A Mitogenic Action for Fibrinogen Mediated through Intercellular Adhesion Molecule-1*

Intercellular adhesion molecule-1 (ICAM-1) is a cell surface ligand for αLβ2 and αMβ2 integrins and has a key role in leukocyte adhesion to the vascular endothelium. The plasma protein fibrinogen has also been shown to interact with ICAM-1. We have investigated the effect of fibrinogen binding to ICAM-1-expressing cells on cell proliferation. The inclusion of 200–800 nmfibrinogen but not fibronectin to the culture medium of Raji induced a 2–4-fold increase in [3H]thymidine incorporation after 8 h. Cell proliferation in cultures containing fibrinogen was also confirmed by direct cell counting. The proliferative response in Raji was abrogated by an anti-ICAM-1 mAb 84H10 which maps to the first Ig domain of ICAM-1. A purified truncated form of ICAM-1 containing the first two Ig-like domains and a peptide with amino acid sequence corresponding to ICAM-1 (8–22) was also able to block the proliferative action of fibrinogen on Raji. 200 nmfibrinogen induced a 3-fold increase in [3H]thymidine incorporation by 293 cells transfected with ICAM-1 cDNA but not control non-transfected 293 cells. Comparable mitogenic effects were achieved with fibrinogen fragments X and D100, and with a synthetic peptide with an amino acid sequence matching fibrinogen γ chain (117–133). These results indicate that interaction between discrete sequences within ICAM-1 and fibrinogen result in cellular proliferation.

Intercellular adhesion molecule-1 (ICAM-1) is a cell surface ligand for ␣ L ␤ 2 and ␣ M ␤ 2 integrins and has a key role in leukocyte adhesion to the vascular endothelium. The plasma protein fibrinogen has also been shown to interact with ICAM-1. We have investigated the effect of fibrinogen binding to ICAM-1-expressing cells on cell proliferation. The inclusion of 200 -800 nM fibrinogen but not fibronectin to the culture medium of Raji induced a 2-4-fold increase in [ 3 H]thymidine incorporation after 8 h. Cell proliferation in cultures containing fibrinogen was also confirmed by direct cell counting. The proliferative response in Raji was abrogated by an anti-ICAM-1 mAb 84H10 which maps to the first Ig domain of ICAM-1. A purified truncated form of ICAM-1 containing the first two Ig-like domains and a peptide with amino acid sequence corresponding to ICAM-1 (8 -

22) was also able to block the proliferative action of fibrinogen on Raji. 200 nM fibrinogen induced a 3-fold increase in [ 3 H]thymidine incorporation by 293 cells transfected with ICAM-1 cDNA but not control nontransfected 293 cells. Comparable mitogenic effects
were achieved with fibrinogen fragments X and D 100 , and with a synthetic peptide with an amino acid sequence matching fibrinogen ␥ chain (117-133). These results indicate that interaction between discrete sequences within ICAM-1 and fibrinogen result in cellular proliferation.
Fibrinogen (Fg) 1 and intercellular cell adhesion molecule-1 (ICAM-1) are prominent ligands for several integrin receptor subclasses (1) and have important roles in facilitating cell-cell associations and adhesions (2,3). ICAM-1 functions as an adhesive ligand for leukocytic ␤ 2 integrins and mediates leukocyte extravasation across vascular endothelial barriers. The expression of ICAM-1 on endothelium is controlled at least in part by inflammatory cytokines (2,4). ICAM-1 has five extracellular Ig-like domains and numerous N-glycosylations contributing to a molecular mass of 95 kDa. The first two Ig-like domains are critical for binding to ␣ L ␤ 2 (LFA-1) (5), whereas the third Ig-like domain of ICAM-1 binds to ␣ M ␤ 2 (Mac-1) (6).
Elevated levels of ICAM-1 have been detected on endothelium lining regions of inflamed tissues, endothelium experiencing shear stress (7), and endothelial cells present within atherosclerotic lesions (8,9).
Fg is an abundant plasma protein with a well recognized role in blood coagulation and hemostasis. Both Fg and fibrin are able to bind to a variety of plasma and matrix proteins (10). Fg and Fg degradation products have been localized in early and late stages of formation of atherosclerotic lesions (11,12). Fg is a homodimer of molecular mass 340 kDa, made up of two sets of ␣, ␤, and ␥ polypeptide chains linked through N-terminal regions by interchain disulfide bonds. Upon the proteolytic action of plasmin, Fg is cleaved first to a 260-kDa form of Fg termed fragment X and subsequently to smaller C-terminal Fg fragments D (100 kDa) and a central Fg fragment E (50 kDa) (10). A direct interaction between ICAM-1 and a discrete region of the Fg ␥ chain has been demonstrated (13). The interaction of Fg with ICAM-1 has been localized to the first Ig-like domain incorporating amino acids 8 -21 of ICAM-1 (14), and Fg: ICAM-1 interactions have been shown to mediate contact between cells (3,13). However, the consequences of Fg binding to ICAM-1 may be even more direct. A recent report has described a Fg-mediated vasorelaxation of saphenous vein endothelium that was abrogated by treatment of the tissue with antibodies directed against ICAM-1 (15), which suggests that Fg association with ICAM-1 is able to effect a direct physiological response. Fg and fragments of Fg are demonstrated to induce mitogenic activity in T and B lymphocyte cell lines (16) and hematopoietic progenitor cells (17); however, the mechanism through which Fg mediates this cellular response remains unclear. This report provides evidence that ICAM-1 is a mitogenic Fg receptor on B lymphocytes, and that the observed mitogenic response can be attributed at least in part to discrete amino acid sequences within the Fg ␥ chain and the first Ig-like domain of ICAM-1. Antibodies-Anti-ICAM-1 polyclonal antibodies R686 and R803 were prepared and purified in our laboratory from the serum of rabbits injected with Escherichia coli-expressed ICAM-1. The anti-ICAM-1 mAb 84H10 was from AMAC International (Westbrook, ME). This antibody has been reported to block ␣ L ␤ 2 /ICAM-1-mediated lymphocyte interactions (20 -22), indicating that this antibody recognizes an epitope within the first Ig domain of ICAM-1. The anti-ICAM-1 mAb QE2 (23) was provided by Dr. R. Faull (St. George Hospital, Kogarah, Australia). The anti-␣ v ␤ 3 antibody LM 609 (24) was a gift from Dr. D. Cheresh (Scripps Research Institute, La Jolla, CA). mAb1980, mAb1969, and mAb1965 mAbs directed against the integrin subunits ␣ v , ␣ 5 , and ␤ 1 , respectively, were obtained from Chemicon International * This work was supported by National Institutes of Health Grant HL 43721 and a grant-in-aid from the American Heart Association National Center (to S. E. D.) and a post-doctoral fellowship from the American Heart Association, NE Ohio affiliate (to E. E. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture-ICAM-1-expressing lymphoblastoid Raji cells and 293 cells of human kidney fibroblast origin were obtained from ATCC. Raji were grown in RPMI 1640 (BioWhittaker, Walkersville, MD) containing 7.5% FCS and 1.0 mM glutamine. 293 cells were maintained in Dulbecco's modified Eagle's medium/F-12 (BioWhittaker) containing 10% FCS and 1.0 mM glutamine. For transfection, 293 cells were washed and plated on 60-mm plastic Petri dishes in medium without FCS. Cells were stably transfected using the calcium-phosphate method with 10 -20 g of purified plasmid DNA isolated from E. coli transfected with pCDM8 containing ICAM-1 cDNA (25). Cells expressing ICAM-1 were detected by incubation with mAb QE2 and FITC-conjugated anti-mouse IgG and isolated in a fluorescence-activated cell sorter (FACS). Levels of ICAM-1 expression were routinely monitored by FACS analysis.
FACS Analyses of Cell Surface Expression of Proteins-Control and ICAM-1-expressing 293 cells were isolated from culture flasks by brief trypsin treatment. Raji were isolated from culture medium by centrifugation. Each cell type was washed twice in Dulbecco's phosphatebuffered saline, pH 7.4, and once in Hanks' balanced salt solution without divalent cations then resuspended in a staining medium of Hanks' balanced salt solution containing 2.0 mM CaCl 2 , 2.0 mM MgCl 2 , 25 mM HEPES, pH 7.4, and 0.1% bovine serum albumin. Cells were counted and resuspended at 10 6 cells/ml, and 0.1-ml aliquots were incubated at 4°C for 30 min with 5.0 g/ml control mouse IgG, anti-␣ M mAb OKM1, anti-␣ V ␤ 3 mAbs LM 609 or mAb1980, anti-␣ X mAb KB90, anti-␣ 5 mAb mAb1969, or anti-ICAM-1 mAbs QE2 or 84H10. Cells were centrifuged through a cushion of FCS and resuspended in staining medium containing 50 g/ml FITC-conjugated goat anti-mouse IgG antibodies (Zymed Laboratories, South San Francisco, CA). Cells were incubated for 30 min at 4°C and then centrifuged and resuspended in 0.5 ml of staining medium. Cell-bound antibodies were detected using a FACScan, and analysis was performed using the LYSIS II program (Becton Dickinson, San Jose, CA).
Preparation of Fibrinogen and Fragments of Fibrinogen-Fg was purified from fresh human plasma by cryoethanol precipitation (26,27). Using electrophoretic conditions that allow the separation of fibrin from fibrinogen monomer and subsequent visualization using Coomassie violet R-250 (18), the isolated material was estimated to comprise greater than 95% fibrinogen. Preparations of Fg were also analyzed for the presence of free fibrinopeptides A and B by elution on a Sep-Pak C 18 high pressure liquid chromatography column using standard preparations of each of the fibrinopeptides (Sigma). At protein concentrations of at least 50-fold greater than those used in these experiments, amounts of fibrinopeptides A and B were below detectable levels. Fg was radiolabeled with 125 I-Na using IODO-BEADS (Pierce) (28,29) and extensively dialyzed against 0.3 M NaCl. The specific activity of 125 I-labeled Fg was 0.45-0.52 mCi/mg. Radiolabeled Fg was aliquoted and stored at Ϫ80°C until use.
For the preparation of plasmin-digested Fg fragments (26), purified Fg (5.0 mg/ml) was incubated at 37°C with 8.3 g/ml plasmin for 1 h (fragment X) and 0.1 mg/ml plasmin for 2 h (fragment D 100 ). Each fragment was purified by ion-exchange chromatography on a column of DEAE-cellulose (Pharmacia, Uppsala, Sweden) using an elution buffer of 50 mM Tris/HCl, pH 7.4, containing 0.4 M NaCl and dialyzed extensively against phosphate-buffered saline. Fg fragment E was isolated as a by-product of fragment D preparations by subjecting these preparations to size exclusion chromatography on a column of Sephacryl S200 (Pharmacia, Uppsala, Sweden) equilibrated with a solution of 50 mM Tris/HCl, pH 7.4, containing 0.15 M NaCl. Fractions containing Fg fragment E were dialyzed extensively against distilled water and lyophilized. The molecular weights of plasmin digestion products of Fg were verified by electrophoresis on SDS gels. Protein concentrations were adjusted to 1 mg/ml using Iscove's modified Dulbecco's Eagle's medium containing penicillin, streptomycin, 1.0 mM glutamine, and 5 g/ml transferrin (IMDM) (Life Technologies, Inc.), and stored at Ϫ20°C.
Recombinant Proteins-The full-length five-domain form of ICAM-1 was ligated into the viral vector PVL1392 (Invitrogen, San Diego, CA) and allowed to recombine with baculoviral genomic DNA (14, 30 -32). Truncated ICAM-1 cDNA was generated by polymerase chain reaction using specific primers which annealed to the extreme 5Ј region of the cDNA and a region 3Ј to cDNA encoding the second IgG-like domain. The 3Ј reverse primer contained appropriate sequences for a stop codon and a NotI restriction enzyme site which was unique for cDNA encoding the first two Ig domains of ICAM-1. The nucleotide sequence of the polymerase chain reaction-amplified DNA was confirmed, and the DNA was ligated into the baculovirus vector pBlueBacIII using XbaI and NotI restriction sites. Recombinant virus containing vector DNA was purified by repeated plaque purification and used for infection of Spodoptera frugiperda (Sf9) insect cells. Sf9 cells were grown in Grace's insect cell media (Life Technologies, Inc.) for 4 days, then washed with phosphate-buffered saline, pH 7.4, and lysed in 1% (w/v) octyl glucoside in phosphate-buffered saline, pH 7.4. Expression of intact ICAM-1 or the two domain form of ICAM-1 (D 1 D 2 ICAM-1) was confirmed by enzyme-linked immunosorbent assay and immunoblotting using ICAM-1-specific antibodies as described previously (14).
Full-length ICAM-1 and D 1 D 2 ICAM-1 were purified from Sf9 cell lysates on a Rotofor preparative isoelectric focusing apparatus (Bio-Rad) using a pH gradient of 6 -8. Portions of SDS gels containing similar, unstained material were excised, and ICAM-1 proteins were extracted in a buffer of 0.2 M Tris/HCl, pH 7.4, containing 2% (w/v) SDS and precipitated with 80% (v/v) acetone. The purity of D 1 D 2 ICAM-1 was greater than 90%, as judged by silver staining of SDS gels, and portions of purified D 1 D 2 ICAM-1 were resuspended in IMDM in 1 M aliquots and stored at Ϫ20°C. In adhesion assays, this protein preparation was able to support the adhesion of 51 Cr-labeled peripheral blood lymphocytes and adhesion could be blocked by the anti-␤ 2 antibody MHM23.
Cell Proliferation Assays-Raji or 293 cells were maintained in appropriate medium without FCS for 24 h prior to commencement of an experiment, to achieve cellular quiescence. Cells were washed twice in IMDM and counted with a hemocytometer. Some cells were preincubated for 30 min with control antibodies or antibodies directed against ICAM-1. Aliquots of 0.2 ml of cells (4 ϫ 10 5 cells/ml) were mixed with 0.2-ml protein solutions diluted with IMDM and 4 l of [ 3 H]thymidine (1.0 Ci/l). Cell suspensions (0.1 ml) were aliquoted into four replicate wells of a 96-well flat-bottomed plate (Becton Dickinson, Franklin Lakes, NJ), and plates were stored at 37°C and 5% CO 2 for 4 -24 h. To measure uptake of [ 3 H]thymidine, the contents of each well were transferred to a 96-well plate with v-shaped wells. Cells were isolated from incubation medium by centrifugation (200 ϫ g/10 min) and washed twice in phosphate-buffered saline, pH 7.4. [ 3 H]Thymidine that had been incorporated into cellular DNA was precipitated by addition of 0.1 ml of 20% (v/v) trichloroacetic acid for 2 h at 4°C. Trichloroacetic acid pellets were rinsed once and resolubilized in 2% (w/v) SDS solution and assayed for radioactivity in a ␤-counter.
Fg Binding to Raji-Raji were centrifuged and resuspended three times in Hanks' balanced salt solution and finally in Hanks' balanced salt solution containing 0.1% bovine serum albumin. The number of cells isolated was ascertained using a hemocytometer, and some cell suspensions were preincubated with 25 g/ml TS1/18 anti-␤ 2 integrin antibody or an anti-ICAM-1 antibody 84H10 for 30 min. Increasing amounts of 125 I-labeled Fg were mixed with 0.1 ml of Raji (5 ϫ 10 6 cells/ml) in a final volume of 0.2 ml. Some cell mixtures also containing a 10 -50-fold excess of either unlabeled Fg or transferrin. All cell suspensions were incubated for 30 min at 22°C. Following incubation, free radioactivity was separated from radioactivity bound to the cell surface by centrifugation at 11,300 ϫ g for 2.5 min through an upper phase of 8% sucrose (0.2 ml) and a lower phase (0.1 ml) containing dibutyl phthalate and dimethyl phthalate (Aldrich) mixed at a proportion of 10:1.05 (33). Radioactivity associated with the cell pellet was assayed using a ␥-counter.
Synthetic Peptides-Peptides with amino acid sequences corresponding to regions of ICAM-1 and fibrinogen were synthesized by the Fmoc

An ICAM-1-dependent Induction of a Mitogenic Response by
Fg-These studies were carried out using the Raji B lymphocyte cell line as these cells demonstrate a constitutive and stable expression of ICAM-1, and are readily grown in suspension culture. FACS analysis of Raji using the anti-ICAM-1 mAb QE2 revealed a constitutive and stable expression of ICAM-1 (Fig. 1A, panel a) which represented approximately 75% of the levels of ICAM-1 found on endothelial cells upon treatment with tumor necrosis factor-␣ using mAb QE2 (14). Raji also demonstrated a constitutive expression of ␣ L (panel d); however, binding of Raji by antibodies directed against the ␣ subunits of integrins reported to bind Fg including ␣ V (panel b), ␣ 5 (panel c), ␣ M (panel e), and ␣ X (panel f) could not be detected, suggesting negligible levels of the integrin heterodimers ␣ M ␤ 2 , ␣ X ␤ 2 , ␣ V ␤ 3 , or ␣ 5 ␤ 1 on Raji. As levels of these Fg-binding proteins (34 -37) were below detectable levels, Raji B lymphocytes provided a suitable means of exclusively studying Fg: ICAM-1 interactions. Fg binding to Raji has been previously reported by others (16), and we have recently demonstrated that Raji adhesion to Fg was ICAM-1-dependent (14). This finding was confirmed using a soluble Fg binding assay. A specific and saturable binding of 125 I-labeled Fg to Raji was observed (Fig. 1B). 125 I-Labeled Fg associated with Raji with a K d of 3 ϫ 10 Ϫ7 M and this association could be competitively blocked by an excess of unlabeled Fg but not an equivalent amount of transferrin. This K d value is comparable to other K d values reported for ICAM-1/Fg interactions (3) and Fg interac-tions with Raji cells (16). Analysis of the 125 I-labeled Fg binding data suggested a single class of binding sites on Raji which bound a total of approximately 40,000 molecules/cell. Inclusion of 25 g/ml anti-ICAM-1 antibody 84H10 completely abrogated specific 125 I-labeled Fg binding to Raji, while ␤ 2 integrin mAb TS1/18 failed to block 125 I-labeled Fg binding to Raji.
Experiments were designed to investigate consequences of soluble Fg binding to Raji. When purified human Fg was included in the culture medium of serum-depleted Raji, DNA synthesis was stimulated in a time-and dose-dependent manner as measured by incorporation of [ 3 H]thymidine into macromolecules (Fig. 2, A and B). This effect was specific to Fg as inclusion of equivalent amounts of purified fibronectin or transferrin did not induce any proliferative response in Raji (Fig.  2B) Fig. 2, A and B, confirming that Fg was the mitogenic component.
A putative mitogenic Fg receptor of approximately 92 kDa was isolated from Raji but not conclusively identified (38). As fibrinogen and ICAM-1 are known to interact, and ICAM-1 is present on Raji cell surfaces, we investigated whether the mitogenic Fg receptor on Raji was ICAM-1. Fig. 2C shows that preincubation of Raji with 20 g/ml anti-ICAM-1 mAb 84H10 before the addition of 200 nM Fg resulted in a 91% inhibition of Fg-induced incorporation of [ 3 H]thymidine by these cells. The mAb 84H10 has been reported to block ␣ L ␤ 2 -dependent binding (20,21), and the epitope for 84H10 is located within the first Ig domain of ICAM-1 (22). Inclusion of normal mouse IgG (not shown) or anti-␤ 2 integrin mAb TS1/18 (20 g/ml) had no effect on the Fg-induced uptake of [ 3 H]thymidine by Raji.
Fg Induces a Mitogenic Response in ICAM-1-transfected Cells-To determine whether Fg could invoke a proliferative response in another ICAM-1-expressing cell type, a stably transfected cell line was used in proliferation assays. Fibroblast 293 cells that had been stably transfected with ICAM-1 cDNA and cloned by FACS as described under "Materials and Methods," were maintained for 24 h in serum-free medium. As non-transfected 293 cells have detectable levels of integrin ␣ 5 ␤ 1 (39), which has been reported to mediate fibrinogen binding to endothelial cells (37), we included non-transfected 293 cells for comparison of fibrinogen-mediated cell proliferation. FACS analysis of cell surface levels of ICAM-1 on transfected cells using either control IgG or the anti-ICAM-1 mAb QE2 and an FITC-conjugated disclosing antibody as described under "Materials and Methods" confirmed the stable expression of ICAM-1 (Fig. 3A, panel b) with mean fluorescence intensities approaching levels observed on Raji. No measurable amounts of ICAM-1 (Fig. 3A, panel a; Ref. 39) could be detected on non-transfected 293 cells.
ICAM-1-transfected 293 cells and control non-transfected cells were removed from tissue culture flasks, washed, and placed in culture medium containing 0 -800 nM Fg and [ 3 H]thymidine for 8 h. Fig. 3B shows that in the presence of increasing The First Ig-like Domain of ICAM-1 Mediates Fg-induced Raji Proliferation-In our previous report (14), a truncated form of ICAM-1 containing the first two Ig domains (D 1 D 2 ICAM-1) was determined to be able to interact with Fg. Because of the importance of this segment of ICAM-1 for associations with Fg, we investigated whether this region of ICAM-1 was important for the Fg-mediated proliferative response of Raji. The cDNA encoding for a truncated form of ICAM-1 that contained the first two Ig-like domains was expressed in Sf9 cells and purified as described under "Materials and Methods." D 1 D 2 ICAM-1 migrated on an SDS gel with an apparent molecular mass of 35 kDa and was recognized by R686 anti-ICAM-1 polyclonal antibody by immunoblotting (Fig. 4A) and enzyme-linked immunosorbent assay (data not included). In addition, D 1 D 2 ICAM-1 supported the adhesion of purified peripheral blood lymphocytes. This adhesion was ␤ 2 -integrindependent as it could be blocked by inclusion of antibody MHM23. When Fg was preincubated at 37°C with increasing amounts of D 1 D 2 ICAM-1, Fg-induced proliferation was blocked in a dose-dependent manner (Fig. 4B). Approximately 350 nM D 1 D 2 ICAM-1 abolished the proliferative action of Fg (Fig. 4B). Control incubations containing Fg treated with equivalent amounts of material purified from Sf9 cell lysates that contained no anti-ICAM-1 immunoreactivity, had no effect on Fginduced Raji cell proliferation. The association of a 125 I-labeled D 1 D 2 ICAM-1 preparation with Raji cells could not be detected in cell binding assays suggesting that this fragment of ICAM-1 did not influence Fg-mediated proliferation by binding directly to Raji; rather, D 1 D 2 ICAM-1 competed with cell surface ICAM-1 for binding sites on Fg. Furthermore, the inability of soluble ICAM-1 to bind to integrin ␣ L ␤ 2 has been reported (40,41).
Previous work from our laboratory has demonstrated that Fg is able to interact with amino acids 8 -22 of ICAM-1, which lie within the first Ig domain of ICAM-1 (14). To further define the regions of ICAM-1 that mediated the proliferative action of Fg, Fg was preincubated with synthetic peptides with sequences corresponding to specific regions of ICAM-1 and used in proliferation assays. Fig. 5 shows that the mitogenic action of 200 nM Fg could be reduced by up to 80% by preincubation of Fg with 300 -500 M amounts of ICAM-1 (8 -22) but not ICAM-1 (130 -139). Concentrations of 100 M ICAM-1 (8 -22) were ineffective in blocking mitogenesis; however, we have noted that such concentrations of this peptide were effective in blocking Raji adhesion to Fg (14). Incubation of Raji with medium containing 500 M ICAM-1 (8 -22) alone did not affect the viability of Raji as judged by uptake of [ 3 H]thymidine by the cells and by trypan blue exclusion (results not shown), suggesting that the peptide was acting externally. In summation these results suggest that the Fg-induced proliferative response in Raji is mediated by ICAM-1, specifically by at least one region within the first IgG-like domain of ICAM-1 containing amino acids 8 -22. A Discrete Region of Fg Contains Mitogenic Activity-We investigated whether the ICAM-1-mediated proliferative property of Fg could be attributed to a portion of Fg. A preparation of Fg was treated with plasmin as described under "Materials and Methods," and the resultant Fg degradation products were purified by ion-exchange chromatography and used in Raji proliferation assays. A similar induction of Raji proliferation could be demonstrated using purified Fg plasmin-degradation products fragment X or fragment D but not fragment E (Fig. 6), suggesting that at least one mitogenic site resided within a region of Fg common to fragments X and D. This is in good agreement with previous studies that report an association between Fg and Raji cell surfaces (16), and a mitogenic proliferation of Raji cells mediated by Fg fragment D (38,42).
At least one region of Fg has been reported to be important for interactions with ICAM-1. The Fg ␥ chain sequence 117-133 was reported to block the Fg-mediated adhesion of leukocytic cells to endothelial cell monolayers and was shown to interact directly with ICAM-1-expressing transfectants (13). We investigated whether this region of Fg encompassing amino acids 117-133 of the ␥ chain could induce a proliferative response in Raji. The peptide ␥(117-133) was compared with Fg for mitogenic activity in Raji proliferation assays. Fig. 7 shows that increasing concentrations of ␥(117-133) in Raji incubation medium were able to increase incorporation of [ 3 H]thymidine by Raji after 8 h. Concentrations of 100 M ␥(117-133) induced a response similar to the observed proliferation of Raji in medium containing 200 nM Fg over the same period of incubation. Control peptides ␥(117-133) scrambled and ␥(124 -133) had no discernible proliferative activity, suggesting that the amino acid sequence and in particular the NH 2 -terminal portion of ␥(117-133) is important for the proliferative activity of this peptide. Inclusion of either the RGD-containing peptide Fg A␣ (571-576) (Fig. 7) or fibrinopeptides A or B (data not included) also had no effect on proliferation of Raji. DISCUSSION This paper documents evidence that ICAM-1 is the receptor responsible for the Fg-mediated mitogenic response in Raji. The mitogenic effect of Fg and Fg fragment D on Raji has been reported (16,42); however, the receptor responsible for this effect has not been clearly defined. The present studies identify ICAM-1 as the candidate protein and delineate the sites within ICAM-1 and Fg that mediate this effect. In addition, our results indicate that Fg mediates a similar mitogenic response in ICAM-1-transfected 293 cells, indicating that other non-lymphoid cells have the internal signaling machinery required for Fg-induced mitogenesis. Previously, the exposure of the B␤  region of fibrin was demonstrated to enhance proliferation of fibroblasts and endothelial cells (43) and the A␣ and B␤ chains of fibrinogen have been reported to induce proliferation in fibroblasts (44,45). A 60-kDa protein resembling calreticulin was demonstrated to be a receptor for the B␤ chain of Fg, and antibodies against calreticulin could inhibit the observed mitogenic response (45). In a separate report, a 130-kDa protein was isolated from endothelial cell lysates by elution from a Sepharose-Fg B␤  affinity column (46). This candidate B␤ receptor was not identified; however, it was apparently not related to PECAM-1 or any known endothelial cell integrin ␣ subunits. Fibrinopeptide A has been reported to induce proliferation in pleural mesothelial cells (47); however, under experimental conditions used for our studies, levels of fibrinopeptides A and B comparable to those used in studies by others did not induce a mitogenic response in Raji. Interestingly, in a report of proliferation of phorbol ester-stimulated tonsillar B lymphocytes induced by mAbs against CD11c/CD18 (␣ X ␤ 2 ), occupancy of the receptor by Fg blocked the mitogenic response in these cells (48).
We have previously reported that the first two Ig domains of ICAM-1 bind directly to Fg (14), and this portion of ICAM-1 contains sites necessary for interaction with ␣ L ␤ 2 , rhinovirus (49 -51), and Plasmodium falciparum-infected erythrocytes (22,52). Results presented here demonstrate that mAb 84H10 which maps to the first Ig domain of ICAM-1 blocked Fg binding to Raji (Fig. 1B) and blocked Fg-induced mitogenesis (Fig.  2C). Consistent with these results, D 1 D 2 ICAM-1 protein specifically blocked Fg-mediated mitogenesis in Raji (Fig. 4B), suggesting that this region of ICAM-1 is also important for initiation of the cellular response to Fg binding. In previous studies, the region of ICAM-1 spanning amino acids 8 -21 was determined to be involved in Fg recognition (14). We used ICAM-1 (8 -22) in the present studies as this peptide has superior solubility properties and has Fg-binding properties similar to ICAM-1 (8 -21). At a concentration of 300 and 500 M, comparable to amounts required to block Raji adhesion to Fg, ICAM-1 (8 -22) also blocked Fg-mediated Raji proliferation (Fig. 6). Lower concentrations of ICAM-1 (8 -22) did not inhibit Fg-mediated mitogenesis, but were able to block adhesion of Raji to Fg bound to plastic (14), implying that ICAM-1-binding sites are more readily accessible within immobilized Fg. Therefore ICAM-1 (8 -22) is implicated in the Fg-mediated response of Raji. The control peptide with amino acid sequence corresponding to ICAM-1 (130 -139) was chosen as it contains a charged, hydrophilic NH 2 -terminal region similar to ICAM-1 (8 -22). Computer modeling of this region of ICAM-1 suggests that several amino acid residues in this defined region are accessible to solvents (14), and ongoing studies expressing recombinant ICAM-1 with mutations in this region will confirm the involvement of this region in mitogenesis.
The physiological degradation of Fg and fibrin by plasmin is key to the process of fibrinolysis and breakdown of a thrombus (10). Purified Fg degradative products fragment X and fragment D 100 also induced a mitogenic response in Raji (Fig. 6). It is important to note that even though the precise affinities of fragments X and D to ICAM-1 have not been reported, the estimated affinity of fragment X for ICAM-1 (8 -21) appears comparable to that of Fg and that fragment D is only 5-fold lower than Fg (14). This is in contrast to the reported affinities of fragments X and D for ␣ IIb ␤ 3 , which are 10-and 100-fold less than Fg, respectively (53). Consistent with the apparent binding affinity of ICAM-1 for Fg and Fg fragments X and D, we observed that the extent of mitogenesis induced by fragment X was at least equal to that by Fg while mitogenesis induced by fragment D was approximately 80% of that observed using intact Fg.
The Fg ␥ chain sequence ␥(117-133) is reported to interfere with Fg binding to ICAM-1, and this region of Fg is directly involved in the transmigration of leukocytes across endothelial cells (13). In our previous studies, ␥(117-133) blocked Fg binding to ICAM-1 (8 -21), suggesting that ␥(117-133) and ICAM-1 (8 -21) may be an important reactive pair in Fg-ICAM-1 interactions (14). The ␥(117-133) peptide was able to induce a mitogenic response in Raji at concentrations of 50 M (Fig. 7), suggesting that this region of the Fg ␥ chain contains a novel mitogenic capacity. This sequence resides within the CS region of Fg and is likely to be at least partially shielded by the C-terminal portion of the A␣ chain, A␣(242-611). Support for this model comes from the observation that the region immediately adjacent, ␥(112-119), is inaccessible to mAb 9F9 in soluble Fg but becomes exposed by proteolytic removal of the C-terminal aspect of the A␣ chain (54). Other groups have reported an association between soluble Fg and ICAM-1 expressing cells (13), and in this report we demonstrate an association between soluble Fg and Raji that was blocked by an anti-ICAM-1 antibody. It is possible that the association between soluble Fg and ICAM-1 through ␥(117-133) forms part of a multistep process. However, it remains to be determined whether some proteolytic processing of soluble Fg is required for complete exposure of the ␥(117-133) sequence.
Elevated levels of ICAM-1 are found on inflamed tissues and ICAM-1 is present in atherosclerotic lesions, and within these regions fibrin, Fg, and Fg degradation products have been detected. As levels of Fg circulating in human plasma normally approach 3.0 mg/ml, it may be anticipated that the ICAM-1mediated mitogenic activity of Fg in vivo is modulated by levels of expression of ICAM-1 on cells and therefore would not occur under normal physiological conditions. However, under conditions of both acute and chronic inflammation, elevated levels of ICAM-1 on endothelial cells and smooth muscle cells (8,55) may mediate a Fg-dependent proliferation of cells. This remains to be determined experimentally; however, a combined role for ICAM-1 and Fg in the adhesion and arrest of infiltrating inflammatory cells within such sites which contributes to the observed pathology of these lesions has been proposed (3).