Intercellular adhesion molecule (ICAM)-1, ICAM-2, and ICAM-3 have been identified as counter-receptors for the leukocyte integrin lymphocyte function-associated antigen 1 (LFA-1). The other leukocyte integrins, Mac-1 and p150,95, also interact with ICAM-1. ICAM-1 and ICAM-3 are highly homologous, and an undefined ligand for Mac-1 is present on neutrophils where ICAM-3 is well expressed. In addition, glycosylation has been shown to affect the interaction of ICAM-1 with Mac-1. We therefore sought to characterize ICAM-3 heterogeneity and determine whether ICAM-3 was a ligand for either Mac-1 or p150,95. Despite extensive differences in N-linked glycosylation, ICAM-3 purified from lymphoid cells and from neutrophils supports adhesion of LFA-1-bearing cells equally well; however, neither supports adhesion of Mac-1 or p150,95-expressing chinese hamster ovary cell transfectants. Similarly, purified Mac-1 does not support adhesion of ICAM-2 or ICAM-3-expressing L cell transfectants. ICAM-3 on neutrophils does not participate in Mac-1-dependent homotypic aggregation. Thus, ICAM-3 is not a counter-receptor for either Mac-1 or p150,95.

The beta 2-integrin (CD18) family members bind to their ligands subsequent to activation of a number of well defined and diverse signal transduction pathways. The precise molecular changes associated with activation of the integrin family members have remained elusive. Here, we characterize a monoclonal, CBR LFA-1/2, that binds to the beta 2-subunit and is able to mimic activation induced upon stimulation by phorbol esters. The Ab induces binding of the LFA-1-expressing cell line, JY, to ICAM-1 (CD54) and ICAM-3 (CD50). Activation of binding by this Ab is independent of Fc interactions and does not occur through cross-linking at the cell surface, because the Fab fragment of the Ab is able to modulate the same effect. Stimulation of neutrophils with CBR LFA-1/2 induces binding to ICAM-1 through activation of both LFA-1 and Mac-1. Activation of Mac-1 by CBR LFA-1/2 was further confirmed by stimulation of neutrophil binding to fibrinogen, a ligand for Mac-1. CBR LFA-1/2 lowers by 10-fold the concentration of Mg2+ required to achieve maximal binding of LFA-1 to ICAM-1. It therefore appears that CBR LFA-1/2 induces a conformational change that directly increases the avidity of beta 2-integrins for ligands.

The mAb YTA-1, which brightly stains CD3-CD16+ large granular lymphocytes (LGL)/natural killer (NK) cells and CD8+ T cells by immunofluorescence, is specific for leukocyte function-associated antigen (LFA)-1. Some mAbs recognizing the LFA-1 alpha chain (CD11a) or LFA-1 beta chain (CD18) inhibited the binding of YTA-1 to peripheral blood mononuclear cells. YTA-1 mAb could be chemically cross-linked to 170 and 96 kDa molecules, whose molecular weights correspond to those of LFA-1 alpha and beta respectively. YTA-1 bound to COS-7 cells co-transfected with CD11a and CD18 cDNAs, but not to untransfected cells. Reactivities of YTA-1 to K562 cells transfected with LFA-1 alpha and beta (CD11a/CD18) cDNAs and to CHO cells transfected with Mac-1 (CD11b/CD18) or p150, 95 (CD11c/CD18) cDNAs strongly suggest that YTA-1 recognizes either LFA-1 alpha or an epitope formed by a combination of LFA-1 alpha and beta. Treatment of fresh CD3-CD16+ LGL with YTA-1 augmented cytolytic activity and induced proliferation. F(ab')2 fragments of YTA-1 augmented NK cytotoxicity, indicating that the NK activating signal was transmitted through LFA-1 without involvement of Fc gamma receptor III. In contrast, the other mAbs against LFA-1 could not activate NK cells. These results collectively indicate that YTA-1 recognizes a unique epitope of LFA-1, which is involved in activation of fresh NK cells.

Previous studies have shown that lymphocyte function-associated antigen-1 (LFA-1) molecules containing the human alpha (CD11a) and human beta (CD18) subunits but not the murine alpha and human beta subunits can bind to human intercellular adhesion molecule 1 (ICAM-1). Using human/mouse LFA-1 alpha subunit chimeras, we mapped regions required for binding to ICAM-1 N-terminal to amino acid (aa) residue 350. Ligand binding sites were mapped in greater detail by scanning this region with murine sequences from 56 down to 17 aa in length and finally by introducing single or few murine aa residue replacements into the human sequence. Replacement of two non-contiguous regions of aa residues 119-153 and 218-248 in the me domain with the corresponding mouse sequences abolished most binding to human ICAM-1, without affecting alpha beta subunit association or expression on the surface of transfected COS cells. Specific residues within the I domain found to be important were Met-140, Glu-146, Thr-243, and Ser-245. Using the recently solved structure of the Mac-1 (CD11b) I domain as a model (Lee, J.-O., Rieu, P., Arnaout, M.A., and Liddington, R. (1995) Cell 80, 631-638), these residues are shown to be located on the surface of the I domain surrounding the site to which Mg2+ is chelated, and fine a ligand binding interface. Mapping of the epitopes of a panel of mouse anti-human and rat anti-mouse monoclonal antibodies gave concordant results. Epitopes were mapped to two different regions in the N-terminal domain, four regions within the I domain, and two regions between the I domain and the EF hand-like repeats. Monoclonal antibodies to epitopes within the mid- to C-terminal portion of the I domain and the N-terminal portion of the region between the I domain and the EF hand-like repeats gave good inhibition of LFA-1-dependent homotypic aggregation with cells that express either ICAM-1 or ICAM-3 as the major LFA-1 ligand.

Eight chemokines were tested for ability to elicit transendothelial chemotaxis of unstimulated peripheral blood T lymphocytes. The C-C chemokines monocyte chemotactic protein (MCP)-2, MCP-3, RANTES, macrophage inflammatory protein (MIP)-1 alpha, MIP-1 beta, and, as previously described, MCP-1 induced significant, dose-dependent transendothelial chemotaxis of CD3+ T lymphocytes. In contrast, the C-X-C chemokines interleukin-8 (IL-8) and interferon-gamma inducible protein-10 (IP-10) failed to induce transendothelial chemotaxis of CD3+ T lymphocytes or T lymphocyte subsets. RANTES, MIP-1 alpha, and MIP-1 beta induced significant transendothelial chemotaxis of CD4+, CD8+, and CD45R0+ T lymphocyte subsets. Phenotyping of mononuclear cells that underwent transendothelial migration to MCP-2, MCP-3, RANTES, or MIP-1 alpha showed both monocytes and activated (CD26 high), memory-type (CD45R0+) T cells. Both CD4+ and CD8+ T lymphocytes were recruited, but not natural killer cells or significant numbers of B cells. MCP-2 was the only C-C chemokine tested that attracted a significant number of naive-type (CD45RA+) T lymphocytes. In the absence of endothelium, IL-8 but not IP-10 promoted modest but significant chemotoxis of CD3+ T lymphocytes. Our data support the hypothesis that C-C, not the C-X-C chemokines IL-8 or IP-10, promote transendothelial chemotaxis of T lymphocytes.