY 153:187-195 (1992)
Neutrophil Adhesion to TN Fa-Activated Endothelial Cells Potentiates Leukotriene B, Production YOSHlKl ISHII, SIU K. LO, AND ASRAR B. M A L I K * Departments of Physiology and Cell Biology, The Albany Medical College, Albany, NY 12208 (Y./., A.B.M.), and Departmrnt of Pathology, Vascular Kcsrarch Division, Brigham and Women's Hospital, Boston, Massachusetts 02 1 1.5 (S.K.L.) Since adhesion of neutrophils (PMN) to endothelial cells may influence PMN activation responses, we examined whether adhesion of PMN to TNF,,-activated human umbilical vein endothelial cells (HUVEC) stimulates leukotrienc B, (LTB,) production. Endothelial adhesivity towards PMN increased after HUVEC pretrestment with TNF,, for 4 h. LTB, production increased markedly in response to stimulation with arachidonic acid (20 pMj when PMN were added to the hyperadhesive HIJVEC. In contrast, stimulation of PMN in suspension did not potentiate LTB, production. LTB, production persisted when PMN were applied to TNF,-pretreated HUVEC fixed with 1 Yo paraformaldehyde excluding the possibilily that metabolic activity of endothelium participates in this response. PMN adhesion to plastic and gelatin also enhanced LTB, indicating that adhesion was a critical event in inducing LTB, production. Wc used monoclonal antibodies (mAb) to adhesion molecules on endothelial cells (i.e., endothelial leukocyte adhesion molecule-l (ELAM-1) and interccllular adhesion molccule-1 (ICAM-1)) or on PMN (CD18) to asses5 the role of PMN adhesion to the activated endothelium on LTB, potentiation. Both anti-ELAM-1 mAb and anti-ICAM-l mAb inhibited PMN adhesion (by 55 and 41 "/.,respectively) as well as LTB, production (by 65 and 50%, respectively). Anti-CD18 rnAb also reduced the adhesion (65%) and the LTB, production (66%). Furtherniore, combination of anti-ELAM-1 mAb (H18/7) and anti-ICAM-l mAb ( R K l I 1 ) or of anti-LLAM-1 rnAb (H18/7) and antiCD18 mAb (IB4) had an additive effect in inhibiting both PMN adhesion as well as LTB, production. PMN adherence to immobolized recombinant soluble rELAM-1 or rlCAM-1 also increased LTB, production, which was prevented with relevant mAbs. However, neither rELAM-1 nor rlCAM-1 stimulated LTB, production of PMN in suspension. We conclude that PMN adhe5ion to TNF,-stimulated endothelial cells rnhances LTB, production by PMN, a response activated by binding of PMN to expressed endothelial cell surface adhesion molecules. @ 1932 Wiley-Liss, Inc
Adhesion of PMN to vascular endothelial cells is an initial step in the inflammatory response. PMN adhesion enables PMN to migrate to the inflammatory site (Moser et al., 1989) and plays an important role in activation of PMN (Dahinden et al., 1983; Fehr et al., 1985; Lo et al., 1991; Nathan et al., 1989; Richter et al., 1990). In recent years, a number of adhesion molecules accounting for PMN adhesion to endothelial cells have come to light. Cytokines such as tumor necrosis factor-a (TNF,) and interleukin-1 (IL-I) increase endothelial adhesiveness by upregulation of endothelial adhesion molecules (Luscinskas et al., 1991; Pober et al., 1986; Smith et al., 1988). Endothelial-leukocyte adhesion molecule-1 (ELAM-1) is inducible on the endothelial surface by cytokines and it binds to specific carbohydrate residues (i.e., sialyl-Lewis x and sialyl Lewis a antigens) on PMN (Bevilacqua et al., 1989; Phillips et al., 1991). The intercellular adhesion molecule-1 (ICAM-1) is also upregulated by cytokines and it pri0 1992 WILEY-LISS, INC.
marily interacts with integrins (i.e., CD18 family) expressed on PMN (Diamond et al., 1990; Smith et al., 1988). ELAM-1 expression is maximum in endothelial cell cultures 4 h after TNF,, activation, whereas ICAM-1 appears t o increase gradually over a 24 h period (Bevilacqua et al., 1989; Luscinskas et al., 1991; Pober et al., 1986). These two adhesion molecules on endothelial cells are essential for PMN adhesion and
Received March 16,1992; accepted May 18,1992
*To whom reprint requestsicorrespondence should be addressed. Dr. Yoshiki Ishii's permanent address is Department of Pulmonary Medicine, Jichi Medical School, Tochigi, Japan 329-04. Dr. Siu K. Lo's current address is Department of Medicine, Division of Hematology and Oncology, Cornell University Medical College, 1300 York Avenue, New York, NY 10021.
188
ISHII ET AL.
migration responses (Luscinskas et al., 1991; Smith et al., 1988). Studies have shown that activation of PMN is dependent on adhesion of PMN to endothelial cells or to specific matrix proteins. The release of H,O, from PMN induced by TNF, stimulation was linked to C D l l i CD18-mediated PMN adherence to matrix proteins (Nathan et al., 1989). TNF,-induced degranulation of adherent human PMN to matrix protein was also dependent on CDllb/CD18 integrin-triggered oscillation of cytosolic free calcium (Richter e t al., 1990).Recently, Lo et al. (1991) have shown that CDllblCD18 activity on PMN is increased by binding of ELAM-1 to its PMN receptor. Another important function of PMN is the production of leukotriene B, (LTB,), a potent mediator involved in the inflammatory response. LTB, mediates PMN chemotaxis and degranulation and induces the production of oxygen radicals (Ford-Hutchinson et al., 1980; Hafstrom et al., 1981). In the present study, we examined whether adhesion of PMN to TNF,-activated endothelial cells signals LTB, production from PMN and whether the LTB, response is dependent on the binding of PMN to adhesion molecules.
MATERIALS AND METHODS Reagents Dulbecco's modified Eagle's medium (DMEM), RPMI 1640, Hanks' balanced salt solution (HBSS), heparin sodium, and fetal bovine serum were obtained from Gibco (Grand Island, NY). Dextran-70 and Ficoll-paque were purchased from Pharmacia (Piscataway, NJ). HPLC grade acetonitrile, methanol, and hexane were obtained from Baxter Healthcare Corp. (Muskegon, MI), and ethanol was purchased from Pharmco Products (Bayonne, NJ). Paraformaldehyde was obtained from J.T. Baker Chemical (Phillipsburg, NJ). Arachidonic acid metabolite standards (LTB,, 20-hydroxyLTB,, 20-carboxy-LTB4, and prostaglandin B,) were purchased from Biomol Research Laboratories (Plymouth Meeting, PA). Arachidonic acid was obtained from Sigma Chemical Co. (St. Louis, MO). Human recombinant TNF, with a specific bioactivity of 25 x 10" U/mg protein was obtained from Cetus Corporation (Emeryville, CA). The endotoxin contamination of TNF, was 0.05 ng/ml by the limulus amebocyte lysate assay. Monoclonal antibodies mAb IB4 (anti-CD18, CgG2a) (Wright et al., 1983) and mAb OKMl (anti-CDllb, IgG2b) (Wright et al., 1983) were provided by Dr. S.D. Wright, Rockefeller University (New York, NY). mAb R R l i l (anti-ICAM-1, IgGl) (Marlin and Springer, 1987; Smith et al., 1988) was provided by Dr. R. Rothlein, Boehringer Ingelheim (Ridgefield, CT). mAb H1817 (anti-ELAM-1, IgG2a) (Bevilacqua et al., 1989) and mAb W6132 (anti-class I histocompatibility antigen, IgG2a) (Barstable et al., 1978) were provided by Dr. M.P. Bevilacqua, University of California School of Medicine (San Diego, CAI. To avoid possible Fc-dependent functional activation (Leeuwenberg e t al., 1990; Macintyre et al., 19891, all antibodies were prepared a s F(ab'), fragments by pep-
sin digestion (Leeuwenberg et al., 1990) and their purity was checked by SDS-polyacrylamide gels.
Isolation of PMN Whole blood was collected from healthy donors by venipuncture into acid-citrate-dextrose and centrifuged to obtain a buffy coat. The buffy coat was separated into PMN-rich fraction by dextran sedimentation, followed by Ficoll-paque gradient centrifugation, and hypotonic lysis of erythrocytes. The washed PMN were suspended in HBSS (without Mg2+and Ca", with 20 mM Hepes, pH 7.4). The purity of PMN was >97% and their viability was >98%. Culture of endothelial cells Human umbilical vein endothelial cells (HUVEC) were prepared by collagenase treatment as described (Obrig et al., 1988) and cultured in RPMI 1640 with 20% fetal bovine serum, retinal-derived growth factor (6.5 pg/ml) (prepared as described in Glasser et al., 1980), and heparin (75 pgiml). Cells in third to fifth passage were plated in 60 mm dishes for measurement LTB, generation and in 24-well plates for the PMN adhesion assay. All culture surfaces were coated with 0.2% gelatin to facilitate endothelial cell adhesion and growth. Neutrophil labeling Freshly isolated PMN were labeled with 51Cr (New England Nuclear, Boston, MA) as described (Bizios et al., 1988). PMN suspension in HBSS containing BSA (1 mgiml) were incubated with 51Cr (0.2-0.5 pCi per lo6 PMN) for 60 min a t 37°C. Free "Cr was removed by washing 3 times with HBSS and the pelleted PMN from the last centrifugation (450g, 5 min, 4°C) were resuspended in a n appropriate volume of DME to yield a cell concentration of 2 x lo7 cellsiml. After labeling, over 98% of the PMN were found to exclude trypan blue and these cells showed minimal basal adherence (
Neutrophil Adhesion to TN Fa-Activated Endothelial Cells Potentiates Leukotriene B, Production YOSHlKl ISHII, SIU K. LO, AND ASRAR B. M A L I K * Departments of Physiology and Cell Biology, The Albany Medical College, Albany, NY 12208 (Y./., A.B.M.), and Departmrnt of Pathology, Vascular Kcsrarch Division, Brigham and Women's Hospital, Boston, Massachusetts 02 1 1.5 (S.K.L.) Since adhesion of neutrophils (PMN) to endothelial cells may influence PMN activation responses, we examined whether adhesion of PMN to TNF,,-activated human umbilical vein endothelial cells (HUVEC) stimulates leukotrienc B, (LTB,) production. Endothelial adhesivity towards PMN increased after HUVEC pretrestment with TNF,, for 4 h. LTB, production increased markedly in response to stimulation with arachidonic acid (20 pMj when PMN were added to the hyperadhesive HIJVEC. In contrast, stimulation of PMN in suspension did not potentiate LTB, production. LTB, production persisted when PMN were applied to TNF,-pretreated HUVEC fixed with 1 Yo paraformaldehyde excluding the possibilily that metabolic activity of endothelium participates in this response. PMN adhesion to plastic and gelatin also enhanced LTB, indicating that adhesion was a critical event in inducing LTB, production. Wc used monoclonal antibodies (mAb) to adhesion molecules on endothelial cells (i.e., endothelial leukocyte adhesion molecule-l (ELAM-1) and interccllular adhesion molccule-1 (ICAM-1)) or on PMN (CD18) to asses5 the role of PMN adhesion to the activated endothelium on LTB, potentiation. Both anti-ELAM-1 mAb and anti-ICAM-l mAb inhibited PMN adhesion (by 55 and 41 "/.,respectively) as well as LTB, production (by 65 and 50%, respectively). Anti-CD18 rnAb also reduced the adhesion (65%) and the LTB, production (66%). Furtherniore, combination of anti-ELAM-1 mAb (H18/7) and anti-ICAM-l mAb ( R K l I 1 ) or of anti-LLAM-1 rnAb (H18/7) and antiCD18 mAb (IB4) had an additive effect in inhibiting both PMN adhesion as well as LTB, production. PMN adherence to immobolized recombinant soluble rELAM-1 or rlCAM-1 also increased LTB, production, which was prevented with relevant mAbs. However, neither rELAM-1 nor rlCAM-1 stimulated LTB, production of PMN in suspension. We conclude that PMN adhe5ion to TNF,-stimulated endothelial cells rnhances LTB, production by PMN, a response activated by binding of PMN to expressed endothelial cell surface adhesion molecules. @ 1932 Wiley-Liss, Inc
Adhesion of PMN to vascular endothelial cells is an initial step in the inflammatory response. PMN adhesion enables PMN to migrate to the inflammatory site (Moser et al., 1989) and plays an important role in activation of PMN (Dahinden et al., 1983; Fehr et al., 1985; Lo et al., 1991; Nathan et al., 1989; Richter et al., 1990). In recent years, a number of adhesion molecules accounting for PMN adhesion to endothelial cells have come to light. Cytokines such as tumor necrosis factor-a (TNF,) and interleukin-1 (IL-I) increase endothelial adhesiveness by upregulation of endothelial adhesion molecules (Luscinskas et al., 1991; Pober et al., 1986; Smith et al., 1988). Endothelial-leukocyte adhesion molecule-1 (ELAM-1) is inducible on the endothelial surface by cytokines and it binds to specific carbohydrate residues (i.e., sialyl-Lewis x and sialyl Lewis a antigens) on PMN (Bevilacqua et al., 1989; Phillips et al., 1991). The intercellular adhesion molecule-1 (ICAM-1) is also upregulated by cytokines and it pri0 1992 WILEY-LISS, INC.
marily interacts with integrins (i.e., CD18 family) expressed on PMN (Diamond et al., 1990; Smith et al., 1988). ELAM-1 expression is maximum in endothelial cell cultures 4 h after TNF,, activation, whereas ICAM-1 appears t o increase gradually over a 24 h period (Bevilacqua et al., 1989; Luscinskas et al., 1991; Pober et al., 1986). These two adhesion molecules on endothelial cells are essential for PMN adhesion and
Received March 16,1992; accepted May 18,1992
*To whom reprint requestsicorrespondence should be addressed. Dr. Yoshiki Ishii's permanent address is Department of Pulmonary Medicine, Jichi Medical School, Tochigi, Japan 329-04. Dr. Siu K. Lo's current address is Department of Medicine, Division of Hematology and Oncology, Cornell University Medical College, 1300 York Avenue, New York, NY 10021.
188
ISHII ET AL.
migration responses (Luscinskas et al., 1991; Smith et al., 1988). Studies have shown that activation of PMN is dependent on adhesion of PMN to endothelial cells or to specific matrix proteins. The release of H,O, from PMN induced by TNF, stimulation was linked to C D l l i CD18-mediated PMN adherence to matrix proteins (Nathan et al., 1989). TNF,-induced degranulation of adherent human PMN to matrix protein was also dependent on CDllb/CD18 integrin-triggered oscillation of cytosolic free calcium (Richter e t al., 1990).Recently, Lo et al. (1991) have shown that CDllblCD18 activity on PMN is increased by binding of ELAM-1 to its PMN receptor. Another important function of PMN is the production of leukotriene B, (LTB,), a potent mediator involved in the inflammatory response. LTB, mediates PMN chemotaxis and degranulation and induces the production of oxygen radicals (Ford-Hutchinson et al., 1980; Hafstrom et al., 1981). In the present study, we examined whether adhesion of PMN to TNF,-activated endothelial cells signals LTB, production from PMN and whether the LTB, response is dependent on the binding of PMN to adhesion molecules.
MATERIALS AND METHODS Reagents Dulbecco's modified Eagle's medium (DMEM), RPMI 1640, Hanks' balanced salt solution (HBSS), heparin sodium, and fetal bovine serum were obtained from Gibco (Grand Island, NY). Dextran-70 and Ficoll-paque were purchased from Pharmacia (Piscataway, NJ). HPLC grade acetonitrile, methanol, and hexane were obtained from Baxter Healthcare Corp. (Muskegon, MI), and ethanol was purchased from Pharmco Products (Bayonne, NJ). Paraformaldehyde was obtained from J.T. Baker Chemical (Phillipsburg, NJ). Arachidonic acid metabolite standards (LTB,, 20-hydroxyLTB,, 20-carboxy-LTB4, and prostaglandin B,) were purchased from Biomol Research Laboratories (Plymouth Meeting, PA). Arachidonic acid was obtained from Sigma Chemical Co. (St. Louis, MO). Human recombinant TNF, with a specific bioactivity of 25 x 10" U/mg protein was obtained from Cetus Corporation (Emeryville, CA). The endotoxin contamination of TNF, was 0.05 ng/ml by the limulus amebocyte lysate assay. Monoclonal antibodies mAb IB4 (anti-CD18, CgG2a) (Wright et al., 1983) and mAb OKMl (anti-CDllb, IgG2b) (Wright et al., 1983) were provided by Dr. S.D. Wright, Rockefeller University (New York, NY). mAb R R l i l (anti-ICAM-1, IgGl) (Marlin and Springer, 1987; Smith et al., 1988) was provided by Dr. R. Rothlein, Boehringer Ingelheim (Ridgefield, CT). mAb H1817 (anti-ELAM-1, IgG2a) (Bevilacqua et al., 1989) and mAb W6132 (anti-class I histocompatibility antigen, IgG2a) (Barstable et al., 1978) were provided by Dr. M.P. Bevilacqua, University of California School of Medicine (San Diego, CAI. To avoid possible Fc-dependent functional activation (Leeuwenberg e t al., 1990; Macintyre et al., 19891, all antibodies were prepared a s F(ab'), fragments by pep-
sin digestion (Leeuwenberg et al., 1990) and their purity was checked by SDS-polyacrylamide gels.
Isolation of PMN Whole blood was collected from healthy donors by venipuncture into acid-citrate-dextrose and centrifuged to obtain a buffy coat. The buffy coat was separated into PMN-rich fraction by dextran sedimentation, followed by Ficoll-paque gradient centrifugation, and hypotonic lysis of erythrocytes. The washed PMN were suspended in HBSS (without Mg2+and Ca", with 20 mM Hepes, pH 7.4). The purity of PMN was >97% and their viability was >98%. Culture of endothelial cells Human umbilical vein endothelial cells (HUVEC) were prepared by collagenase treatment as described (Obrig et al., 1988) and cultured in RPMI 1640 with 20% fetal bovine serum, retinal-derived growth factor (6.5 pg/ml) (prepared as described in Glasser et al., 1980), and heparin (75 pgiml). Cells in third to fifth passage were plated in 60 mm dishes for measurement LTB, generation and in 24-well plates for the PMN adhesion assay. All culture surfaces were coated with 0.2% gelatin to facilitate endothelial cell adhesion and growth. Neutrophil labeling Freshly isolated PMN were labeled with 51Cr (New England Nuclear, Boston, MA) as described (Bizios et al., 1988). PMN suspension in HBSS containing BSA (1 mgiml) were incubated with 51Cr (0.2-0.5 pCi per lo6 PMN) for 60 min a t 37°C. Free "Cr was removed by washing 3 times with HBSS and the pelleted PMN from the last centrifugation (450g, 5 min, 4°C) were resuspended in a n appropriate volume of DME to yield a cell concentration of 2 x lo7 cellsiml. After labeling, over 98% of the PMN were found to exclude trypan blue and these cells showed minimal basal adherence (
