NO303500B1 - Method for Preparing a Therapeutically Active Water ° Selective Functional Derivative of ICAM-1 and Recombinant DNA Molecule capable of Expressing a Water ° Selective Functional Derivative of ICAM-1 - Google Patents

Description

The present invention relates to a method for producing a therapeutically active water-soluble, functional derivative of ICAM-1 as well as recombinant DNA molecule capable of expressing a water-soluble functional derivative of ICAM-1.

Leukocytes must be able to attach to cell substrates in order to adequately defend the host against foreign invaders such as bacteria or viruses. An excellent overview of the defense system is provided by H.W. Eisen, (I: Microbiolo<gy>, 3rd ed., Harper & Row, Philadelphia, PA (1980), pp. 290-295 and 381-418). They must be able to attach to endothelial cells so that they can migrate from the circulation to sites of inflammation. Furthermore, they must be attached to antigen-presenting cells, so that a normal specific immune response can occur, and finally they must be attached to suitable target cells so that virus-infected cells or tumor cells can be lysed.

Leukocyte surface molecules involved in effecting such adhesions have recently been identified using hybridoma technology. Briefly, monoclonal antibodies directed against human T cells were identified (D. Davignon et al., Proe. Nati. Acad. Sei. USA, 78:4535-4539 (1981), and mouse spleen cells

(T. Springer et al., Eur. J. Immunol., 1:301-306 (1979)) which bound to leukocyte surfaces and inhibited the adhesion-related functions described above (T. Springer et al., Fed. Proe. , 44.: 2660-2663' (1985)). The molecules identified by these antibodies were named Mac-1 and lymphocyte function-associated antigen 1 (LFA-1). Mac-1 is a heterodimer found on macrophages, granulocytes and large granular lymphocytes. LFA-1 is a heterodimer found on most lymphocytes (T.A. Springer et al., Immunol. Rev. , 6_8:111-135 (1982)). These two molecules, plus a third molecule, pl50.95 (which has a tissue distribution similar to Mac-1) play a role in cellular adhesion

(G. Keizer et al.. Eur. J. Immunol.. 15:1142-1147 (1985)).

The leukocyte molecules described above were found to be members of a related family of glycoproteins (F. Sanchez-Madrid et al., J. Exper. Med.. 158:1785-1803 (1983); G.D. Keizer et al., Eur. J .Immunol., 15:1142-1147 (1985)). This glycoprotein family consists of heterodimers with one alpha chain and one beta chain. Although the alpha chain in each of the antigens differed from one another, the beta chain was found to be the most highly conserved (F. Sanchez-Madrid et al., J. Exper. Med., 158:1785-1803 (1983)). The beta chain of the glycoprotein family (sometimes referred to as "CD18") was found to have a molecular weight of 95 kd, while the alpha chains were found to range from 150 kd to 180 kd (T. Springer, Fed. Proe., 44=2660-2663 ( 1985)). Although the alpha subunits of the membrane proteins do not have the extensive homology of the beta subunits, careful analysis of the alpha subunits of the glycoproteins has shown that there are significant similarities between them. Overviews of the similarities between the alpha and beta subunits of the LFA-1 related glycoproteins are provided by F. Sanchez-Madrid et al., ( J. Exper. Med., 158:586-602 (1983); J. Exper. Med., 158:1785-1803 (1983)).

A group of individuals has been identified that cannot express normal amounts of any members of this adhesion protein family on their leukocyte cell surface (D.C. Anderson et al., Fed. Proe.. 44=2671-2677 (1985); D.C. Anderson et al. , J. Infect. Dis., 152:668-689 (1985)). Lymphocytes from these patients showed defects in vitro similar to corresponding normal individuals whose LFA-1 molecular family had been counteracted by antibodies. Furthermore, these individuals could not produce a normal immune response due to that their cells were unable to adhere to cell substrates (D.C. Anderson et al., Fed. Proe., 44=2671-2677 (185);' D.C. Anderson et al. , J. Inf ect. Dis . , 152 :668-689 (1985)). These data show that immune reactions are dampened when lymphocytes cannot attach normally due to lack of functional adhesion molecules in the LFA-1 family.

In summary, the ability of lymphocytes to maintain an animal's health and viability requires that they can attach to other cells (such as endothelial cells). This adhesion has been found to require cell-cell contacts that involve the presence of specific receptor molecules on the lymphocyte's cell surface. These receptors enable a lymphocyte to adhere to other lymphocytes or to endothelial cells and other non-vascular cells. The cell surface receptor molecules have been found to be highly related to each other. Humans whose lymphocytes lack these cell surface receptor molecules show chronic and recurrent infections, as well as other clinical symptoms including deficient antibody responses.

Since lymphocyte adhesion is part of the process by which foreign tissue is identified and rejected, an understanding of this process is of considerable value in the fields of organ transplantation, tissue transplantation, allergy and oncology.

The present invention relates to a method for producing intercellular adhesion molecule 1 (ICAM-1) functional derivatives that are therapeutically active and water-soluble. The ICAM-1 derivatives produced according to the invention are a fragment of ICAM-1 which lacks the cytoplasmic field and the transmembrane field of ICAM-1, and it contains field 1 or fields 1 and 2, or fields 1, 2 and 3, selected from the group containing at least one of the mutations referred to in figures 23, 24 and 25, possibly multimeric and chimeric forms of these with corresponding biological activity.

The method according to the invention is characterized in that it comprises the cultivation of a host microorganism that has been transformed with a recombinant vector that codes for a water-soluble, functional derivative of ICAM-1 as defined above in combination with transcriptional and translational regulatory sequences, as well as a sequence encoding a selection signal, under conditions whereby the derivative is expressed and secreted into the culture medium, and then isolation of the derivative.

The invention further includes a recombinant DNA molecule capable of expressing a water-soluble and functional ICAM-1 derivative which is a fragment of ICAM-1, which lacks the cytoplasmic domain and the transmembrane domain of CAM-1, and which contains domain 1 or domains 1 and 2, or fields 1, 2 and 3, selected from the group containing at least one of the mutations referred to in Figures 23, 24 and 25 and optionally chimeric forms thereof.

Preferred method forms according to the invention relate to the production of multimeric forms comprising up to 10 ICAM-1 derivatives and chimeric forms of ICAM-1 derivatives linked to immunoglobulin or a functional part thereof.

In one aspect of the invention, the multimeric form is produced which comprises a chimeric molecule where regions 1 and 2 of the ICAM-1 derivative are linked to a hinge region of the heavy chain of the immunoglobulin.

In another aspect of the invention, the multimeric form comprising a chimeric molecule is produced in which the C-terminal end of ICAM-1 region 2 is linked to a segment of the heavy chain of the immunoglobulin, immediately N-terminal to the hinge region.

Particularly preferably, according to the invention, chimeric forms of ICAM-1 derivatives are produced which are linked to immunoglobulins or parts of immunoglobulins where the immunoglobulin is selected from the group IgA, IgG or IgM.

The therapeutically active ICAM-1 derivatives produced according to the invention can be used in the treatment of inflammation which is the result of a response from the specific defense system in a mammal, which comprises giving an individual in need of such treatment an amount of an anti -inflammatory agent sufficient to suppress the inflammation; wherein the anti-inflammatory agent is selected from the group consisting of: an antibody capable of binding to ICAM-1; a fragment of an antibody, which fragment can bind to ICAM-1; ICAM-1; a functional derivative of ICAM-1; and a non-immunoglobulin antagonist of ICAM-1.

The therapeutically active ICAM-1 derivatives produced according to the invention can be used in the suppression of metastasis of a hematopoietic tumor cell that requires a functional element in the LFA-1 family for migration, which method comprises giving a patient in need of such treatment a an amount of an anti-inflammatory agent sufficient to suppress the metastasis; wherein the anti-inflammatory agent is selected from the group consisting of: an antibody capable of binding to ICAM-1;

a fragment of an antibody, which fragment can bind to ICAM-1; ICAM-1; a functional derivative of ICAM-1; and a non-immunoglobulin antagonist of ICAM-1.

The above-described uses can also occur for the suppression of inflammation or metastasis of a hematopoietic tumor cell, where the non-immunoglobulin antagonist of ICAM-1 is a different non-immunoglobulin antagonist of ICAM-1 than LFA-1.

The therapeutically active ICAM-1 derivatives produced according to the invention can be used for suppressing the growth of an ICAM-1-expressing tumor cell, which comprises giving a patient in need of such treatment an amount of a toxin sufficient to suppress the growth, which toxin is selected from the group consisting of a toxin-derivatized antibody capable of binding to ICAM-1; a toxin-derivatized fragment of an antibody, which fragment can bind to ICAM-1; a toxin-derivatized member of the LFA-1 molecular family;<p>g a toxin-derivatized functional derivative of a member of the LFA-1 molecular family.

They can also be used in suppressing the growth of an LFA-1-expressing tumor cell, which comprises administering to a patient in need of such treatment an amount of toxin sufficient to suppress such growth, which toxin is selected from the group consisting of of a toxin-derivatized ICAM-1; and a toxin-derivatized functional derivative of ICAM-1.

Reference is now made to the drawings:

Figure 1 shows in schematic form the adhesion between a normal cell and a cell lacking LFA-1. Figure 2 shows in schematic form the process of a normal/normal cell adhesion. Figure 3 shows the kinetics of cellular aggregation in the absence (X) or presence of 50 ng/ml PMA (O). Figure 4 shows co-aggregation between LFA-1" and LFA-l<+-> cells. Carboxyfluorescein diacetate-labeled EBV-transformed cells (10<4>) as shown in the figure were mixed with 10<5> unlabeled autologous cells (solid bars). or JY cells (open bars) in the presence of PMA. After 1.5 h, the labeled cells, in aggregates or free, were counted using the qualitative assay of Example 2. The percentage of labeled cells in aggregates is shown. One representative experiments of two are shown. Figure 5 shows immunoprecipitation of ICAM-1 and LFA-1 from JY cells. Triton X-100 lysates of JY cells (lanes 1 and 2) or control lysis buffer (lanes 3 and 4) were immunoprecipitated with antibody that could bind to ICAM-1 (lanes 1 and 3) or antibodies that could bind to LFA-1 (lanes 2 and 4).Part A shows results under reducing conditions; Part B shows results obtained under non-reducing conditions .Molecular weight standards were run in field S. Figure 6 shows the kinetics of IL-1 and gamma interferon effect r on ICAM-1 expression on human skin fibroblasts. Human skin fibroblasts were grown to a density of 8 x 10 cells/0, 3-2 cm well. IL-1 (10 U/ml, closed circles) or recombinant gamma interferon (10 U/ml, open squares) was added, and at the indicated time, the assay was cooled to 4°C, and an indirect binding assay was performed. The standard deviation did not exceed 10%. Figure 7 shows the concentration dependence of IL-1 and gamma interferon effects on ICAM-1. Human skin fibroblasts were grown to a density of 8 x 10<4> cells/0.32 cm<2>/well. IL-2 (open circle), recombinant human IL-1 (open square), recombinant mouse IL-1 (solid square), recombinant human gamma-interferon (solid circles) and recombinant beta-interferon (open triangle) were added in the indicated dilution and was incubated for 4 hours (IL-1) or 16 hours (beta and gamma interferon). The results given are the mean values from quadruplicate determinations; standard deviation did not exceed 10%. Figure 8 shows the nucleotide and amino acid sequence of ICAM-1 cDNA. The first ATG is at position 58. Translated sequences corresponding to tryptic ICAM-1 peptides are underlined. The hydrophobic putative signal peptide and transmembrane sequences are strongly underlined. N-linked glycosylation sites are boxed. A line is drawn over the polyadenylation signal AATAAA at position 2976. The sequence shown is for the HL-60 cDNA clone. Endothelial cell cDNA was sequenced over the majority of its length and showed only minor differences. Figure 9 shows the ICAM-1 homologous regions and the relationship with the immunoglobulin supergene family. (A) arrangement of 5 homologous regions (Dl-5). Two or more identical residues that matched are boxed. Residues conserved two or more times in NCAM regions, as well as residues conserved in regions in sets C2 and C1 were aligned with the internal ICAM-1 repeat units. The location of the predicted fé chains in the ICAM-1 region is marked with bars and lowercase letters above the alignments, and the known location for 6 chains in immunoglobulin C regions is marked with bars and uppercase letters below the alignment. The position of the putative disulfide bridge in ICAM-1 regions is indicated by S-S. (B-D) alignment of protein regions homologous to ICAM-1 regions; the proteins were initially aligned by examining NBRF databases using the FASTP program. The protein sequences are MAG, NCAM, T-cell receptor α-subunit V region, IgM/i chain and α-1-B glycoprotein. Figure 10 shows a schematic comparison of the secondary structures of ICAM-1 and MAG. Figure 11 shows LFA-1-positive EBV-transformed B-lymphoblastoid cells that bind to ICAM-1 in planar membranes. Figure 12 shows LFA-1-positive T-lymphoblasts and T-lymphoma cells that bind to ICAM-1 in plastic-bound vesicles. Figure 13 shows the inhibition of binding of JY-B lymphoblastoid cells that bind to ICAM-1 in plastic-bound vesicles by pre-treatment of cells or vesicles with monoclonal antibodies. Figure 14 shows the effect of temperature on binding of T-lymphoblasts to ICAM-1 in plastic-bound vesicles. Figure 15 shows requirements regarding divalent cations for binding of T-lymphoblasts to ICAM-1 in plastic bound vesicles. Figure 16 shows the effect of anti-adhesion antibodies on the ability of peripheral blood mononuclear cells to proliferate in response to recognition of the T cell-associated antigen OKT3. "0KT3" indicates addition of antigen. Figure 17 shows the effect of anti-adhesion antibodies on the ability of peripheral blood mononuclear cells to proliferate in response to recognition of the non-specific T-cell mitogen, concanavalin A. "CONA" denotes the addition of concanavalin A. Figure 18 shows the effect of anti-adhesion antibodies on the ability of peripheral blood mononuclear cells to proliferate in response to perception of the keyhole limpet hemocyanin antigen. The addition of keyhole limpet hemocyanin to the cells is indicated by "KLH". Figure 19 shows the effect of anti-adhesion antibodies on the ability of peripheral blood mononuclear cells to proliferate in response to perception of the tetanus toxoid antigen. The addition of tetanus toxoid antigen to the cells is indicated by "AGN". Figure 20 shows the binding of monoclonal antibodies RR1/1, R6.5, LB2 and CL203 to ICAM-1 knockout mutants. Figure 21 shows the binding of ICAM-1 knockout mutants to LFA-1. Figure 22 shows the epitopes recognized by monoclonal anti-ICAM-1 antibodies RR1/1, R6.5, LB2 and CL203. Figure 23 shows the binding capacity of ICAM-1 region 2 mutants to LFA-1. Figure 24 shows the binding capacity of ICAM-1 region 3 mutants to LFA-1. Figure 25 shows the binding capacity of ICAM-1 region 1 mutants to LFA-1. Figure 26 shows the alignment of amino-terminal ICAM regions.

One aspect of the background to the present invention concerns the discovery of a natural binding ligand to LFA-1. Molecules such as those in the LFA-1 family, which are involved in the process of cell adhesion, are referred to as "adhesion molecules".

The natural binding ligand is termed "intercellular adhesion molecule" or "ICAM-1". ICAM-1 is a 76-97 kd glycoprotein. ICAM-1 is not a heterodimer. The present invention is directed to the production of "functional derivatives" of ICAM-1. A "functional derivative" of ICAM-1 is a compound that has a biological activity (either functional or structural) that is substantially similar to a biological activity of ICAM-1. The term "functional derivatives" is intended to include the "fragments", "variants", "analogs" or "chemical derivatives" of a molecule. A "fragment" of a molecule such as ICAM-1 is intended to denote any polypeptide subset of the molecule. Fragments of ICAM-1 which have ICAM-1 activity and which are soluble (ie not membrane bound) are particularly preferred. A "variant" of a molecule such as ICAM-1 is intended to denote a molecule with substantially similar structure and function to either the entire molecule or a fragment thereof. A molecule is said to be "substantially similar" to another molecule if both molecules have substantially similar structures, or if both molecules have similar biological activity. On the condition that two molecules have equal activity, they are thus considered variants as this designation is used herein, even if the structure of one of the molecules is not found in the other, or if the sequence of amino acid residues is not identical. An "analog" of a molecule such as ICAM-1 is intended to mean a molecule that has substantially the same function as either the entire molecule or a fragment thereof.

As used herein, a molecule is said to be a "chemical derivative" of another molecule when it contains additional chemical moieties that are not normally part of the molecule. Such parts can improve the molecule's solubility, absorption, biological half-life, etc. The parts can alternatively reduce the molecule's toxicity, eliminate or mitigate any undesirable side effects of the molecule, etc. Parts that can produce such effects are described in Reminaton's Pharmaceutical Sciences (1980) . "Toxin-derivatized" molecules constitute a special class of "chemical derivatives". A "toxin-derivatized" molecule is a molecule (such as ICAM-1 or an antibody) that contains a toxin moiety. The binding of such a molecule to a cell brings the toxin part into close proximity to the cell and thereby accelerates cell death. Any suitable toxin moiety may be used; however, it is preferred to use the toxins such as e.g. ricinus toxin, diphtheria toxin, radioisotopic toxins, membrane-channel-forming toxins, etc. Methods for connecting such parts to a molecule are well known in the field.

An antigenic molecule such as ICAM-1, or members of the LFA-1 family of molecules, is naturally expressed on the surface of lymphocytes. Thus, introduction of such cells into an appropriate animal, as by intraperitoneal injection, etc., will result in the formation of antibodies that can bind to ICAM-1 or members of the LFA-1 family of molecules. If desired, the serum from such an animal can be removed and used as a source of polyclonal antibodies that can bind these molecules. However, it is preferred to remove splenocytes from such animals, to fuse such spleen cells with a myeloma cell line, and to allow such fusion cells to form a hybridoma cell that secretes monoclonal antibodies capable of binding ICAM-1 or members of the LFA-1 family of molecules.

The hybridoma cells obtained in the manner described above can be screened by many different methods to identify desired hybridoma cells that secrete antibody that can bind either to ICAM-1 or to members of the LFA-1 family of molecules. In a preferred screening assay, such molecules are identified by their ability to inhibit the aggregation of Epstein-Barr virus-transformed cells. Antibodies which can inhibit such aggregation are then further examined to determine whether they inhibit such aggregation by binding to ICAM-1, or to a member of the LFA-1 family of molecules. Any aid that can distinguish ICAM-1 from the LFA molecular family can be used in such an investigation. Thus, e.g. the antigen bound by the antibody is analyzed such as by immunoblotting and polyacrylamide gel electrophoresis. If the bound antigen is a member of the LFA-1 family of molecules, the immunoprecipitated antigen will be found to be a dimer, whereas if the bound antigen is ICAM-1, a single molecular weight compound will have been immunoprecipitated. Alternatively, it is possible to distinguish such antibodies that bind to members of the LFA-1 family of molecules from those that bind ICAM-1 by examining the ability of the antibody to bind to cells such as granulocytes, which express LFA-1 but not ICAM-1. The ability of an antibody (which is known to inhibit cell aggregation) to bind to granulocytes indicates that the antibody can bind LFA-1. Absence of such binding indicates an antibody capable of recognizing ICAM-1. The ability of an antibody to bind to a cell such as a granulocyte can be demonstrated by aids that are usually used by those of ordinary skill in the art. Such aids include immunoassays, cell agglutination, filter binding studies, antibody purification, etc.

Alternatively, the anti-aggregation antibodies can be identified by measuring their ability to differentially bind to cells expressing ICAM-1 (such as activated endothelial cells) and their inability to bind to cells that do not express ICAM-1. As will be readily understood by those of ordinary skill in the art, the above assays can be modified or performed in a different sequential order to provide different potential screening assays, each of which can identify and discriminate between antibodies that can bind to ICAM-1 and members of the LFA family of molecules.

The anti-inflammatory agents can be obtained by natural processes (such as by causing an animal, a plant, a fungus, a bacterium_etc. to form a non-immunoglobulin antagonist to ICAM-1, or by causing an animal to to generate polyclonal antibodies that can bind to ICAM-1); by synthetic methods (such as, for example, using the Merrifield method for synthesizing polypeptides to synthesize ICAM-1, functional derivatives of ICAM-1 or protein antagonists of ICAM-1 (either immunoglobulin or non-immunoglobulin)); by hybridoma technology (such as for the generation of monoclonal antibodies that can bind to ICAM-1); or by recombinant technology (such as, for example, for the production of the anti-inflammatory agents produced according to the present invention in different hosts (ie yeast, bacteria, fungi, cultured mammalian cells, etc.), or from recombinant plasmids or virus vectors) . The choice of which method to use will depend on factors such as simplicity, desired yield, etc. It is not necessary to use only one of the above-described methods, processes or technologies for the production of a particular anti-inflammatory agent; the methods, methods and technologies described above can be combined to obtain a particular anti-inflammatory agent.

A. Identification of the LFA-1 binding partner (ICAM-1)

1. Analyzes of LFA-1-dependent aggregation

Many Epstein-Barr virus-transformed cells show aggregation. This aggregation can be increased by the presence of phorbol esters. Such homotypic aggregation (ie, aggregation involving only one cell type) was found to be blocked by anti-LFA-1 antibodies (R. Rothlein et al., J. Ex<p>er. Med., 163:1132-1149

(1986)), which reference is incorporated herein by reference. Thus, the extent of LFA-1-dependent binding can be determined by assessing the extent of spontaneous or phorbol ester-dependent aggregate formation.

An agent that interacts with LFA-1-dependent aggregation can be identified using an assay that can determine whether the agent interacts with either the spontaneous aggregation or the phorbol ester-dependent aggregation of Epstein-Barr virus-transformed cells. Most Epstein-Barr virus-transformed cells can be used in such an analysis as long as the cells can express the LFA-1 receptor molecule. Such cells can be produced according to the technique of T.A. Springer et al.. J. Exper Med., 160:1901-1918 (1984), which reference is incorporated herein by reference. Although any such cell can be used in the LFA-1-dependent binding assay, it is preferred to use cells from the JY cell line (C.T. Terhost et al., Proe. Nati. Acad. Sei. USA. 73.: 910 (1976) ). The cells can be cultured in any suitable culture medium; however, it is most preferred to grow the cells in RMPI 1640 culture medium supplemented with 10% fetal calf serum and 50 µg/ml gentamycin (Gibco Laboratories, NY). The cells should be cultured under conditions suitable for mammalian cell propagation (ie, at a temperature of typically 37°C, in an atmosphere of 5% CO 2 , at a relative humidity of 95%, etc.).

2. LFA-1 binds to ICAM-1

Individuals have been identified among humans whose lymphocytes lack the LFA-1 receptor family of molecules (D.C. Anderson et al., Fed. Proe. 44=2671-2677 (1985); D.C. Anderson et al.. J. Infect. Dis., 152: 668-689 (1985)). Such individuals are said to suffer from leukocyte adhesion deficiency (LAD). EBV-transformed cells from such individuals do not aggregate either spontaneously or in the presence of phorbol esters in the aggregation assay described above. When such cells were mixed with LFA-1 expressing cells, aggregation was observed (R. Rothlein et al., J. Exper. Med., 161:1132-1149 (1986)) (Figure 1). Significantly, these aggregates did not form if these cells were incubated in the presence of anti-LFA-1 antibodies. Thus, although the aggregation required LFA-1, the ability of cells lacking LFA-1 to form aggregates with LFA-1-containing cells demonstrated that the LFA-1 binding partner was not LFA-1, but rather a previously undiscovered cell adhesion molecule. Figure 1 shows the mechanism of cell adhesion.

B. Intercellular adhesion molecule 1 (ICAM-1)

The novel intercellular adhesion molecule ICAM-1 was first identified and partially characterized according to the method of R. Rothlein et al. (J. Immunol., 137:1270-1274 (1986)), which reference is incorporated herein by reference. For detection of the ICAM-1 molecule, monoclonal antibodies were prepared from spleen cells of mice immunized with cells from individuals genetically lacking LFA-1 expression. Resulting antibodies were examined for their ability to inhibit the aggregation of LFA-1 expressing cells (Figure 2). In more detail, mice were immunized with EBV-transformed B cells from LAD patients that do not express the LFA-1 antigen. The spleen cells from these animals were then removed, fused with myeloma cells, and allowed to become hybridoma cells that produced monoclonal antibodies. EBV-transformed B cells from normal individuals expressing LFA-1 were then incubated in the presence of the monoclonal antibody from the hybridoma cell to identify any monoclonal antibody that could inhibit the phorbol ester-induced, LFA-1-dependent, spontaneous aggregation of the EBV-transformed B cells. Since the hybridoma cells were derived from cells that had never encountered the LFA-1 antigen, no monoclonal antibody to LFA-1 was formed. Consequently, any antibody found to inhibit aggregation must be able to bind to an antigen that, although not LFA-1, participates in the LFA-1 adhesion process. Although any method of obtaining such monoclonal antibodies may be used, it is preferred to obtain ICAM-1 binding monoclonal antibodies by immunization of Balb/C mice using the methods and programs described by R. Rothlein et al. , ( J. Immunol., 137:1270-1274 (1986)) with Epstein-Barr virus-transformed peripheral blood mononuclear cells from individuals deficient in LFA-1. Such cells are described by T.A. Springer et al., (J. Exper Med., 160:1901-1918 (1984)).

In a preferred method for the formation and detection of antibodies that can bind to ICAM-1, mice are immunized with either EBV-transformed B cells that express both ICAM-1 and LFA-1, or more preferably with TNF-activated endothelial cells that express ICAM -1, but not LFA-1. In a most preferred method for generating hybridoma cells that produce anti-ICAM-1 antibodies, a Balb/C mouse was sequentially immunized with JY cells and with differentiated U937 cells (ATCC CRL-1593). The spleen cells from such animals are removed, fused with myeloma cells and allowed to develop into antibody-forming hybridoma cells. The antibodies are tested for their ability to inhibit the LFA-1-dependent, phorbol ester-induced aggregation of an EBV-transformed cell line, such as JY cells, which express both the LFA-1 receptor and ICAM-1. As shown by R. Rothlein et al. , ( J. Immunol., 137 :1270-1274 (1987)), antibodies which can inhibit such aggregation are then examined, with regard to their ability to inhibit the phorbol ester-induced aggregation of a cell line, such as SKW3 (M. Dustin et al., J. Exper Med., 165:672-692 (1987)), whose ability to spontaneously aggregate in the presence of a phorbol ester is inhibited by antibody capable of binding LFA-1, but is not inhibited by anti-ICAM-1 antibodies . Antibodies which can inhibit the phorbol ester-induced aggregation of cells such as JY cells, but which cannot inhibit the phorbol ester-induced aggregation of cells such as SKW3 cells, are likely to be anti-ICAM-1 antibodies. Alternatively, antibodies that can bind to ICAM-1 can be identified by screening for antibodies that can inhibit the LFA-1-dependent aggregation of LFA-expressing cells (such as JY cells) but that cannot bind to LFA-expressing cells 1, but little or no ICAM-1 (such as normal granulocytes), or that can bind to cells expressing ICAM-1 but not LFA-1 (such as TNF-activated endothelial cells). Another option is to immunoprecipitate from cells expressing ICAM-1, LFA-1, or both, using antibodies that inhibit the LFA-1-dependent aggregation of cells, such as JY cells, and by SDS-PAGE or an equivalent method determine one or another molecular characteristic of the molecule that is precipitated with the antibody. If the characteristic is the same as that of ICAM-1, the antibody can be assumed to be an anti-ICAM-1 antibody.

Using monoclonal antibodies prepared as described above, the ICAM-1 cell surface molecule was purified and characterized. ICAM-1 was purified from human cells or tissues using monoclonal antibody affinity chromatography. In such a method, a monoclonal antibody that can react with ICAM-1 is coupled to an inert column matrix. Any method of performing such coupling may be used; however, it is preferred to use the method according to H.C. Oett<g>en et al.. J. Biol. Chem.. 259:12034 (1984)). When a cell lysate is passed through the matrix, the ICAM-1 molecules present are adsorbed and retained by the matrix. By changing the pH or the ion concentration in the column, the bound ICAM-1 molecules can be eluted from the column. Although any suitable matrix can be used, it is preferred to use sepharose (Pharmacia) as matrix material. The formation of column matrices and their use in protein purification is well known in the art.

In a manner understood by those of ordinary skill in the art, the above-described assays can be used for the identification of compounds that can weaken or inhibit the degree or extent of cell adhesion.

ICAM-1 is a cell surface glycoprotein expressed on non-hematopoietic cells such as vascular endothelial cells, thymic epithelial cells, certain other epithelial cells and fibroblasts, and on hematopoietic cells such as tissue macrophages, mitogen-stimulated T-lymphocyte blasts and nucleated B cells and dendritic cells in tonsils, lymph nodes and Peyer's patches. ICAM-1 is highly expressed on vascular endothelial cells in T-cell areas of lymph nodes and tonsils showing reactive hyperplasia. ICAM-1 is expressed in small amounts on lymphocytes in peripheral blood. Phorbol ester-stimulated differentiation of some myelomonocytic cell lines greatly increases ICAM-1 expression. Thus, ICAM-1 is preferentially expressed at sites of inflammation and is not usually expressed by resting cells. ICAM-1 expression on skin fibroblasts is increased three- to fourfold by either interleukin 1 or gamma interferon at levels of 10 U/ml over a period of 4 or 10 hours, respectively. The induction is dependent on protein and mRNA synthesis and is reversible.

ICAM-1 shows molecular weight heterogeneity in different cell types with a molecular weight of 97 kd on fibroblasts, 114 kd on the myelomonocytic cell line U937, and 90 kd on the lymphoblastoid B cell JY. ICAM-1 biosynthesis has been found to involve an intracellular precursor of approx. 73 kd. The non-N-glycosylated form resulting from tunicamycin treatment (which inhibits glycosylation) has a molecular weight of 55 kd.

ICAM-1 isolated from phorbol ester-stimulated U937 cells or from fibroblast cells yields an identical major product with a molecular weight of 60 kd after chemical deglycosylation. Monoclonal ICAM-1 antibodies interact with the adhesion of phytohemagglutinin blasts to cell lines lacking LFA-1. Pretreatment of fibroblasts, but not lymphocytes, with monoclonal antibodies capable of binding ICAM-1 inhibits lymphocyte-fibroblast adhesion. Pretreatment of lymphocytes, but not fibroblasts, with antibodies to LFA-1 has also been found to inhibit lymphocyte-fibroblast adhesion.

ICAM-1 is thus the binding ligand of the CD-18 complex on leukocytes. It can be applied to fibroblasts and endothelial cells in vitro by inflammatory mediators such as IL-1, gamma interferon and tumor necrosis factor within a time frame consistent with the infiltration of lymphocytes into inflammatory lesions in vivo (M'.L. Dustin et al. , J , Immunol, 137:245-254

(1986); J. S. Prober et al., J. Immunol. 137:1893-1896. Furthermore, ICAM-1 is expressed on non-hematopoietic cells such as vascular endothelial cells, thymic epithelial cells, other epithelial cells and fibroblasts, and on hematopoietic cells such as tissue macrophages, mitogen-stimulated T-lymphocyte blasts and nuclear center B cells and dendritic cells in tonsils, lymph nodes and Peyer's stains (M.L. Dustin et_al., J. Immunol. 137:245-254 (1986)). ICAM-1 is expressed on keratinocytes in benign inflammatory lesions such as allergic eczema, lichen planus, exanthema, urticaria and blistering diseases. Allergic skin reactions produced by application of a hapten to the skin to which the patient is allergic also showed a strong ICAM-1 expression on the keratinocytes. On the other hand, toxic spots on the skin did not show ICAM-1 expression on the keratinocytes. ICAM-1 is present on keratinocytes from biopsies of skin lesions from various dermatological disorders, and ICAM-1 expression is induced on allergic patch test lesions, whereas keratinocytes from toxic patch test lesions did not express ICAM-1.

ICAM-1 is therefore a cell substrate to which lymphocytes can attach, so that the lymphocytes can migrate to sites of inflammation and/or perform various effector functions that contribute to this inflammation. Such functions include the formation of antibody, lysis of virally infected target cells, etc. The term "inflammation" as used herein is intended to include reactions in the specific or non-specific defense system. As used herein, the term "specific defense system" is intended to denote the component of the immune system that responds to the presence of specific antigens. Inflammation is said to be the result of a response from the specific defense system if the inflammation is caused by, mediated by or associated with a response from the specific defense system. Examples of inflammation resulting from a response of the specific defense system include the response to antigens such as rubella virus, autoimmune diseases, delayed-type hypersensitivity response, mediated by T cells (such as would be seen in individuals giving "positive" test in the Mantaux test), psoriasis, etc.

A "non-specific defense system reaction" is a response mediated by leukocytes that do not have immunological memory. Such cells include granulocytes and macrophages. As used herein, an inflammation is said to be the result of a response of the non-specific defense system if the inflammation is caused by, mediated by, or associated with a response of the non-specific defense system. Examples of inflammation that results, at least in part, from a response of the non-specific defense system include inflammation associated with conditions such as: asthma; adult respiratory distress syndrome (ARDS) or multiple organ injury syndromes secondary to sepsis or trauma; reperfusion injury to myocardial or other tissues; acute glomerulonephritis; reactive arthritis; dermatoses with acute inflammatory components; acute purulent meningitis or other inflammatory disorders of the central nervous system; thermal damage; hemodialysis; leukapheresis; ulcerative colitis; Chron's disease; necrotizing enterocolitis; granulocyte-transfusion-associated syndromes; and cytokine-mediated toxicity.

According to the present invention, functional ICAM-1 derivatives, and in particular such derivatives comprising fragments or mutant variants of ICAM-1 having regions 1, 2 and 3, can be used in the treatment or therapy of such reactions in the non- specific defense system. More preferred for such treatment or therapy are ICAM-1 fragments or mutant variants containing region 2 of ICAM-1. Most preferred for such treatment or therapy are ICAM-1 fragments or mutant variants containing region 1 of ICAM-1.

C. Cloning of the ICAM-1 gene

Any of many different methods can be used to clone the ICAM-1 gene. One such method involves analyzing a shuttle vector library of cDNA inserts (derived from an ICAM-1 expressing cell) for the presence of an insert containing the ICAM-1 gene. Such an analysis can be performed by cells being transfected with the vector and then analyzed for ICAM-1 expression. The preferred method for cloning this gene involves determining the amino acid sequence of the ICAM-1 molecule. To accomplish this task, ICAM-1 protein can be purified and analyzed by automated sequence analyzers. Alternatively, the molecule can be fragmented such as with cyanogen bromide or with proteases such as papain, chymotrypsin or trypsin (Y. Oike et al., J. Biol. Chem., 257:9751-9758 (1982); C. Liu et al., Int. J .Pept.Protein Res., 21:209-215 (1983)). Although it is possible to determine the entire amino acid sequence of ICAM-1, it is preferred to determine the sequence of peptide fragments in the molecule. If the peptides have a length of more than 10 amino acids, the sequence information is usually sufficient to clone a gene such as the gene for ICAM-1.

The sequence of amino acid residues in a peptide is designated herein either by the use of their commonly used 3-letter designations or by their single-letter designations. A listing of these 3-letter and 1-letter designations can be found in textbooks such as Biochemistry, A. Lehninger, Worth Publishers, New York, NY (1970). When such a sequence is listed vertically, the amino-terminal residue is meant to be at the top of the list, and the carboxy-terminal residue in the peptide is meant to be at the bottom of the list. In horizontal entry, similarly, the amino-terminal residue is intended to be at the left end, while the carboxy-terminal residue is intended to be at the right end. The residues of amino acids in a peptide can be separated by hyphens. Such hyphens are only intended to make the presentation of a sequence easier. As a purely illustrative example, the amino acid sequence denoted:

-Gly-Ala-Ser-Phe-

that an Ala residue is bound to the carboxy group on Gly, and that a Ser residue is bound to the carboxy group on the Ala residue and to the amino group on a Phe residue. The designation further indicates that the amino acid sequence contains the tetrapeptide Gly-Ala-Ser-Phe. The term is not intended to limit the amino acid sequence of this one tetra-peptide, but is intended to include (1) the tetrapeptide with one or more amino acid residues attached to either its amino or carboxy terminus, (2) the tetrapeptide with one or more amino acid residues attached to both its amino and its carboxy end, (3) the tetrapeptide without any further amino acid residues.

When one or more suitable peptide fragments have been sequenced, the DNA sequences that can encode them are examined. Because the genetic code is degenerate, more than one codon can be used to code for a particular amino acid (J.D. Watson, I: Molecular Biology of the Gene, 3rd ed., W.A. Benjamin, Inc., Menlo Park, Ca (1977 ), pp. 356-357). The peptide fragments are analyzed for the identification of sequences of amino acids that can be encoded by oligonucleotides with the lowest degree of degeneracy. This is preferably carried out by identifying sequences containing amino acids coded for by only a single codon. Although such amino acid sequences may occasionally be encoded by only a single oligonucleotide, often the amino acid sequence may be encoded by any of a set of similar oligonucleotides. It is significant that although all the elements in the set contain oligonucleotides that can code for the peptide fragment, and thus potentially contain the same nucleotide sequence as the gene that codes for the peptide fragment, only one element in the set contains a nucleotide sequence that is identical to the nucleotide sequence in this gene. Because this element is contained within the kit and can hybridize to DNA even in the presence of the other elements of the kit, it is possible to use the non-fractionated set of oligonucleotides in the same way as one would use a single oligonucleotide to clone the gene that codes for the peptide.

In a manner exactly analogous to that described above, one can use an oligonucleotide (or set of oligonucleotides) having a nucleotide sequence that is complementary to the oligonucleotide sequence or set of sequences that can code for the peptide fragment.

A suitable oligonucleotide, or set of oligonucleotides, which may encode a fragment of the ICAM-1 gene (or which is complementary to such an oligonucleotide, or set of oligonucleotides) is identified (using the method described above), synthesized and hybridized by aids which are well known in the field, against a DNA, or more preferably, a cDNA preparation originating from human cells which can express ICAM-1 gene sequences. Techniques for nucleic acid hybridization are described by R. Maniatis et al., In: Molecular Clonin<g>, a Laboratory Manual, Coldspring Harbor, NY (1982), and by B.D. Haymes et al., I. Nucleic Acid Hybridization, a Practical Approach, IRL Press, Washington, DC (1985), which references are incorporated herein by reference. The source of the DNA or cDNA used will preferably have been enriched with respect to ICAM-1 sequences. Such enrichment is most easily obtained by cDNA obtained by extracting RNA from cells grown under conditions that induce ICAM-1 synthesis (such as U937 grown in the presence of phorbol esters, etc.).

Techniques such as, or similar to, those described above have successfully enabled the cloning of genes for human aldehyde dehydrogenases (L.C. Hsu et al.. proe. Nat. Acad. Sei. USA, 82:3771-3775 (1985) , fibronectin (S. Suzuki et al.. Eur. Mol. Biol. Organ. J., 4:2519-2524 (1985)), the human estrogen receptor gene (P. Walter et al., Proe. Nati. Acad. Sci. USA, 82.: 7889-7893 (1985)), tissue type plasminogen activator (D. Pennica et al., Nature, 301:214-221 (1983)) and human placental alkaline phosphatase complementary DNA (W. Kam et al. ., Proe. Nat. Acad. Sci. USA, 82:8715-8719 (1985)).

In a preferred alternative way to cloning the ICAM-1 gene, a library of expression vectors is prepared by cloning DNA, or more preferably cDNA, from a cell which can express ICAM-1, into an expression vector. The library is then examined for elements that can express a protein that binds to anti-ICAM-1 antibody, and that has a nucleotide sequence that can code for polypeptides that have the same amino acid sequence as ICAM-1 or fragments of ICAM-1.

The cloned ICAM-1 gene, obtained by the methods described above, can be operably linked to an expression vector and introduced into bacterial or eukaryotic cells to produce ICAM-1 protein. Techniques for such manipulations are described by T. Maniatis et al., as above, and are well known in the art.

D. Applications of assays of LFA-1-dependent aggregation

The above-described assay, which can measure LFA-1-dependent aggregation, can be used for the identification of agents that act as antagonists to inhibit the extent of LFA-1-dependent aggregation. Such antagonists can work by weakening the ability of LFA-1 or ICAM-1 to mediate aggregation. Thus, such agents include immunoglobulins such as an antibody that can bind to either LFA-1 or ICAM-1. In addition, non-immunoglobulin agents (ie, chemical) can be tested using the assay described above to determine whether they are antagonists of LFA-1 aggregation.

E. Applications of antibodies that can bind to ICAM-1 receptor proteins

1. Anti-inflammatory agents

Monoclonal antibodies to elements of the CD-18 complex inhibit many adhesion-dependent functions of leukocytes including binding to endothelium (D. Haskard et al., J. Immunol., 137:2901-2906 (1986)), homotypic adhesions (R. Rothlein et al., J. Exp. Med., 163:1132-1149 (1986)), antigen- and mitogen-induced proliferation of lymphocytes (D. Davignon et al., Proe. Nati. Acad. Sei. USA, 78:4535 -4539 (1981)), antibody formation (A. Fischer et al., J. Immunol., 136:3198-3203 (1986)), and effector functions of all leukocytes such as lytic activity of cytotoxic T cells (A.M. Krenskv et al. , J. Immuno., 132:2180-2182 (1984)), macrophages (G. Strassman et al. , J. Immunol. , 136.:4328-4333

(1986)), and all cells involved in antibody-dependent cell cytotoxicity reactions (S. Kohl et al., J. Immunol., 133:2972-2978 (1984)). In all of the above functions, the antibodies inhibit the leukocyte's ability to attach to the appropriate cell substrate, which in turn inhibits the final outcome.

As discussed above, the binding of ICAM-1 molecules to the members of the LFA-1 family of molecules is of central importance in cellular adhesion. During the adhesion process, lymphocytes can constantly monitor an animal for the presence of foreign antigens. Although these processes are normally desirable, they are also the cause of organ transplant rejection, tissue transplant rejection and many autoimmune diseases. Consequently, any aid that could weaken or inhibit cellular adhesion would be highly desirable for recipients of organ transplants, tissue transplants, or for autoimmune patients.

Monoclonal antibodies that can bind to ICAM-1 are very well suited as anti-inflammatory agents in a mammal. It is significant that such agents differ from conventional anti-inflammatory agents in that they can selectively inhibit adhesion and do not produce other side effects such as kidney toxicity found in the case of conventional agents. Monoclonal antibodies that can bind to ICAM-1 can therefore be used to prevent organ or tissue rejection, or to modify autoimmune responses without fear of such side effects in the mammal.

It is significant that the use of monoclonal antibodies that can perceive ICAM-1 can make it possible to perform organ transplants even between individuals who have HLA mismatch.

2. Suppressors of delayed type hypersensitivity reaction

Since ICAM-1 molecules are mostly expressed at sites of inflammation, such as sites involved in delayed-type hypersensitivity reactions, antibodies (especially monoclonal antibodies) that can bind to ICAM-1 molecules have therapeutic potential in attenuation or elimination of such reactions. This potential therapeutic application can be exploited in one of two ways. First, a composition containing a monoclonal antibody against ICAM-1 can be administered to a patient experiencing a delayed-type hypersensitivity reaction. E.g. such compositions can be administered to an individual who has been in contact with antigens such as poison ivy, poison oak, etc. In the second embodiment, the monoclonal antibody capable of binding to ICAM-1 is administered to a patient in association with an antigen for the prevention of a subsequent inflammatory reaction. Thus, the additional administration of an antigen with an ICAM-1-binding monoclonal antibody may enable an individual to temporarily tolerate subsequent presentation of this antigen.

3. Therapy for chronic inflammatory disease

Since LAD patients lacking LFA-1 do not produce an inflammatory response, it is hypothesized that antagonism of LFA-1's natural ligand, ICAM-1, would also inhibit an inflammatory response. The ability of antibodies against ICAM-1 to inhibit inflammation provides the basis for their therapeutic use in the treatment of chronic inflammatory diseases and autoimmune diseases such as lupus erythematosus, autoimmune thyroiditis, experimental allergic encephalomyelitis (EAE), multiple sclerosis, some forms of diabetes Reynaud's syndrome, rheumatoid arthritis etc. Such antibodies can also be used as therapy in the treatment of psoriasis. Generally, the monoclonal antibodies that can bind to ICAM-1 can be used in the treatment of such diseases that can usually be treated by steroid therapy.

4. Diagnostic and prognostic applications

Since ICAM-1 is mostly expressed at sites of inflammation, monoclonal antibodies that can bind to ICAM-1 can be used as a means of imaging or visualizing the sites of infection and inflammation in a patient. In such use, the monoclonal antibodies are detectably labeled using radioisotopes, affinity markers (such as biotin, avidin, etc.), fluorescent markers, paramagnetic atoms, etc. Methods for carrying out such labeling are well known in the field. An overview of the clinical use of antibodies in diagnostic imaging of H.B. Grosman, Uroi. Clin. North Amerk., 13:465-474 (1986), E.C. Unger et al., Invest-Radiol.. 20: 693:700

(1985) and B.A. Khaw et al., Science, 209:295-297 (1980).

The presence of inflammation can also be detected using binding ligands such as mRNA, cDNA or DNA, which bind to ICAM-1 gene sequences or to ICAM-1 mRNA sequences,

in cells expressing ICAM-1. Techniques for performing such hybridization assays are described by T. Maniatis (as above).

Detection of centers for such detectably labeled antibodies indicates a site of inflammation or tumor development. In one embodiment, this investigation of inflammation is performed by removing tissue samples or blood and incubating such samples in the presence of detectably labeled antibodies. In a preferred embodiment, this technique is performed in a non-invasive manner using magnetic imaging, fluorography, etc. Such a diagnostic test can be used to screen organ transplant recipients for early signs of potential tissue rejection. Such analyzes may also be performed in efforts to determine an individual's susceptibility to rheumatoid arthritis or other chronic inflammatory diseases.

5. Supplement to the introduction of antigenic material administered for therapeutic or diagnostic purposes

Immune responses to therapeutic or diagnostic agents such as e.g. bovine insulin, interferon, tissue-type plasminogen activator, or mouse monoclonal antibodies substantially impair the therapeutic or diagnostic value of such agents and may actually cause diseases such as serum sickness.

Such a situation can be remedied by using the antibodies. In this embodiment, such antibodies would be administered in combination with the therapeutic or diagnostic agent. The addition of the antibodies prevents the recipient from perceiving the agent, and therefore prevents the recipient from initiating an immune response against it. Absence of such an immune response results in the patient receiving further administrations of the therapeutic or diagnostic agent.

F. Applications of intercellular adhesion molecule 1 (ICAM-1)

ICAM-1 is a binding partner of LFA-1. As such, ICAM-1 or its functional derivatives can be used interchangeably with antibodies that can bind to LFA-1 in the treatment of disease.

In solubilized form, such molecules can thus be used to inhibit inflammation, organ failure, graft failure, etc. ICAM-1 or its functional derivatives can be used in the same way as anti-ICAM antibodies to reduce the immunogenicity of therapeutic or diagnostic agents.

ICAM-1, its functional derivatives and its antagonists can be used to block metastasis or proliferation of tumor cells expressing either ICAM-1 or LFA-1 on their surfaces. Many different methods can be used to achieve such a goal. E.g. migration of hematopoietic cells requires LFA-1-ICAM-1 binding. Antagonists of such binding therefore suppress this migration and block the metastasis of tumor cells originating from leukocytes. Alternatively, toxin-derivatized molecules capable of binding either ICAM-1 or a member of the LFA-1 family of molecules can be administered to a patient. When such toxin-derivatized molecules bind to tumor cells expressing ICAM-1 or a member of the LFA-1 family of molecules, the tumor cell is killed in the presence of the toxin, thereby inhibiting the proliferation of the tumor.

G. Applications of non-immunoglobulin antagonists to ICAM-1-dependent adhesion

ICAM-1-dependent adhesion can be inhibited by non-immunoglobulin antagonists that can bind either to ICAM-1 or to LFA-1. An example of a non-immunoglobulin antagonist of ICAM-1 is LFA-1. An example of a non-immunoglobulin antagonist that binds to LFA-1 is ICAM-1. By using the assays described above, additional non-immunoglobulin antagonists can be identified and purified. Non-immunoglobulin antagonists to ICAM-1-dependent adhesion can be used for the same purpose as antibodies against LFA-1 or antibodies against ICAM-1.

H. Administration of the compositions prepared according to the present invention

The therapeutic effects of ICAM-1 can be achieved by giving a functional part of the ICAM-1 molecule that is therapeutically active to a patient.

ICAM-1 and its functional derivatives can be obtained either synthetically, using recombinant DNA technology or by proteolysis. The therapeutic benefits of ICAM-1 can be enhanced by the use of functional derivatives of ICAM-1 that have additional amino acid residues added to increase binding to carrier or to increase the activity of ICAM-1. The scope of the present invention is further intended to include functional derivatives of ICAM-1 which lack certain amino acid residues or which contain altered amino acid residues, as long as such derivatives show the capacity to influence cell adhesion.

Both the antibodies and the ICAM-1 molecule described herein are said to be "substantially free of natural contaminants" if preparations containing them are substantially free of materials with which these products are normally and naturally found together.

Antibodies and biologically active fragments thereof (either polyclonal or monoclonal) which can bind to ICAM-1 can be produced either by an animal or by tissue culture or by recombinant DNA adjuvants.

When a patient receives antibodies or fragments thereof that can bind to ICAM-1, or when a recipient patient receives ICAM-1 (or a fragment, variant or derivative thereof), the dosage of agent administered will vary depending on such factors as the patient's age, weight, height, sex, general medical condition, previous medical history, etc. Usually it is desirable that the recipient receives a dosage of antibody that is in the range from approx. 1 pg/kg to 10 mg/kg (body weight of patient), although a lower or higher dosage may be administered. When a patient receives ICAM-1 molecules or their functional derivatives, it is preferred to administer such molecules in a dosage that is also in the range of about 1 pg/kg to 10 mg/kg (body weight of patient), although a lower or higher dosage can also be administered. As discussed above, the therapeutically effective dose may be lowered if the anti-ICAM-1 antibody is co-administered with an anti-LFA-1 antibody. As used herein, a compound is said to be administered in addition to a second compound when the administration of the two compounds is so close in time that both compounds can be detected at the same time in the patient's serum.

Both the antibody that can bind to ICAM-1, and ICAM-1 itself, can be administered to patients intravenously, intramuscularly, subcutaneously, enterally, or parenterally. When administering antibody or ICAM-1 by injection, the administration can take place by continuous infusion or by single or multiple single doses.

The anti-inflammatory agent prepared according to the present invention is intended to be administered to recipients in an amount sufficient to suppress inflammation. An amount is said to be sufficient to "suppress" inflammation if the dosage, mode of administration, etc. of the agent is sufficient to attenuate or prevent inflammation.

Anti-ICAM-1 antibody or a fragment thereof can be administered either alone or in combination with one or more additional immunosuppressive agents (particularly to a recipient of an organ or tissue transplant). The administration of such compound(s) may be for either a "prophylactic" or "therapeutic" purpose. In prophylactic administration, the immunosuppressive compound(s) is given before any inflammatory response or symptom (e.g. before, at, or just after) the time of an organ or tissue transplant, but before any symptoms of organ failure.

The prophylactic administration of the compound(s) serves to prevent or attenuate any subsequent inflammatory response (such as rejection of a transplanted organ or tissue, etc.). In therapeutic administration, the immunosuppressive compound(s) is given at (or shortly after) the onset of a symptom of true inflammation (such as, for example, organ or tissue rejection). Therapeutic administration of the compound(s) serves to attenuate any true inflammation (such as rejection of a transplanted organ or tissue). The anti-inflammatory agents according to the present invention can thus be given either before the onset of inflammation (for suppression of a predicted inflammation) or after the start of inflammation.

A composition is said to be "pharmacologically satisfactory" if its administration can be tolerated by a recipient patient. Such an agent is said to be administered in a "therapeutically effective amount" if the amount administered is physiologically significant. An agent is physiologically significant if its presence results in a detectable change in the physiology of a recipient patient.

The antibody and the ICAM-1 molecules can be designed according to known methods for the preparation of pharmaceutically suitable mixtures, whereby these materials, or their functional derivatives, are combined in mixture with a pharmaceutically satisfactory carrier-excipient. Suitable excipients and their design, including other human proteins, e.g. human serum albumin, is e.g. described in Remington's Pharmaceutical Sciences (16th ed., A. Osol, ed., Mack, Easton PA (1980)). To form a pharmaceutically satisfactory composition suitable for effective administration, such compositions will contain an effective amount of anti-ICAM antibody or ICAM-1 molecule, or their functional derivatives, together with an appropriate amount of carrier excipient.

Additional pharmaceutical methods can be used to regulate the duration of the effect. Controlled release preparations can be obtained by using polymers for complex binding or absorption of anti-ICAM-1 antibody or ICAM-1, or their functional derivatives. The controlled release can be achieved by selecting appropriate macromolecules (e.g. polyesters, polyamino acids, polyvinyl, pyrrolidone, ethylene vinyl acetate, methyl cellulose, carboxymethyl cellulose or protamine sulfate), and the concentration of macromolecules as well as the methods of incorporation to control release. Another possible method for regulating the duration of action of controlled release preparations is to incorporate anti-ICAM-1 antibody or ICAM-1 molecules, or their functional derivatives, into particles of a polymeric material such as polyesters, polyamino acids, hydrogels, poly(lactic acid) or ethylene vinyl acetate copolymers. Instead of incorporating these agents into polymeric particles, it is alternatively possible to enclose these materials in microcapsules such as e.g. is prepared by co-aggregation techniques or by interfacial polymerization, e.g. respectively hydroxymethyl cellulose or gelatin microcapsules and poly(methyl methacrylate) microcapsules, or in colloidal drug delivery systems, e.g. liposomes, albumin microspheres, microemulsions, nanoparticles and nanocapsules or in macroemulsions. Such techniques are described in Remington's Pharmaceutical Sciences (1980).

While the invention has now been generally described, it will be more easily understood by reference to the following examples, which are given by way of illustration.

EXAMPLE 1

Cultivation of mammalian cells

Typically, the EBV-transformed and hybridoma cells were maintained in RMPI 1640 culture medium, supplemented with 20 mM L-glutamine, 50 µg/ml gentamycin, and 10% fetal bovine (or fetal calf) serum. The cells were grown at 37°C in a humidified atmosphere of 5% CO 2 and 95% air.

For generation of Epstein-Barr virus (EBV) transformants, T cell-depleted peripheral blood mononuclear cells were plated at 10<6>/ml in RPMI 1640 medium supplemented with 20% fetal calf serum (FCS), and 50 µg/ml gentamycin incubated for 16 hours with EBV-containing supernatant from B95-8 cells (D.A. Thorley-Lawson et al., J. Exper. Med., 146:495 (1977)). Cells in portions of 0.2 ml were placed in 10 microtiter wells. The medium was replaced with RPMI 1640 medium (supplemented with 20% fetal calf serum and 50 µg/ml gentamycin) until cell growth was observed. The cells grew in most wells and were expanded in the same medium. Phytohemagglutinin (PHA) blasts were established at 10<6> cells/ml in RPMI 1640 medium (supplemented with 20% fetal calf serum) containing a 1:800 dilution of PHA-P (Difco Laboratories, Inc. , Detroit, MI). PHA lines were expanded with interleukin 2-(IL-2)-conditioned medium and pulsed weekly with PHA (D.A. Cantrell et al., J. Exper. Med., 158 :1895 (1983)). The above procedure was described by T. Springer et al., J. Exper. Med., 160:1901-1918

(1984), which reference is incorporated herein by reference. Cells obtained by the above procedure are then examined with anti-LFA-1 antibodies to determine whether they express the LFA-1 antigen. Such antibodies are described by F. Sanchez-Madrid et__al. , J. Ex<p>er. With. , 158:1785 (1983).

EXAMPLE 2

Assays regarding cellular aggregation and adhesion

For assessment of the extent of cellular adhesion, it was

applied aggregation analyses. Cell lines used in such assays were washed twice with RPMI 1640 medium containing 5 mM Hepes buffer (Sigma Chemical Co., St. Louis) and resuspended to a concentration of 2 x 10 cells/ml. In flat-bottomed 96-well microtiter plates (No. 3596; Costar Cambridge, MA),

added 50 µl of appropriate monoclonal-antibody supernatant or 50 µl of complete medium with or without purified monoclonal antibodies, 50 µl of complete medium containing 200 ng/ml of the phorbol ester-phorbol myristate acetate (PMA) and 100 µl of cells at a concentration of 2x 10<6>cells/ml in complete medium. This gave a final concentration of 50 ng/ml PMA and 2 x 10<5> cells/well. The cells aggregated spontaneously, and the degree of aggregation was noted at different time points. Scores ranged from 0 to 5+, where 0 indicated essentially no cells clustered; 1+ indicated that less than 10% of the cells were in aggregates; 2+ indicate that

less than 50% of cells were aggregated; 3+ indicated that up to 100% of cells were in small, loose clusters; 4+ indicated that up to 100% of the cells were aggregated into larger clusters; and 5+ indicated that 100% of the cells were in large, very compact aggregates. To achieve a more quantitative assessment of cell adhesion, reagents and cells were added to 5 ml polystyrene tubes in the same order as above. The tubes were placed in a rack on a rotary shaker at 37°C. After 1 hour at approx. 200 rpm (revolutions per minute), 10 µl of the cell suspension was placed in a hemocytometer, and the number of free cells was quantified. Percent aggregation was determined by the following equation:

number of free cells

% aggregation = 100 x (1 )

number of entered cells

The number of cells entered in the above formula is the number of cells per ml in a control tube containing only cells and complete medium that had not been incubated. The number of free cells in the above equation is equal to the number of non-aggregated cells per ml from test tubes. The above procedures are described by R. Rothlein et__al. , J. Exper. Med.. 163 :1132-1149 (1986).

EXAMPLE 3

LFA-1-dependent cell aggregation

The qualitative aggregation assay described in Example 2 was performed using the Epstein-Barr transformed cell line JY. When PMA was added to the culture medium in the microtiter plates, aggregation of cells was observed. Time-lapse video recordings showed that the JY cells at the bottom of the microtiter wells were motile and showed active membrane ruffling and pseudopodia movement. Contact between the pseudopodia of neighboring cells often resulted in cell-cell adhesion. If the adhesion was long-lasting, the area of cell contact moved to the uropod. Contact could be maintained despite strong cell movements and jerking of the cells in opposite directions. The main difference between PMA-treated and untreated cells appeared to be in the stability of these contacts once formed. With PMA, cell clusters developed that grew in size as additional cells adhered to their periphery.

As a second aid for measuring adhesion, the quantitative analysis described in example 2 was used. The cell suspensions were shaken at 200 rpm for 2 hours, transferred to a hemocytometer, and cells not in aggregates were counted. In the absence of PMA, 42% (SD = 20%, N = 6) of JY cells were in aggregates after 2 h, whereas JY cells incubated under identical conditions with 50 ng/ml PMA had 87% (SD = 8%, N = 6) of the cells in aggregates. Kinetic investigations of the aggregation showed that PMA increased the amount and size of the aggregation at all the investigated time points (Figure 3).

EXAMPLE 4

Inhibition of aggregation of cells during use

of monoclonal anti-LFA-1 antibodies

To investigate the effects of anti-LFA-1 monoclonal antibodies on PMA-induced cell aggregation, such antibodies were added to cells incubated according to the qualitative aggregation assay of Example 2. The monoclonal antibodies were found to inhibit the formation of aggregates of cells either in presence or absence of PMA. Both the F(ab')2 and Fab' fragments of monoclonal antibodies against the alpha chain of LFA-1 could inhibit cell aggregation. Although approximately 100% of cells formed aggregates in the absence of anti-LFA-1 antibody, less than 20% of cells were found to be in aggregates when antibody was added. The results of this experiment are described by R. Rothlein et al.,

( J. Exper. Med. 163:1132-1149 (1986)).

EXAMPLE 5

Cell aggregation requires the LFA-1 receptor

EBV-transformed lymphoblastoid cells were prepared from patients as described in Example 1. Such cells were probed for monoclonal antibodies capable of recognizing LFA-1, and the cells were found to be deficient in LFA-1.

The qualitative aggregation assay described in Example 2 was applied using the LFA-1 deficient cells described above. Such cells did not spontaneously aggregate, even in the presence of PMA.

EXAMPLE 6

Discovery of ICAM-1

The LFA-1-deficient cells of Example 5 were labeled with carboxyfluorescein diacetate (M. Patarroyo et al., Cell. Immunol., 62:237-248 (1981)). The labeled cells were mixed at a ratio of 1:10 with autologous or JY cells, and the percentage of fluorescein-labeled cells in aggregates was determined according to the method of R. Rothlein et al., J. Ex<p>er. Med., 163:1132-149 (1986). The LFA-1-deficient cells were found to be able to co-aggregate with LFA-1-expressing cells (Figure 4).

To determine whether LFA-1 was only important in the formation of aggregates, or in maintaining them, antibodies that could bind to LFA-1 were added to the formed aggregates described above. The addition of antibody was found to strongly disrupt the formed aggregation. Time-lapse video recording confirmed that addition of the monoclonal antibodies to formed aggregates began to cause disruption within 2 hours (Table 1). After addition of monoclonal antibodies against LFA-1, pseudopodia movements and changes in the shape of individual cells in aggregates continued unchanged. Individual cells were gradually detached from the periphery of the aggregate; after 8 hours the majority of cells were dispersed. According to time-lapse video recording, the disruption of aggregates formed by monoclonal LFA-1 antibodies appeared equivalent to the aggregation process in the absence of monoclonal LFA-1 antibody when going back in time.

EXAMPLE 7

Requirement of divalent ions for LFA-1-dependent aggregation

LFA-1-dependent adhesions between cytotoxic T cells and target substances require the presence of magnesium (E. Martz, J. Cell. Biol., 84=584-598 (1980)). PMA-induced JY cell aggregation was examined for dependence on divalent cations. JY cells did not aggregate (using the assay of Example 2) in medium free of calcium or magnesium ions. Addition of divalent magnesium supported aggregation at concentrations as low as 0.3 mM. Addition of calcium ions alone had little effect. However, calcium ions were found to enhance the ability of magnesium ions to support PMA-induced aggregation. When 1.25 mM calcium ions were added to the medium, magnesium ion concentrations as low as 0.02 millimolar were found to support aggregation. These data show that the LFA-1-dependent aggregation of cells requires magnesium ions, and that calcium ions, although they are insufficient in themselves, can act synergistically with magnesium ions so that aggregation is made possible.

EXAMPLE 8

Isolation of hybridoma cells capable of expressing monoclonal anti-ICAM-1 antibodies

Monoclonal antibodies capable of binding to ICAM-1 were isolated according to the method of R. Rothlein et al., J. Immunol., 137:1270-1274 (1986), which reference is incorporated herein by reference. Thus, 3 BALB/C mice were immunized intraperitoneally with EBV-transformed peripheral blood mononuclear cells from an LFA-1-deficient individual (T.A. Springer et al., J. Ex<p>er. Med., 160:1901 ( 1984)). About. 10 cells in 1 ml of RPMI 1640 medium were used for each immunization. The immunization doses were administered 45, 29 and 4 days before spleen cells were removed from the mouse to generate the desired hybridoma cell lines. On day 3 before removal of the spleen cells, the mice received additional 10<7> cells in 0.15 ml of medium (intravenously).

Isolated spleen cells from the animals described above were fused with P3X73Ag8.653 myeloma cells at a ratio of 4:1 according to the protocol of G. Galfre et al., Nature, 266:550

(1977). Aliquots of the resulting hybridoma cells were introduced into 96-well microtiter plates. The hybridoma supernatants were screened for inhibition of aggregation, and one inhibitory hybridoma (among over 600 wells screened) was cloned and subcloned by limiting dilution. This subclone was designated RR1/1.1.1 (hereinafter referred to as "RR1/1").

Monoclonal antibody RR1/1 was consistently found to inhibit PMA-stimulated aggregation of the LFA-1-expressing cell line JY. The RR1/1 monoclonal antibody inhibited aggregation similarly to, or slightly less than, some monoclonal antibodies against the LFA-1 alpha or beta subunits. In contrast, aggregation was not inhibited by control monoclonal antibody against HLA, which is abundantly expressed on JY cells. The antigen that was bound by monoclonal antibody RR1/1 is defined as the intercellular adhesion molecule 1 (ICAM-1).

EXAMPLE 9

Use of monoclonal anti-ICAM-1 antibodies

for characterization of the ICAM-1 molecule

To determine the nature of ICAM-1, and particularly to determine whether ICAM-1 was distinct from LFA-1, cellular proteins were immunoprecipitated using monoclonal antibody RR1/1. The immunoprecipitation was carried out according to the method according to R. Rothlein et al., ( J Immunol., 137:1270-1274

(1986)). JY cells were lysed at 5 x 10<7> cells/ml in 1% Triton X-100, 0.14 mM NaCl, 10 mM Tris, pH 8.0, with freshly added 1 mM phenylmethylsulfonyl fluoride, 0.2 units per . ml trypsin inhibitor aprotinin (lysis buffer) for 20 minutes at 4°C. The lysates were centrifuged at 10,000 x g for 10 minutes and pre-clarified with 50 µl of a 50% suspension of CNBr-activated, glycine-stopped sepharose CL-4B for 1 hour at 4°C. 1 ml of lysate was immunoprecipitated with 20 µl of a 50% suspension of monoclonal antibody RR1/1 coupled to sepharose CL-4B (1 mg/ml) overnight at 4°C (T.A. Springer et al., J. Exper. Med ., 160:1901 (1984)). Sepharose-bound monoclonal antibody was prepared using DNBr activation of sepharose CL-4B in carbonate buffer according to the method of S. March et al., ( Anal. Biochem., 60:149

(1974)). Washed immunoprecipitates were subjected to SDS-PAGE and silver staining according to the method of J.H. Morrissey, Anal. Biochem., 117:307 (1981).

After elution of proteins with SDS sample buffer (M.K. Ho et al., J. Biol. Chem., 258:636 (1983)) at 100°C, the samples were divided into two halves and subjected to electrophoresis (SDS-8% PAGE) under reducing (Figure 5A) or non-reducing conditions (Figure 5B). Bands with molecular weight of 50 kd and 25 kd corresponded to the heavy and light chains of immunoglobulins from the Sepharose with the monoclonal antibody (Figure 5A, lane 3). Variable amounts of other bands in the 25-50 kd weight range were also observed, but were not seen in precipitates from hairy leukemia cells, which yielded only one band at molecular weight 90 kd. The 177 kd alpha subunit and the 95 kd beta subunit of LFA-1 were found to migrate differently than ICAM-1 under both reducing (Figure 5A, lane 2) and non-reducing (Figure 5B, lane 2) conditions.

To determine the effect of monoclonal antibody RR1/1 on PHA lymphoblast aggregation, the quantitative aggregation assay described in Example 2 was used. Thus, T cell blast cells were stimulated for 4 days with PHA, washed thoroughly and then cultured for 6 days in the presence of IL-2 conditioned medium. PHA was found to be internalized during this 6 day culture and did not contribute to the aggregation assay. In three different assays with different T-cell blast preparations, monoclonal ICAM-1 antibodies inhibited aggregation consistently (Table 2).

<a>Aggregation of PHA-induced lymphoblasts stimulated with 50 ng/ml PMA was quantified indirectly by microscopic counting of the number of non-aggregated cells as described in Example 2.

<b>Percent inhibition relative to cells treated with PMA and monoclonal X63 antibody.

<c>The aggregation was measured 1 hour after the simultaneous addition of monoclonal antibody and PMA. The cells were shaken at 175 rpm.

dA negative number indicates a percentage increase in the aggregation.

<e>The aggregation was measured 1 hour after the simultaneous addition of monoclonal antibody and PMA. The cells were pelleted at 200 x G for 1 minute, incubated at 37°C for 15 minutes, gently resuspended and shaken for 45 minutes at 100 rpm.

<f>The cells were pre-treated with PMA for 4 hours at 37°C. After monoclonal antibody was added, the tubes were incubated at 37°C quiescent for 20 minutes and shaken at 75 rpm for 100 minutes.

Monoclonal LFA-1 antibodies were consistently more inhibitory than monoclonal ICAM-1 antibodies, while monoclonal HLA-A, B and LFA-3 antibodies were without effect. These results show that among the investigated monoclonal antibodies, only those that could bind to LFA-1 or ICAM-1 could inhibit cell adhesion.

EXAMPLE 10

Preparation of monoclonal antibody against ICAM-1

Immunization

A Balb/C mouse was immunized intraperitoneally (i.p.) with 0.5 ml of 2 x IO<7>JY cells in RPMI medium 103 days and 24 days before fusion. On days 4 and 3 before fusion, mice were immunized i.p. with 10<7> cells of PMA-differentiated U937 cells in 0.5 ml RPMI medium.

Differentiation of U937 cells

U937 cells (ATCC CRL-1593) were differentiated by incubation at 5 x 10<5>/ml in RPMI with 10% fetal bovine serum, 1% glutamine and 50 µg/ml gentamycin (complete medium) containing 2 ng/ml phorbol-12-myristate acetate (PMA) in a sterile polypropylene container. On the third day of this incubation, half of the medium volume was removed and replaced with fresh complete medium containing PMA. On day 4, the cells were removed, washed and prepared for immunization.

Fusion on

Spleen cells from the immunized mice were fused with P3x63 Ag8 "653 myeloma cells at a ratio of 4:1 according to Galfre et al., (Nature, 266:550 (1977)). After the fusion, the cells were plated in flat-bottomed 96-well microtiter plates with IO<5>spleen cells/ well.

Selection with regard to anti-ICAM-1-positive cells

After one week, 50 µl of supernatant was examined by the qualitative aggregation assay according to example 2 using both JY and SKW-3 as aggregating cell lines. Cells from supernatants that inhibited JY cell aggregation but not SKW-3 were selected and cloned 2 times using limiting dilution.

This experiment resulted in the identification and cloning of three separate hybridoma lines that produced monoclonal anti-ICAM-1 antibodies. The antibodies formed by these hybridoma lines were IgG2a, <Ig>G2bog IgM, respectively. The hybridoma cell line that produced the IgG2a anti-ICAM-1 antibody was designated R6'5'D6'E9'B2. The antibody produced by the preferred hybridoma cell line was designated R6'5'D6'E9'B2 (hereinafter referred to as "R6-5-D6".

EXAMPLE 11

Expression and regulation of ICAM-1

For the measurement of ICAM-1 expression, a radioimmunoassay was developed. In this assay, purified RR1/1 was iodinated using iodogen to a specific activity of 10 µCi/ixg. Endothelial cells were grown in 96-well plates and treated as described for each experiment. Plates were cooled at 4°C by placing in a cold room for 0.5-1 hour, not immediately on ice. The monolayers were washed 3 times with cold complete medium and then incubated for 30 minutes at 4°C with 125 I-RR1/1. The monolayers were then waxed 3 times with complete medium. The bound 12<5>1 was released using 0.1 N NaOH and counted. The specific activity for<125>I-RR1/1 was adjusted using unlabeled RR1/1 to obtain a linear signal over the antigen density range found in this study. Non-specific binding was determined in the presence of a thousand-fold excess of unlabeled RR1/1 and was subtracted from the total binding to obtain specific binding.

ICAM-1 expression, measured using the radioimmunoassay described above, is increased on human umbilical vein endothelial cells (HUVEC) and human saphenous vein endothelial cells (HSVEC) by IL-1, TNF, LPS and IFN-gamma (Table 3 ). Saphenous vein endothelial cells were used in this study to confirm the results from umbilical vein endothelial cells in cultured adult large vein endothelial cells. The basal expression of ICAM-1 is twofold higher on saphenous vein endothelial cells than on umbilical vein endothelial cells. Exposure of umbilical vein endothelial cells to recombinant IL-l-alpha, IL-l-beta and TNF-gamma increases ICAM-1 expression 10-20 fold. IL-1-alpha, TNF and LPS were the most potent inducers, and IL-1 was less potent on a weight basis and also at saturating concentrations for the response (Table 3). IL-1-beta at 100 ng/ml increased ICAM-1 expression 9-fold on HUVEC and 7.3-fold on HSVEC, the half-maximal increase occurring at 15 ng/ml. rTNF at 50 ng/ml increased ICAM-1 expression 16-fold on HUVEC and 11-fold on HSVEC with half-maximal effects at 0.5 ng/ml. Interferon-gamma produced a significant increase in ICAM-1 expression of 5.2-fold on HUVEC or 3.5-fold on HSVEC at 10,000 U/ml. The effect of LPS at 10 µg/ml was of the same order of magnitude as the effect of rTNF. Pairwise combinations of these mediators resulted in additive or slightly less than additive effects on ICAM-1 expression (Table 3). Cross-titration of rTNF with rIL-l-beta or rIFN-gamma showed no synergism between these at suboptimal or optimal concentrations.

Since LPS increased ICAM-1 expression on endothelial cells at levels found in some cells in culture media, the possibility that the basal ICAM expression could be due to LPS was investigated. When several serum aliquots were examined, it was found that serum with low endotoxin content resulted in lower basal ICAM-1 expression by 25%. All results presented here were for endothelial cells cultured in low endotoxin serum. However, incorporation of the LPS-neutralizing antibiotic polymyxin B at 10 µg/ml reduced ICAM-1 expression by only an additional 25% (Table 3). The increase in ICAM-1 expression upon treatment with IL-1 or TNF was not affected by the presence of 10 µg/ml polymyxin B, which is consistent with the low endotoxin levels in these preparations (table 3).

Upregulation of ICAM-1 expression on HUVEC and HSVEC. HUVEC or HSVEC were seeded in 96-well plates at 1:3 from a confluent monolayer and allowed to grow to confluence. The cells were then treated with the indicated materials or media for 16 hours, and the RIA was performed according to the described procedure. All points were performed with four parallels.

EXAMPLE 12

Kinetics of interleukin-1 and

gamma interferon induction of ICAM-1

The kinetics of interleukin-1 and gamma interferon effects on ICAM-1 expression on skin fibroblasts were determined using the<125>I goat anti-mouse IgG binding assay according to M.L. Dustin et al., (J. Immunol., 137:245-254 (1986); which reference is incorporated herein by reference). To perform this binding assay, human skin fibroblasts were cultured in a 96-well microtiter plate to a density of 2-8 x 10<4> cells/well (0.32 cm<2>). The cells were washed twice with RPMI 1640 medium supplemented as described in Example 1. The cells were additionally washed once with Hank's Balanced Salt Solution (HBSS), 10 mM HEPES, 0.05% NaN 3 and 10% heat-inactivated fetal bovine serum. Washing with this binding buffer was performed at 4°C. In each well was added 50 µl of the binding buffer described above and 50 µl of the appropriate hybridoma supernatant with X63 and W6/32 as negative and positive control, respectively. After incubation for 30 minutes at 4°C with gentle agitation, the wells were washed twice with binding buffer, and the second antibody 125I-goat - anti-mouse IgG was added at 50 nCi in 100/xl.<125>I-goat. The -antimouse antibody was prepared using iodogen (Pierce) according to the method of P.J. Fraker et al., (Biochem. Biophys. Res. Commun., £0:849 (1978)). After 30 minutes at 4°C, the wells were washed twice with 200 µl binding buffer, and the cell layer was solubilized by adding 100 µl 0.1 N NaOH. This and a 100/xl wash were counted in a Beckman 5500 gamma counter. The specific counts per minute was calculated as [cpm with monoclonal antibody]-[cpm with X63]. All steps, including induction with specific reagents, were performed in four parallels.

The effect of interleukin 1 with a half-life for ICAM-1 induction of 2 hours was faster than the effect of gamma interferon with a half-life of 3.75 hours (Figure 6). The time course for the return to resting levels of ICAM-1 was found to depend on the cell cycle or the rate of cell growth. In quiescent cells, interleukin 1 and gamma interferon effects are stable for 2-3 days, while ICAM-1 expression in log-phase cultures is close to baseline 2 days after removal of these inducers.

The dose-response curves for the induction of ICAM-1 by recombinant mouse and human interleukin 1, and for recombinant human gamma interferon, are shown in Figure 7. Gamma interferon and interleukin 1 were found to have the same concentration dependences with almost identical effects at 1 ng / ml. The recombinant human and mouse interleukin 1 also have similar curves, but are much less effective than preparations of human interleukin 1 in inducing ICAM-1 expression.

Cyclohexamide, an inhibitor of protein synthesis, and actinomycin D, an inhibitor of mRNA synthesis, abrogate the effects of both interleukin 1 and gamma interferon on ICAM-1 expression on fibroblasts (Table 4). Furthermore, tunicamycin inhibited a

inhibitor of N-linked glycosylation, only the interleukin 1 effect by 43%. These results demonstrate that protein and mRNA synthesis, but not N-linked glycosylation, is required for interleukin 1- and gamma interferon-stimulated increases in ICAM-1 expression.

<a>Human fibroblasts were grown to a density of 8 x 10<4>cells/0.32 cm<2>well. The treatments were carried out in a final volume of 50 µl containing the indicated reagents. Cycloheximide, actinomycin D and tunicamycin were added at concentrations of 20 µg/ml, 10 µg/ml and 2 µg/ml, respectively, at the same time as the cytokines. All points are mean values of four parallel wells + SD.

EXAMPLE 13

Tissue distribution of ICAM-1

Histochemical studies were performed on frozen tissues from human organs to determine the distribution of ICAM-1 in the thymus, lymph nodes, intestines, skin, kidneys and liver. To perform such an analysis, frozen tissue sections (4 µm thick) of normal human tissues were fixed in acetone for 10 minutes and stained with the monoclonal antibody RR1/1 by an immunoperoxidase technique using the avidin-biotin complex method (Vector Laboratories, Burlingame, CA) described by N. Cerf-Bensussan et al., (J. Immunol., 130:2615 (1983)). After incubation with the antibody, the sections were sequentially incubated with biotinylated horse anti-mouse IgG and avidin-biotinylated peroxidase complexes. The sections were finally dipped in a solution containing 3-amino-9-ethylcarbazole (Aldrich Chemical Co., Inc. Milwaukee WI) to produce a color reaction. The sections were then fixed in 4% formaldehyde for 5 minutes and counterstained with hematoxylin. Control included sections incubated with unrelated monoclonal antibodies instead of the RRl/l antibody.

ICAM-1 was found to have a distribution most similar to that of major histocompatibility complex (MHC) class II antigens. Most of the blood vessels (both small and large) in all tissues showed staining of endothelial cells with ICAM-1 antibody. Vascular-endothelial staining was more intense in interfollicular (paracortical) areas in lymph nodes, tonsils and Peyer's patches, compared to vessels in kidney, liver and normal skin. In the liver, staining was mostly restricted to sinusoidal-covering cells; the hepatocytes and endothelial cells that line most of the portal veins and

-arteries, were not stained.

Diffuse staining of large cells and a dendritic staining pattern were observed in the thymus marrow. In the cortex, the staining pattern was focal and mainly dendritic. Thymocytes were not stained. In the peripheral lymphoid tissue, the germinal center cells in the secondary lymphoid follicles were intensely stained. In some lymphoid follicles, the staining pattern was mostly dendritic, with no recognizable staining of the lymphocytes. Weak staining of cells in the cover zone was also observed. In addition, dendritic cells with cytoplasmic extensions ("middle finger" reticulum cells) and a small number of lymphocytes in the interfollicular or paracortical areas were stained with the ICAM-1 binding antibody.

Cells resembling macrophages were stained in the lymph nodes and lamina propria of the small intestine. Fibroblast-like cells (spindle-shaped cells) and dendritic cells scattered in the stroma of most examined organs were stained with the ICAM-1-binding antibody. No staining could be seen in the long-glove/indeterminate cells of the epidermis. No staining was observed in smooth muscle tissue.

The staining of epithelial cells was also seen in the mucous membranes of the tonsils. Although hepatocytes, bile duct epithelium, intestinal epithelial cells, and tubular epithelial cells of the kidney were not stained under most circumstances, sections of normal kidney tissue from a specimen of resected kidney with renal cell carcinoma showed staining of many proximal tubule cells for ICAM-1 . These tubule epithelial cells were also stained with an anti-HLA-DR binding antibody.

In summary, ICAM-1 is expressed on non-hematopoietic cells such as vascular endothelial cells, and on hematopoietic cells such as tissue macrophages and mitogen-stimulated T-lymphocyte blasts. ICAM-1 was found to be expressed in small amounts on peripheral blood lymphocytes.

EXAMPLE 14

Purification of ICAM-1 by monoclonal-antibody affinity chromatography

General cleaning plan

ICAM-1 was purified from human cells or tissues using monoclonal antibody affinity chromatography. Monoclonal antibody RR1/1, which was reactive with ICAM-1, was first purified and coupled to an inert column matrix. This antibody is described by R. Rothlein et al.. J. Immunol., 137:1270-1274 (1986), and M.L. Dustin et al., J. Immunol.. 137:245 (1986). ICAM-1 was solubilized from cell membranes by lysing the cells in a nonionic detergent, Triton X-100, at a near-neutral pH. The cell lysate containing solubilized ICAM-1 was then passed through pre-columns designed to remove materials that bind non-specifically to the column matrix material, and then through the column matrix with the monoclonal antibody, allowing ICAM-1 to bind to the antibody. The antibody column was then washed with a series of wash buffers of increasing pH up to pH 11.0. During these washes, ICAM-1 remained bound to the antibody matrix, while non-binding and weakly binding contaminants were removed. The bound ICAM-1 was then specifically eluted from the column using a detergent buffer of pH 12.5.

Purification of monoclonal antibody RRl/ 1 and covalent coupling to sepharose CL-4B

The anti-ICAM-1 monoclonal antibody RR1/1 was purified from the peritoneal fluid of hybridoma-bearing mice, or from hybridoma culture supernatants by standard techniques of ammonium sulfate precipitation and protein A affinity chromatography (Ey et al., Immunochem., 15:429 (1978)). The purified IgG, or rat IgG (Sigma Chemical Co., St. Louis, MO) was covalently coupled to sepharose CL-4B (Pharmacia, Upsala, Sweden) using a modification of the method of March et al., ( Anal .Biochem., 6.0:149 (1974)). Briefly, sepharose CL-4B was washed in distilled water, activated with 40 mg/ml CNBr in 5 M K 2 HPO 4 (pH ca. 12) for 5 min, and then washed thoroughly with 0.1 mM HCl at 4°C. The filtered, activated sepharose was resuspended with an equal volume of purified antibody (2-10 mg/ml in 0.1 M NaHCO 3 , 0.1 M NaCl). The suspension was incubated for 18 hours at 4°C with gentle end-to-end rotation. The supernatant was then checked for unbound antibody by absorbance at 280 nm, and remaining reactive sites on the activated sepharose were saturated by addition of glycine to 0.05 M. The coupling efficiency was usually greater than 90%.

Detergent solubilization of membranes prepared from human spleen

All procedures were performed at 4°C. Frozen human spleen (200 g fragments) from patients with hairy cell leukemia was thawed on ice in 200 ml Tris-saline solution (50 mM Tris, 0.14 M NaCl, pH7.4 at 4°C) containing 1 mM phenylmethylsulfonyl fluoride ( PMSF), 0.2 U/ml aprotinin and 5 mM iodoacetamide. The tissue was cut into small pieces and homogenized at 4°C with a Tekmar machine homogenizer. The volume was then brought to 300 ml with Tris-saline solution, and 100 ml of 10% Tween 40 (polyoxyethylene sorbitan monopalmitate) in Tris-saline solution was added to achieve a final concentration of 2.5% Tween 40.

For the preparation of membranes, the homogenate was extracted using three strokes of a Teflon Dounce or more preferably Potter-Elvehjem homogenizer, and then centrifuged at 1000 x g for 15 minutes. The supernatant was retained and the clot was re-extracted with 200 ml of 2.5% Tween 40 in Tris-saline. After centrifugation at 1000 x g for 15 min, the supernatants from both extractions were mixed and centrifuged at 150,000 x g for 1 h to pellet the membranes. The membranes were washed by resuspension in 200 ml Tris-saline solution, centrifuged at 150,000 x g for 1 hour. The membrane pellet was resuspended in 200 ml Tris-saline solution and homogenized with a motorized homogenizer and Teflon pestle until the suspension was uniformly cloudy. The volume was then brought up to 900 ml with Tris-saline solution, and N-lauroylsarcosine was added to a final concentration of 1%. After stirring at 4°C for 30 minutes, insoluble material in the detergent lysate was removed by centrifugation at 150,000 x g for 1 hour. Triton X-100 was then added to the supernatant to a final concentration of 2%, and the lysate was stirred at 4°C for 1 hour.

Detergent solubilization of JY-B lymphoblastoid cells

The EBV-transformed B lymphoblastoid cell line JY was cultured in RPMI-1640 containing 10% fetal calf serum (FCS)

and 10 mM HEPES to an approximate density of 0.8-1.0 x 10 cells/ml. To increase the cell surface expression of ICAM-1, phorbol-12-myristate-13-acetate (PMA) was added at 25 ng/ml for 8-12 hours before harvesting the cells. Sodium vanadate (50/xM) was also added to the cultures during this time. The cells were pelleted by centrifugation at 500 x g for 10 minutes and washed twice in Hank's Balanced Salt Solution (HBSS) by resuspension and centrifugation. The cells (approx. 5 g per 5 liter culture) were lysed in 50 ml lysis buffer (0.14 M NaCl, 50 mM Tris pH 8.0, 1% Triton X-100, 0.2 U/ml aprotinin, 1 mM PMSF, 50fiM

sodium vanadate) by stirring at 4°C for 30 minutes. Non-lysed nuclei and insoluble cell debris were removed by centrifugation at 10,000 x g for 15 min, followed by centrifugation of the supernatant at 150,000 x g for 1 h, and filtration of the supernatant through Whatman 3 mm filter paper.

Affinity chromatography of ICAM-1 for structural investigations

For large-scale purification of ICAM-1 for use in structural studies, a 10 ml column of RRl/1-sepharose CL-4B (coupled with 2.5 mg antibody/ml gel) and two 10 ml pre-columns were used of CNBr-activated glycine-stopped sepharose CL-4B and rat IgG coupled to sepharose CL-4B (2 mg/ml). The columns were connected in series and pre-washed with 10 column volumes of lysis buffer, 10 column volumes of pH 12.5 buffer (50 mM triethylamine, 0.1% Triton X-100, pH 12.5 at 4°C), followed by equilibration with 10 column volumes of light buffer. One liter of the human spleen detergent lysate was applied at a flow rate of 0.5-1.0 ml/minute. The two pre-columns were used to remove non-specific binding material from the lysate prior to passage through the RR1/1 Sepharose column.

After loading, the column of RRl/l Sepharose and bound ICAM-1 was washed sequentially at a flow rate of 1 ml/min with a minimum of 5 column volumes of each of the following: 1) lysis buffer, 2) 20 mM Tris, pH 8', 0 /0.14 M NaCl/0.1% Triton X-100, 3) 20 mM glycine pH 10.0/0.1% Triton X-100 and 4) 50 mM triethylamine, pH 11.0/0.1% Triton X-100. All wash buffers contained 1 mM PMSF and 0.2 U/ml aprotinin. After washing, the remaining bound ICAM-1 was eluted with 5 column volumes of elution buffer (50 mM triethylamine/0.1% Triton X-100/pH 12.5 at 4°C) at a flow rate of 1 ml/3 minutes. The eluted ICAM-1 was collected in 1 ml fractions and immediately neutralized by the addition of 0.1 ml of 1 M Tris, pH 6.7. Fractions containing ICAM-1 were identified by SDS-polyacrylamide electrophoresis of 10 µl aliquots (J. Springer et al., Exp. Med., 160:1901 (1984)), followed by silver staining (J.H. Morrissey, Anal Biochem., 117:307 (1981)). Under these conditions, the main amount of ICAM-1 was eluted in approx. 1 column volume and had a purity of more than 90%, as judged by silver-stained electropherograms (a small amount of IgG, eluted from the affinity matrix, was the main contaminant). The fractions containing ICAM-1 were mixed and concentrated approx. 20 times using Centricon 30 microconcentration devices (Amicon, Danvers, MA). The purified ICAM-1 was quantified by Lowry protein analysis of an ethanol-precipitated portion of the mixture: approx. 500 µg of pure ICAM-1 was prepared from the 200 g of human spleen.

About. 200/xg purified ICAM-1 was subjected to a second-step purification by preparative SDS-polyacrylamide gel electrophoresis. The band representing ICAM-1 was visualized by loading the gel in 1 M KCl. The region of the gel containing ICAM-1 was then excised and electroeluted according to the method of Hunkapiller et al., Meth. Enzymol., 91:227:236 (1983). The purified protein had a purity greater than 98%, as judged by SDS-PAGE and silver plating.

Affinity purification of ICAM-1 for functional studies

ICAM-1 for use in functional studies was purified from detergent lysates of JY cells as described above, but on a smaller scale (a 1 ml column of RRl/l Sepharose), and with the following modifications. All solutions contained 50 µM sodium vanadate. After washing the column with pH 11.0 buffer containing 0.1% Triton X-100, the column was washed again with five column volumes of the same buffer containing 1% n-octyl-beta-D-glucopyranoside (octylglucoside) instead for 0.1% Triton X-100. The octylglucoside detergent replaces Triton X-100 bound to ICAM-1, and unlike Triton X-100 can be removed afterwards by dialysis. ICAM-1 was then eluted with buffer of pH 12.5 containing 1% octylglucoside instead of 0.1% Triton X-100, and was analyzed and concentrated as described above.

EXAMPLE 15

Characterization of purified ICAM-1

ICAM-1 purified from human spleen migrates in SDS-polyacrylamide gels as a broad band with Mr 72,000-91,000. ICAM-1 purified from JY cells also migrates as a broad band with Mr 76,500-97,000. These Mr are within the reported range for ICAM-1 immunoprecipitated from different cell sources: Mr=90,000 for JY cells, 114,000 for the myelomonocytic cell line U937, and 97,000 for fibroblasts (Dustin et al., J. Immunol., 137:245 ( 1986)). This wide range of Mr has been attributed to an extensive but variable degree of glycosylation. The non-glycosylated precursor has an Mr of 55,000 (Dustin et al.). The protein purified from either JY cells or human spleen retains its antigenic activity, as evidenced by its ability to rebind to the original affinity column, and by immunoprecipitation with RRl/l Sepharose and SDS-polyacrylamide electrophoresis.

For the production of peptide fragments of ICAM-1, approx. 200 µg reduced with 2 mM dithiothreitol/2% SDS, followed by alkylation with 5 mM iodoacetic acid. The protein was precipitated with ethanol, and redissolved in 0.1 M NH4CO3/0.1 mM CaCl2/0.1% zwittergent 3-14 (Calbiochem), and cleaved with 1 wt% trypsin at 37°C for 4 hours , followed by a further digestion with 1% trypsin for 12 hours at 37°C. The tryptic peptides were purified by reversed-phase HPLC using a 0.4 x 15 cm C4 column (Vydac). The peptides were eluted with a linear gradient from 0 to 60% acetonitrile in 0.1% trifluoroacetic acid. Selected peptides were subjected to sequence analysis on a gas-phase micro-sequencer (Applied Biosystems). The sequence information obtained in this study is shown in Table 5.

EXAMPLE 16

Cloning of the ICAM-1 gene

The gene for ICAM-1 can be cloned using various methods. E.g. the amino acid sequence information obtained by sequencing the tryptic fragments of ICAM-1 (table 5) can be used for the identification of an oligonucleotide sequence that would correspond to the ICAM-1 gene. Alternatively, the ICAM-1 gene can be cloned using anti-ICAM-1 antibody to detect clones that produce ICAM-1.

Cloning of the gene for ICAM-1 using oligonucleotide probes

Using the genetic code (J.D. Watson, I: Molecular Biolo<g>y of the Gene, 3rd ed., W.A. Benjamin, Inc., Menlo Park, CA (1977)), one or more different oligonucleotides can be identified, each of which would encode the tryptic ICAM-1 peptides. The likelihood that a particular oligonucleotide will actually constitute the true ICAM-1 coding sequence can be assessed by considering abnormal base-pairing ratios and the frequency with which a particular codon is actually used (to code for a particular amino acid) in eukaryotic cells. Such "codon usage rules" are described by R. Lathe et al., J. Molec. Biol., 183:12 (1985). By applying the "codon usage rules" according to Lathe, a single oligonucleotide, or a set of oligonucleotides, is identified which contains a theoretically "most likely" nucleotide sequence (i.e. the least abundant nucleotide sequence) that can code for the tryptic ICAM-1 -peptide sequences.

The oligonucleotide, or the set of oligonucleotides, which contains the theoretically "most likely" sequence that can code for the ICAM-1 fragments, is used to identify the sequence of a complementary oligonucleotide or set of oligonucleotides that can be hybridized to the "most likely" sequence, or set of sequences. An oligonucleotide containing such a complementary sequence can be used as a probe for the identification and isolation of the ICAM-1 gene (T. Maniatis et al.. Molecular Cloning A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor, NY (1982) .

As described under part C above, it is possible to clone the ICAM-1 gene from eukaryotic DNA preparations presumed to contain this gene. For identification and cloning of the gene encoding the ICAM-1 protein, a DNA library is screened for its ability to hybridize with the oligonucleotide probes described above. Because of. that it is likely that there will only be two copies of the gene for ICAM-1 in a normal diploid cell, and because that it is possible that the ICAM-1 gene may have large non-transcribed intervening sequences (introns) if cloning is not desirable, it is preferred to isolate ICAM-1 coding sequences from a cDNA library prepared from mRNA of an ICAM- 1-forming cell, rather than from genomic DNA. Suitable DNA or cDNA preparations are enzymatically cleaved, or randomly cut and ligated into recombinant vectors. The ability of these recombinant vectors to hybridize to the above described oligonucleotide probes is then measured. Procedures for hybridization are e.g. described in T. Maniatis Molecular Cloninq A Laboratory Manual. Cold Spring Harbor Press, Cold Spring Harbor, NY (1982) or in B.T. Haymes et al., Nucleic Acid Hybridization a Practical Approach, IRL Press, Oxford, England (1985). Vectors found to be capable of such hybridization are then analyzed to determine the extent and nature of the ICAM-1 sequences they contain. Based on purely statistical considerations, a gene such as that encoding the ICAM-1 molecule could be unequivocally identified (via hybridization assay) using an oligonucleotide probe of only 18 nucleotides.

Thus, in brief, the actual identification of ICAM-1 peptide sequences enables the identification of a theoretical "most likely" DNA sequence, or set of such sequences, which may code for such a peptide. By constructing an oligonucleotide that is complementary to this theoretical sequence (or by constructing a set of oligonucleotides that is complementary to the set of "most likely" oligonucleotides), one obtains a DNA molecule (or a set of DNA molecules) that can function as a probe for the identification and isolation of the ICAM-1 gene.

Using the ICAM-1 peptide sequences in Table 5, the sequence of the "most likely" sequence of an oligonucleotide that could encode the AA and J peptides was determined (Tables 6 and 7, respectively). Oligonucleotides complementary to these sequences were synthesized and purified for use as probes for the isolation of ICAM-1 gene sequences. Appropriate size-selected cDNA libraries were generated from poly(A)<+->RNA for both PMA-induced HL-60 cells and from PS-stimulated umbilical vein endothelial cells. A size-selected cDNA library was prepared using poly(A)<+->RNA from PMA-induced HL-60 cells according to the method of U. Gubler et al., ( Gene, 25:263-269 (1983)) and A. Corbi et al., (EMBO J., 6:4023-4028 (1987)), which references are incorporated herein by reference.

A size-selected cDNA library was prepared using poly(A)<+->RNA from umbilical vein endothelial cells that had been stimulated for 4 hours with PS 5/tg/ml. RNA was extracted by homogenizing the cells in 4 M guanidine isothiocyanate and the supernatant was subjected to ultracentrifugation through a CsCl gradient (J.M. Chirgwin et al., Biochem., 18:5294-5299 (1979)). Poly(A)<+->RNA was isolated from the mixture of the total amount of RNA species using oligo-(dT)-cellulose chromatography (type 3, Collaborative Research) (H. Aviv et al., Proe. Nati. Acad. Sei.

( USA), 69:1408-1412 (1972)).

First strand cDNA was synthesized using 8 fig poly(A)<+->RNA, avian myeloblastosis virus reverse transcriptase (Life Sciences) and an oligo(dT) primer. The DNA-RNA hybrid was digested with RNase H (BRL), and the second strand was synthesized using DNA polymerase I (New England Biolabs). The product was methylated with Eco R1 methylase (New England Biolabs), blunt-end ligated to EcoR1 linkers (New England Biolabs), cleaved with Eco Ri and size-selected on a low-melting point agarose gel. cDNA greater than 500 bp was ligated to gt10 that had previously been Eco R1 digested and dephosphorylated (Stratagen). The product of the ligation was then packaged (Stratagen gold).

Umbilical vein endothelial cell and HL-60 cDNA libraries were then plated at 20,000 PFU/150 mm plate. Recombinant DNA was transferred in duplicate to nitrocellulose filters, denatured in 0.5 M NaOH/1.5 M NaCl, neutralized in 1 M Tris, pH 7.5/1.5 M NaCl and baked at 80°C for 2 h ( W.D. Benton et al., Science, 196:180-182 (1977) ). The filters were prehybridized and hybridized in 5X SSC containing 5X Denhardt's solution, 50 mM NaPO 4 and 1 µg/ml salmon sperm DNA.

Pre-hybridization was performed at 45°C for 1 hour.

Hybridization was performed using 32 bp (5'-TTGGGCTGGTCACAGGAGGTGGAGCAGGTGAC) or 47 bp (5'-GAGGTGTTCTC-AAACAGCTCCAGGCCCTGGGGCGCCAGGTCCAGCTC) anti-sense oligonucleotides based, as discussed above, on the ICAM-1 tryptic peptides, respectively. and AA (Tables 6 and 7) (R. Lathe, J. Molec. Biol.. 183:1-12 (1985)). The oligonucleotides were end-labeled with Y"(<32p>(AT<p>) using T4 polynucleotide kinase and conditions recommended by the manufacturer (New England Biolabs). After overnight hybridization, the filters were washed twice with 2X SSC/0.1% SDS for 30 minutes at 45° C. Phage were isolated from the plaques showing hybridization and were purified by successive replating and reprobing.

Cloning of the gene for ICAM-1 using anti-ICAM-1 antibody f

Alternatively, the gene for ICAM-1 can be cloned using anti-ICAM-1 antibody. DNA, or more preferably cDNA, is extracted and purified from a cell capable of expressing ICAM-1. The purified cDNA is fragmented (by cutting, endonuclease cleavage, etc.) to form a mixture of DNA or cDNA fragments. DNA or cDNA fragments from this mixture are then cloned into an expression vector to form a genomic library of expression vectors whose elements each contain a unique cloned DNA or cDNA fragment.

EXAMPLE 17

Analysis of the cDNA clones

Phage DNA from positive clones was digested with EcoRl and examined by Southern analysis using a cDNA from one clone as a probe. Maximum size cDNA inserts that cross-hybridized were subcloned into the EcoR1 site in plasmid vector pGEM 4Z (Promega). HL-60 subclones containing cDNA in both orientations were eliminated by exonuclease III digestion (S. Henikoff, Gene, 28.:351-359 (1984)) according to the manufacturer's recommendations (Erase-a-Base, Promega) . Progressively eliminated cDNA molecules were then cloned and subjected to dideoxynucleotide chain termination sequencing (F. Sanger et al., Proe. Nati. Acad. Sei ( USA). 74:5463-5467 (1977)) according to the manufacturer's recommendations (Sequenase, U.S. Biochemical). The HL-60 cDNA 5' and coding regions were completely sequenced on both strands, and the 3' region was sequenced approx. 70% on both chains. A representative endothelial cDNA was sequenced over the majority of its length by rapid cloning of 4 bp recognition restriction enzyme fragments.

The cDNA sequence from one HL-60 and one endothelial cell cDNA was generated (Figure 8). The 3023 bp sequence contains a short 5' untranslated region and a 1.3 kb 3' untranslated region with a consensus polyadenylation signal at position 2966. The longest open reading frame begins with the first ATG at position 58 and ends with a TGA termination triplet at position 1653. The identity between the translated amino acid sequence and sequences determined from 8 different tryptic peptides totaling 91 amino acids (underlined in Figure 8) confirmed that authentic ICAM-1 cDNA clones had been isolated. The amino acid sequences of tryptic ICAM-1 peptides are shown in Table 8.

Hydrophobicity analysis (J. Kyte et al., J. Molec. Biol., 157:105-132 (1982)) suggests the presence of a 27-residue signal sequence. The determination of the +1-glutamine is consistent with our inability to obtain the N-terminal sequence for three different ICAM-1 protein preparations; glutamine can cyclize to pyroglutamic acid, resulting in a blocked N-terminus. The translated sequence from 1 to 453 is mainly hydrophilic, followed by a 24-residue hydrophobic putative transmembrane domain. The transmembrane field is immediately followed by several charged residues found in a 27-residue putative cytoplasmic field.

The predicted size of the fully unfolded polypeptide chain is 55,219 daltons, in excellent agreement with the observed size of 55,000 for deglycosylated ICAM-1 (M.L. Dustin et al.,

J. Immunol.. 137:245-254 (1986)). Eight N-bonded

glycosylation sites are predicted. Absence of asparagine in the tryptic peptide sequences in two of these sites confirms their glycosylation and their extra-cellular orientation. Assuming 2,500 daltons per N-linked carbohydrate with a high content

of mannose, a size of 75,000 daltons is predicted for the ICAM-1 precursor, compared to the observed six of 73,000 daltons (M.L. Dustin et__al. , J. Immunol. , 137:245-254 (1986)). After conversion of high-mannose to complex carbohydrate, the fully developed ICAM-1 glycoprotein is 76-114 kd, depending on the cell type (M.L. Dustin et al., J. Immunol., 137:245-254 (1986)). Thus, ICAM-1 is a highly glycosylated, but otherwise typical, integral membrane protein.

EXAMPLE 18

ICAM-1 is an integrin-binding member of the immunoglobulin supergene family

Alignment of internal ICAM-1 repeats was performed using the Microgenie protein alignment program (C. Queen et al., Nucl. Acid Res., 12:581-599 (1984)) followed by inspection. Alignment of ICAM-1 with IgM, N-CAM and MAG was performed using the Microgenie and ALIGN program (M.O. Dayhoff et al., Meth. Enzvmol. , .91:525-545 (1983)). Four protein sequence databases maintained by the National Biomedical Research Foundation were examined for protein sequence similarities using the FASTP program of Williams and Pearson (D.J. Lipman et al., Science, 227:1435-1439 (1985)).

Since ICAM-1 is a ligand of an integrin, it was unexpected that it would be a member of the immunoglobulin supergene family. However, inspection of the ICAM-1 sequence shows that it fulfills all criteria proposed for membership in the immunoglobulin supergene family. These criteria are discussed below.

The entire extracellular domain of ICAM-1 is constructed of 5 homologous immunoglobulin-like regions shown aligned in Figure 9A. Fields 1-4 are respectively 88, 97, 99 and 99 residues long and thus have typical Ig field size; field 5 is truncated within 68 residues. Examination of the NBRF database using the FASTP program revealed significant homologies with members of the immunoglobulin supergene family including IgM and IgG C domains, variable T cell receptor α subunit domain, and alpha-l-beta-glycoprotein (Figure 9B-D ).

Using the above information, the amino acid sequence of ICAM-1 was compared with the amino acid sequences of other members of the immunoglobulin supergene family.

Three types of Ig superfamily domains, V, Cl and C2 have been differentiated. Both V and C fields are constructed of 2 6 layers

linked together by the intrafield disulfide bond; V domains contain 9 anti-parallel S chains, while C domains have 7. Constant domains were divided into C1 and C2 sets based on characteristic residues shown in Figure 9A. The Cl set includes proteins that are part of wood

antigen perception. The C2 set includes several Fc receptors and proteins involved in cell adhesion including CD2, LFA-3, MAG and NCAM. ICAM-1 fields were found to be most homogeneous with fields in the C2 set when placing ICAM-1 in this set; this is reflected by stronger similarity to conserved residues in C2 than Cl fields as shown for S-chains B-F in Figure 9. Moreover, ICAM-1 fields matched much better with S-chains A and G in C2 fields than with these chains in V and Cl fields, which made possible good arrangements over the entire C2 area. Arrangements with C2 fields from NCAM, MAG and alpha-1-S-glycoprotein are shown in Figures 9B and 9C; the identity was in the range from 28 to 33%. Arrays with a T-cell receptor Va 27% identity and IgM C-field 3-34% identity are also shown (Figures 9B, 9D).

One of the most important characteristics of immunoglobulin fields is the disulfide-bonded cysteines that bind together the B and F-S chains that stabilize the S-layer formation; in ICAM-1, the cysteines are conserved in all cases except in chain f in field 4, where a leucine amino acid is present that may face the layer formation and stabilize the contact as suggested for some other V and C2 set fields. The distance between the cysteine amino acids (43, 50, 52 and 37 residues) is as described for the C2 set.

To examine the presence of intrachain disulfide bonds in ICAM-1, endothelial cell ICAM-1 was subjected to SDS-PAGE under reducing and non-reducing conditions. Endothelial cell ICAM-1 was used because it shows less glycosylation heterogeneity than JY or hair-cell-spleen ICAM-1 and enables greater sensitivity to shifts in Mr. ICAM-1 was therefore purified from cultures of umbilical vein endothelial cells stimulated in 16 hours with LPS

(5 izg/ml), by immuno-affinity chromatography as described above. Acetone-precipitated ICAM-1 was resuspended in sample buffer (U.K. Laemmli, Nature. 227:680-685 (1970)) with 0.25% 2-mercaptoethanol or 25 mM iodoacetamide and brought to 100°C for 5 minutes. The samples were subjected to SDS-PAGE 4670 and silver staining 4613. Endothelial cell ICAM-1 had an apparent Mr of 100 kd under reducing conditions and 96 kd under non-reducing conditions, strongly suggesting the presence of intrachain disulfides in native ICAM-1 .

Application of the primary sequence for secondary structure prediction (P.Y. Chou et al., Biochem., 11:211-245 (1974)) showed the 7 expected fé chains in each ICAM-1 domain, labeled a-g in Figure 9A top, which exactly fulfilled the prediction of an immunoglobulin field, corresponding to the positions of chains A-H in immunoglobulins (Figure 9A, lower). Field 5 lacks the A and C chains, but since these form edges of the layers, the layers could still be formed, possibly replacing chain D with chain C, as suggested for some other C2 fields; and the characteristic disulfide bond between the B and F chains would be unaffected. Thus, all the criteria for field size, sequence homology, conserved cysteine amino acids forming the putative intrafield disulfide bond, presence of disulfide bonds and presumed S-layer structure are met for incorporation of ICAM-1 into the immunoglobulin supergene family.

ICAM1 was found to be most highly homologous to NCAM and MAG glycoproteins in the C2 set. This is of particular interest since both NCAM and MAG mediate cell-cell adhesion. NCAM is important in neuron-neuron and neuromuscular interactions (B.A. Cunningham et al., Science, 236:799-806 (1987)), while MAG is important in neuron-oligodendrocyte and oligodendrocyte-oligodendrocyte interactions during myelination (M (Poltorak et al., J. Cell Biol.. 105:1893-1899 (1987)). The cell surface expression of NCAM and MAG is developmentally regulated during nervous system formation and myelination, respectively, in analogy with the regulated induction of ICAM-1 during inflammation (T.A. Springer et al., Ann. Rev. Immunol., 5:223-252 (1987) ). ICAM-1, NACM (B.A. Cunningham et al., Science, 236:799-806(1987)) and MAG (J.L. Salzer et al.. J. Cell. Biol.. 104:957-965 (1987)) are similar in the overall structure as well as homologous, since each is an integral membrane glycoprotein constructed of 5 C2 domains that form the N-terminal extracellular region, although an additional non-Ig-like sequence is present in NCAM between the last C2 domain and the transmembrane field. ICAM-1 matches over its entire length including the transmembrane and cytoplasmic domains, with MAG with 21% identity; the same percentage identity is found when comparing the 5 fields of ICAM-1 and NCAM-1. A schematic comparison of the secondary structures of ICAM-1 and MAG is shown in figure 10. Field-by-field comparisons show that the level of homology between fields in ICAM-1 and NCAM molecules (respectively x + s.d. 21 + 2.8% and 18.6 ± 3.8%) is the same as the level of homology when comparing ICAM-1 fields and NCAM and MAG fields (20.4 + 3.7 and 21.9 + 2.7, respectively) . Although there is evidence for alternative splicing in the C-terminal regions of NCAM (B.A. Cunningham et al., Science, 236:799-806 (1987); D. Barthels et al., EMBO J.. 6:907-914 ( 1987)) and MAG (C. Lai et al., Proe. Nati. Acad. Sei. (USA). 84:4377-4341 (1987)), no evidence for this has been found in the sequencing of endothelial or HL-60 ICAM-1 clones or by studies of the ICAM-1 protein skeleton and precursor in many different cell types (M.L. Dustin et al., J. Immunol.. 137:245-254 (1986)).

ICAM-1 acts as a ligand for LFA-1 in lymphocyte interactions with a variety of different cell types. Lymphocytes bind to ICAM-1 incorporated into artificial membrane bilayers, and this requires LFA-1 on the lymphocyte, demonstrating direct LFA-1 interaction with ICAM-1 (S.D. Marlin et al., Cell. 51:813-819

(1987)). LFA-1 is a leukocyte integrin and has no immunoglobulin-like features. Leukocyte integrins comprise one integrin subfamily. The other two subfamilies mediate cell-matrix interactions and recognize the sequence RGD in their ligands which include fibronectin, vitronectin, collagen and fibrinogen (R.O. Hynes, Cell, 48.:549-554 (1987); E. Ruoslathi et al., Science , 238: 491-497 (1987)). The leukocyte integrins are expressed only on leukocytes, are involved in cell-cell interactions, and the only known ligands are ICAM-1 and iC3b, a fragment of the complement component C3 that shows no immunoglobulin-like features and is perceived by Mac- 1 (T.K. Kishimoto et al., I; Leukocvte Typing III. M. McMichael (ed.), Springer-Verlag, New York (1987) ; T.A. Springer et al. , Ann. Rev. Immunol.. 5_:223-252

(1987); D.C. Anderson et al., Ann. Fox. Med.. 38 = 175-194 (1987)). Based on sequence analysis, any peptides in the ICAM-1 sequence that are recognized by LFA-1 are shown in Table 9.

ICAM-1 is the first example of a member of the immunoglobulin supergene family binding to an integrin. Although both of these families play an important role in cell adhesion, interaction between them was not previously expected. In contrast, interactions within the immunoglobulin gene superfamily are quite common. It is very possible that further examples of interactions between the integrin and immunoglobulin families will be found. LFA-1 perceives a ligand distinct from ICAM-1 (T.A. Springer et al.. Ann. Rev. Immunol., 5:223-252 (1987)), and the leukocyte integrin Mac-1 perceives a ligand distinct from C3bi at neutrophil-neutrophil -adhesion (D.C. Anderson et al., Ann. Rev. Med.. 3_8:175-194 (1987)). Furthermore, purified MAG-containing vesicles are bound to neurites which are MAG, and thus MAG must be able to interact heterophilically with a special receptor (M. Poltorak et al.. J. Cell Biol.. 105:1893-1899 (1987)).

The role of NCAM in neutral-neutral and neutral-muscle cell interactions has been suggested to be due to homologous NCAM-NCAM interactions (B.A. Cunningham et al., Science, 236:799-806

(1987)). The important role of MAG in interactions between neighboring turn loops of Schwann cells enclosing axons during myelin sheath formation may be due to interaction with a special receptor, or homophilic MAG-MAG interaction. The homology with NCAM and the frequent occurrence of field-field interactions within the immunoglobulin supergene family raise the possibility that ICAM-1 could participate in homophilic interactions as well as heterophilic ICAM-1-LFA-1 interactions. However, binding of B-lymphoblast cells co-expressing similar densities of LFA-1 and ICAM-1 to ICAM-1 in artificial or cellular monolayers can be completely inhibited by pretreatment of the B-lymphoblast with LFA-1-MAb, while the adhesion is unaffected by fi-lymph oblast-f or treatment with ICAM-1-MAb. Pre-treatment of the monolayer with ICAM-1-MAb completely abrogates binding M.L. Dustin et al., J. Immunol., 137:245-254 (1986); S. D. Marlin et al., Cell, 51:813-819 (1987)). These findings demonstrate that if homophilic ICAM-1 interactions occur at all, they must be weaker than heterophilic interactions with LFA-1.

The possibility that the leukocyte integrins perceive ligands in a fundamentally different way is consistent with

the presence of a 180-residue sequence in their α-subunits which may be important in ligand binding and which is not present in the RGD-sensing integrins (A. Corbi et al. , ( EMBO J. , 6_:4023-4028

(1987)). Although Mac-1 has been suggested to perceive the RGD sequence present in iC3b 5086, there is no RGD sequence in ICAM-1 (Figure 8). This is consistent with the fact that the fibronectin peptide GRGDSP and the control peptide GRGESP do not inhibit ICAM-1-LFA-1 adhesion (S.D. Marlin et al., Cell, 51:813-819 (1987)). However, related sequences such as PRGGS and RGEKE are present in ICAM-1 in regions predicted to form a loop between fé chains a and b in field 2, and c and d in field 2 (Figure 9), and may thus be available for perception . It is of interest that the homologous MAG molecule contains an RGD sequence between fields 1 and 2 (M. Poltorak et al.. J. Cell Biol., 105:1893-1899 (1987); J.L. Salzer et al., J .Cell.Biol., 104:957-965 (1987)).

EXAMPLE 19

"Southern" and "Northern Blots"

Southern blots were performed using 5 µg of genomic DNA extracted from three cell lines: BL2, a Burkitt lymphoma cell line (a gift from Dr. Gilbert Lenoir); JY and Er-LCL, EBV-transformed B lymphoblastoid cell lines are.

The DNA molecules were digested with 5 times the manufacturer's recommended amount of BamHl and EcoRl endonucleases (New England Biolabs). After electrophoresis through a 0.8% agarose gel, the DNA molecules were transferred to a nylon membrane (Zeta Probe, BioRad). The filter was prehybridized and hybridized according to standard procedures using ICAM-1 cDNA from HL-60 labeled with α-(32P)d-XTP molecules by random labeling (Boehringer Mannheim). Northern blots were performed using 20 µg total RNA or 6 µg poly (A) + RNA. RNA was denatured and electrophoresed through a 1% agarose-formaldehyde gel and electro-transferred to Zeta-Probe. The filters were prehybridized and hybridized as described previously (D.E. Staunton et al., Embo J. , 6_:3695-370l (1987)) using the HL-60 cDNA probe of<32>P-labeled oligonucleotide- probes (described above).

Southern blots using the 3 kb cDNA probe and genomic DNA digested with BamHl and EcoRl showed single-dominant hybridization fragments of 20 and 8 kb, respectively, suggesting a single gene and suggesting that the majority of the coding information is present within 8 kb. In blots of three different cell lines, there is no evidence of restriction fragment polymorphism.

EXAMPLE 20

Expression of the ICAM-1 gene

An "expression vector" is a vector which (due to the presence of the appropriate transcripts and/or translational control sequences) can express a DNA (or cDNA) molecule that has been cloned in the vector, thereby forming a polypeptide or protein. Expression of the cloned sequences occurs when the expression vector is introduced into a suitable host cell. If a prokaryotic expression vector is used, the appropriate host cell will be any prokaryotic cell capable of expressing the cloned sequences. Similarly, if a eukaryotic expression vector is used, the appropriate host cell will be any eukaryotic cell capable of expressing the cloned sequences. It is significant that since eukaryotic DNA can contain introns, and since such sequences cannot be correctly processed in prokaryotic cells, it is preferred to use ~cDNA from a cell that can express ICAM-1 for generating a prokryotic genome expression vector library. Methods for the production of cDNA and for the production of a genome library are described by T. Maniatis et al., (Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor, NY (1982)).

The expression vector genomic library described above is used to create a bank of host cells (each of which contains an element of the library). The expression vector can be introduced into the host cell by any of many different means (ie, transformation, transfection, protoplast fusion, electroporation, etc.). The bank of expression vector-containing cells is clonally propagated and its elements analyzed individually (using an immunoassay) to determine whether they form a protein that can bind to anti-ICAM-1 antibody f .

The expression vectors for the cells that make a protein that can bind to anti-ICAM-1 antibody are then further analyzed to determine whether they express (and thus contain) the entire ICAM-1 gene, whether they express (and contain) only a fragment of the ICAM-1 gene, or whether they express (and contain) a gene whose product, although immunologically related to ICAM-1, is not ICAM-1. Although such analysis can be performed in any convenient manner, it is preferred to determine the nucleotide sequence of the DNA or cDNA fragment that had been cloned into the expression vector. Such nucleotide sequences are then examined to determine whether they may code for polypeptides with the same amino acid sequence as the tryptic cleavage fragments of ICAM-1 (Table 5).

An expression vector containing a DNA or cDNA molecule encoding the ICAM-1 gene can thus be understood by: (i) the ability to control the expression of a protein that can bind to anti-ICAM-1 antibody; and (ii) presence of a nucleotide sequence that can encode each of the tryptic fragments of ICAM-1. The cloned DNA molecule in such an expression vector can be removed from the expression vector and isolated in pure form.

EXAMPLE 21

Functional activities of purified ICAM-1

In cells, ICAM-1 normally functions as a surface protein associated with the cell membrane. The function of purified ICAM-1 was therefore investigated after the molecule had been restored in artificial lipid membranes (liposomes or vesicles) by dissolving the protein in detergent-solubilized lipids, followed by removal of the detergent by dialysis. ICAM-1, purified from JY cells and eluted in the detergent octylglucoside as described above, was reconstituted into vesicles, and the ICAM-1-containing vesicles were fused to glass coverslips or plastic culture wells to enable detection of bound cells to the protein.

Production of planar membranes and plastic-bound vesicles

Vesicles were prepared by the method of Gay et al., (J. Immunol., 136:2026 (1986)). Briefly, egg phosphatidylcholine and cholesterol were dissolved in chloroform and mixed in a molar ratio of 7:2. The lipid mixture was dried to a thin film while rotating under a stream of nitrogen gas, and then lyophilized for 1 hour while removing all traces of chloroform. The lipid film was then dissolved in 1% octyl glucoside/0.14 M NaCl/20 mM Tris (pH 7.2) to a final concentration of phosphatidylcholine of 0.1 mM. About. 10 µg of purified ICAM-1, or human glycophorin (Sigma Chemical Co., St. Louis, MO) as a control membrane glycoprotein, was added to each ml of dissolved lipids. The protein-lipid-detergent solution was then dialyzed at 4°C against 3 shifts of 200 volumes of 20 mM Tris/0.14 M NaCl, pH 7.2, and one shift of HBSS.

Planar membranes were prepared by the method of Brian et al.. Proe. Nati. Acad. Sci., 8JL:6159 (1984). Coverslips (diameter 11 mm) were boiled for 15 minutes in a 1:6 dilution of 7X detergent (Linbro), washed overnight in distilled water, immersed in 70% ethanol and air dried. An 80 µl drop of vesicle suspension containing either ICAM-1 or glucophorin was placed in the bottom of wells in a 24-well cluster plate, and the prepared coverslips were carefully floated on top. After 20-30 minutes of incubation at room temperature, the wells were filled with HBSS, and the coverslips were inverted so that the flat membrane would face up. The wells were then washed thoroughly with HBSS to remove unbound vesicles. The flat membrane surface was never exposed to air.

During experiments with planar membrane fused to glass surfaces, vesicles containing ICAM-1 were found to bind directly to the plastic surface of multiwell tissue culture plates, retaining functional activity evidenced by specific cell binding. Such vesicles are hereinafter referred to as "plastic-bound vesicles" (PBV), since the nature of the lipid vesicles bound to the plastic has not been determined. Plastic-bound vesicles were prepared by adding 30 µl vesicle suspension directly to the bottom of wells in 96-well tissue culture plates (Falcon), followed by incubation and washing as described for planar membranes.

Cell adhesion assays

Cell adhesion assays using planar membranes or plastic bound vesicles were both performed in essentially the same manner, except that the cell numbers and volume fractions for PBV assays were reduced to 1/5 of that used in planar membrane assays.

T lymphocytes from normal controls and a patient with leukocyte adhesion deficiency (LAD) whose cells do not express LFA-1 (D.C. Anderson et al., J. Infect. Dis., 152:668 (1985)) were prepared by culturing mononuclear cells from peripheral blood with 1/xg/ml Concanavalin-A (Con-A) in RPMI-1640 plus 20% FCS with 5 x 10<5> cells/ml for 3 days. The cells were then washed twice with RPMI and once with 5 mM methyl-alpha-D-mannopyranoside to remove residual lecithin from the cell surface. The cells were cultured in RPMI/20% FCS containing 1 ng/ml recombinant IL-2, and were used between 10 and 22 days after starting the culture.

For the detection of cell binding to planar membranes or PBV, Con-A blasts, the T-lymphoma SKW-3 and the EBV-transformed B lymphoblastoid cell lines JY (LFA-1 positive) and LFA-1 deficient lymphoblastoid cell line ( BBN) (from patient 1, T.A. Springer et al., J. Exper Med., 160:1901-1918 (1986)) labeled by incubating 1 x 10<7> cells in 1 ml RPMI-1640/10% FCS with 100 ttCi Na<51>CrO4 for 1 hour at 37°C, followed by four washes with RPMI-1640 to remove unincorporated label. In monoclonal-antibody-blocking experiments, cells or plastic-bound vesicles were pretreated with 20 µg/ml of purified antibody in RPMI-1640/10% FCS at 4°C for 30 minutes, followed by four washes to remove unbound antibody. In experiments regarding the effects of divalent cations on cell attachment, cells were washed once with Ca<2+->, Mg<2+->freeHBSS plus 10% dialyzed FCS, and CaCl and MgCl were added to the indicated concentrations. In all experiments, cells and planar membranes or PBV were pre-equilibrated at the appropriate temperature (4°C, 22°C or 37°C) in the appropriate assay buffer.

For measurement of cell-bound to purified ICAM-1,<51>Cr-labeled cells (5 x IO<5>EBV transformants in flat-membrane assays; 1 x IO<5>EBV transformants or SKW-3 cells , 2 x IO<5>Con-A vesicles in PBV assay) centrifuged for 2 min at 25 x g on planar membranes or PBV, followed by incubation at 4°C, 22°C or 37°C for 1 h. After the incubation, unbound cells were removed by eight cycles of filling and aspiration with buffer pre-equilibrated to the appropriate temperature. Bound cells were quantified by solubilizing the well contents with 0.1 N NaOH/1% Triton X-100 and counting in a gamma counter. Percent cell binding was determined by dividing cpm from bound cells by input cell-associated cpm. In flat-membrane assays, the entered cpm was corrected for the ratio of the surface area of coverslips compared to the surface area of the culture wells.

In these assays, EBV-transformed B-lymphoblastoid cells, SKW-3 T-lymphoma cells and Con-A-T lymphoblasts bound specifically to ICAM-1 in artificial membranes (Figures 11 and 12). The binding was specific since the cells bound very weakly to planar control membranes or vesicles containing equivalent amounts of another human cell surface glycoprotein, glycophorin. Furthermore, LFA-1-positive EBV transformants and Con-A blasts bound, whereas their LFA-1-negative counterparts did not bind to any significant extent, showing that binding depends on the presence of LFA-1 on the cells .

Both the specificity of the cell binding and the dependence on cellular LFA-1 were confirmed by monoclonal antibody blocking experiments (Figure 13). The binding of JY cells could be inhibited by 97% when the ICAM-1-containing PBVs were pre-treated with a monoclonal anti-ICAM-1 antibody RR1/1. Pre-treatment of the cells with the same antibody had little effect. Conversely, the anti-LFA-1 monoclonal antibody TS1/18 inhibited binding by 96%, but only when the cells, but not PBV, were pretreated. A control antibody TS2/9 that reacted with LFA-3 (another lymphocyte surface antigen) had no significant inhibitory effect when either cells or PBV were pretreated. This experiment shows that ICAM-1 itself in the artificial membranes, and not a minor contaminant, produces the observed cell adhesion, and that the adhesion is dependent on LFA-1 on the binding cell.

The binding of cells to ICAM-1 in artificial membranes also showed two other characteristics of the LFA-1-dependent adhesion system: temperature dependence and a requirement for divalent cations. As shown in Figure 14, Con-A blasts bound to ICAM-1 in PBV most efficiently at 37°C, partially at 22°C and very poorly at 4°C. As shown in Figure 15, the binding was completely dependent on the presence of divalent cations. At physiological concentrations, Mg<2+>alene was sufficient for maximum cell binding, while Ca<2+>alene gave very low levels of binding. However, Mg<2+> at 1/10 of the normal concentration, combined with Ca<2+>, was synergistic and gave maximum binding.

Briefly stated, the specificity of cell binding to purified ICAM-1 incorporated in artificial membranes, the specific inhibition with monoclonal antibodies and the temperature and divalent cation requirements show that ICAM-1 is a specific ligand for the LFA-1-dependent adhesion system.

EXAMPLE 22

Expression of ICAM-1 and HLA-DR by

allergic and toxic patch test reactions

Skin biopsies from five normal individuals were examined for expression of ICAM-1 and HLA-DR. It was found that although the endothelial cells of some blood vessels normally expressed ICAM-1, there was no ICAM-1 expression on keratinocytes from normal skin. No staining for HLA-DR on any keratinocyte from normal skin biopsies was observed. The expression kinetics of ICAM-1 and class II antigens were then examined for cells in biopsies from allergic and toxic skin lesions. Half of the six subjects examined were found to have keratinocytes expressing ICAM-1 4 hours after application of the hapten (Table 10). There was an increase in the percentage of individuals expressing ICAM-1 on their keratinocytes with exposure time to the hapten, as well as an increase in color intensity, showing more ICAM-1 expression per keratinocyte up to 48 hours. Indeed, at this time, a certain proportion of keratocytes in all biopsies stained positive for ICAM-1. At 72 hours (24 hours after the hapten was removed), seven of the eight subjects still had ICAM-1 expressed on their keratinocytes, while for one subject the expression of ICAM-1 decreased between 48 and 72 hours.

<a>Samples were considered positive if at least small clusters of keratinocytes were stained.

bAll patches were removed at this point.

Histologically, the staining pattern for ICAM-1 on keratinocytes from biopsies taken 4 hours after application of the hapten was usually in small clusters. After 48 hours, ICAM-1 was expressed on the surface of the majority of keratinocytes, with no difference being seen between the center and periphery of the lesion. The staining intensity decreased as the keratinocytes approached the stratum corneum. This was found for biopsies taken from both the center and the periphery of the lesions. Also at this time the patch test was positive (infiltration, erythema and vesicles). No difference in ICAM-1 expression was observed when different haptens were applied to susceptible individuals. In addition to keratinocytes, ICAM-1 was also expressed on some mononuclear cells and endothelial cells at the lesion site.

The expression of HLA-DR on keratinocytes in the allergic skin lesions was less frequent than the expression of ICAM-1. None of the subjects examined had lesions with keratinocytes that stained positive for HLA-DR up to 24 hours after application of the hapten. Indeed, only four biopsy specimens had keratinocytes expressing HLA-DR, and no biopsy had keratinocytes positive for HLA-DR and not ICAM-1 (Table 10).

In contrast to the allergic patch lesion, the toxic patch lesion induced with croton oil or sodium lauryl sulfate had keratinocytes showing no, if any, ICAM-1 on their surface at all time points examined (Table 11). Indeed, at 8 hours after patch application, which was the optimal time for ICAM-1 expression in the allergic patch test subjects, only one of the 14 toxic patch test subjects had keratinocytes expressing ICAM-1 in their lesions. Also in contrast to the allergic patch biopsies, there was no HLA-DR expressed on keratinocytes in toxic patch lesions.

These data show that ICAM-1 is expressed in immune-based inflammation, and not in toxic-based inflammation, and thus the expression of ICAM-1 can be used to distinguish between immune-based and toxic-based inflammation, such as acute renal failure in kidney transplant patients, where the it is difficult to determine whether the failure is due to rejection or renal toxicity of the immunosuppressive therapeutic agent. Renal biopsy and assessment of upregulation of ICAM-1 expression will enable differentiation of immune-based rejection and non-immune-based toxicity reaction.

EXAMPLE 23

Expression of ICAM-1 and HLA-DR i

benign skin diseases

Cells from skin biopsies of lesions from patients with different types of inflammatory skin diseases were examined for their expression of ICAM-1 and HLA-DR. A certain proportion of keratinocytes in biopsies from allergic-contact eczema, pemphigoid/pemphigus and lichen planus expressed ICAM-1. Lichen planus biopsies showed the most intense staining with a pattern similar to, or even stronger than, that seen in 48-hour allergic patch test biopsies (Table 12). Consistent with results seen with the allergic patch test, the most intense ICAM-1 staining was seen at sites of heavy mononuclear cell infiltration. Furthermore, 8 of the 11 examined lichen planus biopsies were positive with regard to HLA-DR expression on keratinocytes.

The expression of ICAM-1 on keratinocytes from skin biopsies from patients with exanthema and urticaria was less pronounced. Only four of the seven examined patients with these diseases had keratinocytes expressing ICAM-1 at the lesion site. HLA-DR expression was found in only one patient, and this was in association with ICAM-1.

Endothelial cells and a certain proportion of the infiltrate of the mononuclear cells from all the investigated benign inflammatory skin diseases expressed ICAM-1 to varying extents.

EXAMPLE 24

Expression of ICAM-1 on keratinocytes in psoriatic skin lesions

The expression of ICAM-1 in skin biopsies from 5 patients with psoriasis was examined before the start of, and periodically during, a course of PUVA treatment. Biopsies were taken from 5 patients with classic psoriasis confirmed by histology. Biopsies were taken sequentially before and during the indicated time for PUVA treatment. PUVA was given 3-4 times weekly. Biopsies were taken from the periphery of the psoriatic plaques in 5 patients, and in addition, biopsies were taken from clinically normal skin in 4 of the patients.

Fresh skin biopsy specimens were frozen and stored in liquid nitrogen. Cryostat sections of 6 were air-dried overnight at room temperature, fixed in acetone for 10 minutes and either stained immediately or wrapped in aluminum foil and stored at -80°C until staining.

The staining was carried out in the following way. Sections were incubated with monoclonal antibodies and stained by a 3-step immunoperoxidase method (H. Stein et. , Adv. Cancer Res. , 42.:67-147 (1984)), using a diaminobenzidine-H 2 O 2 substrate . Tonsils and lymph nodes were used as positive controls for anti-ICAM1 and HLA-DR staining. Tissues stained in the absence of primary antibody were negative controls.

The monoclonal antibodies against HLA-DR were supplied by Becton Dickinson (Mountainview, California). The monoclonal anti-ICAM-1 antibody was R6-5-D6. Peroxidase-conjugated rabbit anti-mouse Ig and peroxidase-conjugated swine anti-rabbit Ig were supplied by DAKAPATTS, Copenhagen, Denmark. Diaminobenzidine tetrahydrochloride was supplied from Sigma (St. Louis, Mo.).

The results of the investigation show that the endothelial cells in some blood vessels express ICAM-1 in both diseased and normal skin, but the color intensity and the number of blood vessels that expressed ICAM-1 were increased in the psoriatic skin lesions. Furthermore, the expression pattern of ICAM-1 in keratinocytes in untreated psoriatic skin lesions from the 5 patients varied from staining of only small cell clusters to staining of many keratinocytes. During PUVA treatment, ICAM-1 expression for two of the patients (patients 2 and 3) showed a marked reduction that preceded, or was concurrent with, clinical remission (table 13). Patients 1, 4 and 5 had decreases and increases of ICAM-1 expression during PUVA treatment that corresponded to clinical remissions and relapses, respectively. There was no ICAM-1 expression on keratinocytes from normal skin before or after PUVA treatment. This shows that PUVA does not induce ICAM-1 on keratinocytes from normal skin.

The observation that the density of the infiltrate of mononuclear cells corresponded to the amount of ICAM-1 expression on keratinocytes was remarkable. This applied to both a decreased number of mononuclear cells in lesions during PUVA treatment when ICAM-l expression also decreased, and an increased number of mononuclear cells during PUVA treatment when ICAM-l expression on keratinocytes was more predominant. Endothelial cells and skin mononuclear cells are also ICAM-1 positive. In clinically normal skin, ICAM-1 expression was restricted to endothelial cells without any labeling of keratinocytes.

The expression of HLA-DR on keratinocytes was variable. There was no HLA-DR positive biopsy that was not also ICAM-1 positive.

Briefly stated, these results show that before treatment ICAM-1 expression is pronounced on the keratinocytes and corresponds to a dense infiltrate of mononuclear cells. During PUVA treatment, a pronounced reduction in ICAM-1 staining is seen to parallel the clinical improvement. Histologically, the skin infiltrate also became smaller. When a clinical relapse was evident during treatment, the expression of ICAM-1 on the keratinocytes increased, as well as the density of the skin infiltrate. When a clinical remission was seen during treatment, there was a corresponding decrease in ICAM-1 staining on the keratinocytes as well as a decrease in the skin infiltrate. Thus, the expression of ICAM-1 on keratinocytes responded to the density of the infiltrate of mononuclear cells in the dermis. These data show that clinical response to PUVA treatment resulted in a pronounced decrease in ICAM-1 expression on keratinocytes in parallel with a more moderate decrease in the mononuclear cells. This shows that ICAM-1 expression on keratinocytes is responsible for starting and maintaining the skin infiltrate, and that PUVA treatment down-regulates ICAM-1, which in turn reduces the skin infiltrate and the inflammatory response. The data also show that there was variable HLA-DR expression on keratinocytes during PUVA treatment.

The expression of ICAM-1 on keratinocytes in psoriatic lesions correlates with the clinical severity of the lesion as well as the size of the skin infiltrate. Thus, ICAM-1 plays a central role in psoriasis, and inhibition of its expression and/or inhibition of its interaction with the CD 18 complex on mononuclear cells would be an effective treatment of the disease. Furthermore, control of ICAM-1 expression on keratinocytes will be an effective tool for diagnosis and prognosis, as well as

as an assessment of the progress of the treatment of psoriasis.

EXAMPLE 25

Expression of ICAM and HLA-DR by

malignant skin diseases

In contrast to lesions from benign skin conditions, the expression of ICAM-1 on keratinocytes from malignant skin lesions was much more variable (Table 14). Among the 23 cutaneous T-cell lymphoma studies, ICAM-1-positive keratinocytes were identified in only 14 cases. There was a tendency for keratinocytes from biopsies of mycosis fungoides lesions to lose their ICAM-1 expression with the progression of the disease to more advanced stages. However, ICAM-1 expression was observed on a varying proportion of the infiltrate of mononuclear cells from most of the cutaneous T-cell lymphoma lesions. Among the remaining lymphomas examined, 4 of 8 had keratinocytes expressing ICAM-1. Among the 29 examined patients with malignant skin diseases, 5 had keratinocytes expressing HLA-DR without expressing ICAM-1 (Table 14).

EXAMPLE 26

Effect of anti-ICAM-l antibodies on

the proliferation of human peripheral blood mononuclear cells

Human peripheral blood mononuclear cells are induced to proliferate by the presence and perception of antigens or mitogens. Certain molecules, such as the mitogen concanavalin A or the T-cell-binding antibody 0KT3, cause a non-specific proliferation of peripheral blood mononuclear cells.

Human peripheral blood mononuclear cells are heterogeneous in that they consist of subpopulations of cells that can perceive specific antigens. When a peripheral blood mononuclear cell capable of perceiving a particularly specific antigen encounters the antigen, proliferation of this subpopulation of mononuclear cells is effected. Tetanus toxoid and keyhole limpet hemocyanin are examples of antigens that are recognized by subpopulations of peripheral mononuclear cells, but not recognized by all peripheral mononuclear cells in sensitized individuals.

The ability of anti-ICAM-1 monoclonal antibody R6-5-D6 to inhibit proliferative responses of human peripheral blood mononuclear cells in systems known to require cell-cell adhesions was tested.

Mononuclear cells from peripheral blood were purified on Ficoll-Paque (Pharmacia) gradients according to the manufacturer's instructions. After collection of the interfacial layer, the cells were washed three times with RPMI 164 0 medium and cultured in flat-bottomed 96-well microtiter plates at a concentration of 10S cells/ml in RPMI 164 0 medium supplemented with 10% fetal bovine serum, 2 mM glutamine and gentamycin (50 µg/ml).

Antigen, either the T-cell mitogen, concanavalin A (0.25/xg/ml); the T-cell binding antibody, 0KT3 (0.001/xg/ml); keyhole limpet hemocyanin (10 g/ml) or tetanus toxoid (1:100 dilution from source) were added to cells cultured as described above either in the presence or absence of anti-ICAM antibody (R6-5-D6; final concentration 5 g/ml). Cells were cultured for 3.5 days (concanavalin A assay), 2.5 days (OKT3 assay), or 5.5 days ("keyhole limpet" hemocyanin and tetanus toxoid assay) before assays were terminated. 18 hours before the end of the analysis, 2.5/xCi<3>H-thymidine was added to the cultures. Cell proliferation was analyzed by measuring the incorporation of thymidine into DNA by the mononuclear cells from peripheral blood. Incorporated thymidine was collected and counted in a liquid scintillation counter (Merluzzi et al., J. Immunol., 139:166-168 (1987)). The results of these tests are shown in Figure 16 (concanavalin A test), Figure 17 (OKT3 test), Figure 18 (keyhole limpet hemocyanin test) and Figure 19 (tetanus toxoid test).

Anti-ICAM-1 antibody was found to inhibit proliferative responses to the non-specific T-cell mitogen,

ConA; the non-specific T cell-associated antigen, OKT-3;

and the specific antigens, "keyhole limpet" hemocyanin and

tetanus toxoid, in mono-nuclear cells. The inhibition by anti-ICAM-1 antibody was comparable to the inhibition by anti-LFA-1 antibody, suggesting that ICAM-1 is a functional ligand of LFA-1 and that antagonism of ICAM-1 will inhibit specific defense system responses.

EXAMPLE 27 Effect of anti-ICAM-1 antibody on

the mixed lymphocyte reaction

As discussed above, ICAM-1 is required for efficient cell-cell interactions during an immune response mediated by LFA-1-dependent cell adhesion. The induction of ICAM-1 during immune responses or inflammatory disease enables leukocyte interaction between and with endothelial cells.

When lymphocytes from two unrelated individuals are cultured in the presence of each other, blast transformation and cell proliferation of the lymphocytes is observed. This response, of one population of lymphocytes to the presence of another population of lymphocytes, is known as the mixed-lymphocyte reaction (MLR), and is analogous to the response of lymphocytes to the addition of mitogens ( Immunology The Science of Self-Nonself Discrimination, (J. Klein, John Wiley & Sons, NY (1982), pp. 453-458).

Experiments were performed to determine the effect of monoclonal anti-ICAM antibodies on the human MLR. These experiments were carried out as follows. Peripheral blood was collected from normal, healthy donors by venipuncture. The blood was collected in heparinized tubes diluted 1:1 at room temperature with Puck's G (GIBCO) balanced salt solution (BSS). The blood mixture (20 ml) was layered onto 15 ml of a Ficoll/hypaque density gradient (Pharmacia, density = 1.078, room temperature) and centrifuged at 1000 x g for 20 minutes. The interface layer was then collected and washed 3 times in Puck's G. Cells were counted on a hemacytometer and resuspended in RPMI 1640 culture medium (GIBCO) containing 0.5% gentamycin, 1 mM L-glutamine (GIBCO) and 5 % heat-inactivated (56°C, 30 minutes) human AB serum (Flow Laboratories) (hereafter referred to as RPMI culture medium).

Mouse anti-ICAM-1 (R6-5-D6) was used in these experiments. All monoclonal antibodies (prepared from ascites by Jackson ImmunoResearch Laboratories, Boston, MA) were used as purified IgG preparations.

Peripheral blood mononuclear cells (PBMC) were cultured in medium with 6.25 x 10 cells/ml in round-bottom microtitre plates of the Linbro type (no. 76-013-05). Stimulator cells from a separate donor were irradiated at 1000 R and cultured with the responder cells at the same concentration. The total volume per culture was 0.2 ml. Controls included responder cells alone as well as stimulator cells alone. The culture plates were incubated at 37°C in a 5% CO 2 atmosphere with humidified air for 5 days. The wells were pulsed with 0.5 tiCi tritiated thymidine (3HT) (New England Nuclear) for the last 18 hours of cultivation. In some cases, a two-way MLR was performed. The protocol was the same, except that the other donor's cells were not inactivated by irradiation.

The cells were harvested on glass fiber filters using an automatic multiple sample harvester (Skatron, Norway), and rinsing with water and methanol. Filters were oven dried and counted in Aquasol in a Beckman (LS-3801) liquid scintillation counter. Results are given as mean CPM + standard error for 6 individual cultures.

Table 15 shows that purified monoclonal anti-ICAM-1 antibodies inhibited MLR in a dose-dependent manner with significant suppression apparent at 20 ng/ml. Purified mouse IgG had little or no suppressive effect. Suppression of MLR by the monoclonal anti-ICAM-1 antibody occurs when the antibody is added within the first 24 hours of culture (Table 16).

Briefly, the ability of antibody against ICAM-1 to inhibit MLR shows that monoclonal ICAM-1 antibodies are therapeutically useful in acute graft rejection. Monoclonal ICAM-1 antibodies are also therapeutically useful in related immune-mediated disorders dependent on LFA-1/ICAM-1 regulated cell-cell interactions.

The experiments described here show that the addition of monoclonal antibodies against ICAM-1 inhibits the mixed-lymphocyte reaction (MLR) when added during the first 24 hours of the reaction. Furthermore, ICAM-1 is upregulated on human monocytes from peripheral blood during in vitro culture.

Furthermore, it was found that ICAM-1 is not expressed on resting human lymphocytes or monocytes from peripheral blood. ICAM-1 is upregulated on monocytes in cultured cells alone or cells co-cultured with unrelated donor cells in a mixed-lymphocyte reaction using standard flow cytometric assays. This upregulation of ICAM-1 on monocytes can be used as an indicator of inflammation, especially if ICAM-1 is expressed on healthy monocytes in individuals with acute or chronic inflammation.

The specificity of ICAM-1 towards activated monocytes, and the ability of anti-ICAM-1 antibody to inhibit MLR, suggests that monoclonal ICAM-1 antibodies may have diagnostic and therapeutic potential in acute graft rejection and related immune-mediated disorders requiring cell-cell- exchange effects.

EXAMPLE 2 8

Synergistic effects of combined administration

of anti-ICAM-1 and anti-LFA-1 antibodies

As shown in example 27, MLR is inhibited by anti-ICAM-1 antibody f. MLR can also be inhibited by anti-LFA-1 antibody. To determine whether combined administration of anti-ICAM-1 and anti-LFA-1 antibodies would have an additive or synergistic effect, an MLR assay (performed as described in Example 27) was performed in the presence of various concentrations of the two antibodies.

This MLR assay showed that the combination of anti-ICAM-1 and anti-LFA-1 at concentrations where neither antibody alone dramatically inhibits MLR is significantly more potent in inhibiting the MLR response (Table 17). This result shows that therapies that additionally include the administration of anti-ICAM-1 antibody (or fragments thereof) and anti-LFA-1 antibody (or fragments thereof) have the capacity to provide an enhanced anti-inflammatory therapy. Such improved therapy enables the administration of lower doses of antibody than would otherwise be therapeutically effective, and is important in conditions where

EXAMPLE 2 9

Additive effects of combined administration of sub-optimal doses of anti-ICAM-1 and

other immunosuppressive agents in MLR

As shown in Example 28, MLR is inhibited by combinations of anti-ICAM-1 and anti-LFA-1 antibodies. To determine whether the combined administration of anti-ICAM-1 and other immunosuppressive agents (such as dexamethasone, azathioprine, cyclosporine A or steroids (such as prednisone, etc.)) would also have increased effects, MLR- assays performed using sub-optimal concentrations (ie, concentrations that would be lower than the optimal concentration at which the agent alone would be administered to an individual) of R6-5-D6 in conjunction with other immunosuppressive agents according to the protocol of Example 27.

The data show that the inhibitory effects of R6-5-D6 are at least additive with the inhibitory effects of suboptimal doses of dexamethasone (Table 18), azathioprine (Table 19) and cyclosporine A (Table 20). This suggests that anti-ICAM-1 antibodies may be effective in reducing the required doses of known immunosuppressive agents, thereby reducing their toxic side effects. When using an anti-ICAM-1 antibody (or a fragment thereof) to achieve such immunosuppression, it is possible to combine the administration of the antibody (or fragment thereof) with either a further single immunosuppressive agent, or with a combination of more than one additional immunosuppressive agent.

EXAMPLE 3 0

Effect of anti-ICAM-1 antibody on suppression

of rejection of transplanted allogeneic organs

To demonstrate the effect of anti-ICAM-1 antibody in suppressing rejection of an allogeneic transplanted organ, cynomolgus monkeys were transplanted with allogeneic kidneys according to the procedure described by Cosimi et al., ( Transplant. Proe.. 13 :499-503 (1981)) with the modification that valium and ketamine were used as anesthetics.

Thus, the kidney transplant was performed essentially as follows. Heterotropic kidney allotransplantations were performed on 3-5 kg cynomolgus monkeys, essentially as described by Marquet (Marquet et al., Medical Primatolo<g>y, part II, Basel, Krager, p. 125 (1972)) after anesthesia with valium and ketamine. End-to-side anastomoses of donor renal vessels on a patch of aorta or vena cava were constructed using 7-0 Prolene suture. The donor ureter was spatulated and implanted in the bladder by the extravesical procedure (Y. Taguchi et al., in Dausset et al., (ed.), in: Advances in Transplantation, Baltimore, Williams & Williams, p. 339 ( 1968)). Renal function was assessed by serum creatinine determinations weekly or twice weekly. In addition, frequent allograft biopsies were taken for histopathological examination, and full autopsies were taken on all recipients who did not survive. For most recipients, bilateral nephrectomy was performed at the time of transplantation, and subsequent death due to urinary intoxication was considered the endpoint of allograft survival. For some recipients, unilateral native kidney removal and contralateral ureteral ligation were performed at the time of transplantation. When allograft rejection occurred, the ligation of the autologous ureter was removed, resulting in the restoration of normal renal function and the opportunity to continue immunological monitoring of the recipient animal.

Monoclonal antibody R6-5-D6 was administered daily for 12 days, starting two days before transplantation at a dose of 1-2 mg/kg/day. Serum levels of creatinine were examined periodically to monitor rejection. The effect of anti-ICAM-1 antibody on the immune system's rejection of the allogeneic kidneys is shown in Table 21.

The results show that R6-5-D6 was effective in prolonging the life of monkeys receiving allogeneic kidney transplants.

EXAMPLE 31

Effect of anti-ICAM-1 antibody on suppression

of acute rejection of transplanted organs

To demonstrate that anti-ICAM-1 antibody is effective in an acute model of graft rejection, R6-5-D6 was also tested in a therapeutic or acute kidney failure model. In this model, monkey kidneys were transplanted (using the protocol described in Example 30) and given perioperatively 15 mg/kg cyclosporin A (CyA) i.m. until stable renal function had been achieved. The dose of CyA was then reduced twice weekly in increments of 2.5 mg/kg until rejection occurred, as shown by a rise in blood creatinine levels. At this point, R6-5-D6 was administered for 10 days, and the length of survival was monitored. It is important to note that in this protocol the dose of CyA remains suboptimal, since it is not changed once the acute rejection episode takes place. In this model, history controls (N=5) without any antibody help survive 5-14 days from the start of the rejection episode. To date, six animals have been studied using R6-5-D6 in this protocol (Table 22). Two of these animals are still surviving (M12: 31 days and M5: 47 days after the administration of R6-5-D6). Two animals lived 38 and 55 days after initiation of R6-5-D6 therapy, and two animals died of causes other than acute rejection (one animal died of CyA toxicity, and the other died while receiving R6-5-D6 under anesthesia). This model is closer to the clinical situation where R6-5-D6 would be used from the beginning.

EXAMPLE 32

Genetic construction and expression of truncated derivatives of ICAM-1

In its native state, ICAM-1 is a cell membrane-bound protein that contains an extracellular region with 5 immunoglobulin-like domains, a transmembrane domain, and a cytoplasmic domain. It was desirable to construct functional derivatives of ICAM-1 that lacked the transmembrane field and/or the cytoplasmic field in that a soluble, secreted form of ICAM-1 could be formed. These functional derivatives were constructed by oligonucleotide-directed mutagenesis of the ICAM-1 gene, followed by expression in monkey cells after transfection with the mutant gene.

Mutations in the ICAM-1 gene resulting in amino acid substitutions and/or truncated derivatives were generated by the method of T. Kunkel, (Proe. Nati. Acad. Sei. (U.S.A), .82:488-492 (1985)) . The ICAM-1 cDNA prepared as described above was digested with restriction endonucleases Sal 1 and Kpn 1, and the resulting 1.8 kb DNA fragment was subcloned into the plasmid vector CDM8 (B. Seed et al., Proe. Nat. Acad . Sei. ( U. S. A.). 84:3365-3369 (1987)). A duf1 young strain of E. coli (BW313/P3) was then transformed with this construct, designated pCD1.8C. A single-chain uracil-containing template was extracted from the transformants by infection with helper phage R408 (Stratagene<R>). Mutant ICAM-1 cDNA molecules were then generated by priming second-strand synthesis with an oligonucleotide having mismatched bases, and subsequent transformation of a young<+-> host (MC1061/P3) with the resulting heteroduplex. isolated by screening for newly formed endonuclease restriction sites introduced by the mutant oligonucleotide The mutant ICAM-1 protein was expressed by transfection of Cos-7 cells with the mutant DNA in the eukaryotic expression vector CDM8 using standard DEAE-dextran procedures (R.F. Seiden et al., In: Current Protocols in Molecular Biology,

F.M. Ausubel et al.. ed.) pages 9.2.1-9.2.6 (1987)).

A truncated functional derivative of ICAM-1 lacking the transmembrane and cytoplasmic domains, but containing the extracellular region possessing all 5 immunoglobulin-like domains, was prepared. A 30 bp mutant oligonucleotide (CTC TCC CCC CGG TTC TAG ATT GTC ATC ATC) was used to transform the codons for the amino acids tyrosine (Y) and glutamic acid (E) in positions 452 and 453, respectively, to a phenylalanine (F) and et. translational stop codon (TAG). The mutant was isolated at its unique Xba 1 restriction site and was designated Y<452>E/F, TAG.

For expression of the mutant protein, COS cells were transfected with three mutant subclones (no. 2, no. 7 and no. 8). Three days after transfection with the three mutant subclones, culture supernatants and cell lysate were analyzed by immunoprecipitation with monoclonal anti-ICAM-1 antibody RR1/1 and SDS-PAGE. ICAM-1 was precipitated from the culture supernatants of cells transfected with mutant subclones #2 and #8, but not from detergent lysates of these cells. The molecular weight of ICAM-1 found in the culture supernatant was reduced by approx. 6 kd relative to the membrane form of ICAM-1, which is consistent with the size predicted for the mutant DNA. Thus, this functional derivative of ICAM-1 is secreted as a soluble protein. In contrast, ICAM-1 was not immunoprecipitated from control culture supernatants from cells transfected with native ICAM-1, demonstrating that the membrane form of ICAM-1 is not diffused from Cos cells. Furthermore, no ICAM-1 was immunoprecipitated from any culture supernatants or cell lysates from "mock" transfected negative control cells.

The truncated ICAM-1 secreted from transfected cells was purified by immunoaffinity chromatography with a specific ICAM-1 antibody (R6-5-D6) and examined for functional activity in a cell binding assay. After purification in the presence of the detergent octylglucoside, preparations containing native ICAM-1, or the truncated, secreted form, were diluted to a final concentration of 0.25% octylglucoside (a concentration below the critical micelle concentration for the detergent). These preparations of ICAM-1 were allowed to bind to the surfaces of 96-well plastic plates (Nunc), forming ICAM-1 bound to a solid phase. After washing out unbound material, approx. 75-80% and 83-88% of SKW-3 cells that had LFA-1 on their surface bound specifically to the native and the truncated forms of ICAM-1, respectively. These data demonstrate that the secreted, truncated, soluble functional ICAM-1 derivative retained both the immunological reactivity and the ability to mediate ICAM-1-dependent adhesion characteristic of native ICAM-1.

A functional derivative of ICAM-1 lacking only the cytoplasmic domain was prepared by similar methods.

A 25 bp oligonucleotide (TC AGC ACG TAC CTC TAG AAC CGC CA) was used to change the codon for amino acid 476 (Y) to a translational TAG stop codon. The mutant was designated Y<476>/TAG. Immunoprecipitation analysis and SDS-PAGE of Cos cells transfected with the mutant revealed a membrane-bound form of ICAM-1 with a molecular weight of approx. 3 kd less than native ICAM-1. Indirect immunofluorescence of the mutant-transfected Cos cells showed a

punctate staining pattern similar to native ICAM-1 expressed on LPS-stimulated human endothelial cells. Finally, cells transfected with the mutant DNA specifically bound to purified LFA-1 on plastic surfaces in a manner similar to Cos cells transfected with native ICAM-1 DNA (Table 23).

EXAMPLE 33

Mapping of functional ICAM-1 fields

Studies of ICAM-1 have shown that the molecule has seven fields.

Five of these fields are extracellular (field 5 is closest to the cell surface, field 1 is furthest from the cell surface), one field is a transmembrane field, bg one field is cytoplasmic (ie inside the cell). For determining which fields contribute to ICAM-1's ability to bind LFA-1, epitope mapping studies can be used. To carry out such studies, various deletion mutants are prepared and these are characterized with respect to their capacity to bind to LFA-1. Alternatively, the assays can be performed using anti-ICAM antibody known to interfere with the capacity of ICAM-1 to bind LFA-1. Examples of such suitable antibody include RR1/1 (R. Rothlein et al.,

J. Immunol.. 137:1270-1274 (1986)), R6 . 5 (T.A. Springer et al., U.S. Patent Application Serial No. 07/250,446), LB-2 (E.A. Clark et al., Leukocvte Ty<p>in<g>I, A. Bernard et al., eds.), Springer Verlag

pp. 339-346 (1984)), or CL203 D.E. Staunton et al., Cell. 56:849-853 (1989)).

Deletion mutants of ICAM-1 can be generated in any of a number of different ways. However, it is preferred to produce such mutants via site-directed mutagenesis, or in other recombinant ways (such as by constructing ICAM-1-expressing gene sequences where sequences coding for particular protein regions have been deleted). Methods which can be used for the production of such mutants are well known in the field. Using such methods, three ICAM-1 deletion mutants were produced. The first mutant lacks amino acid residues F185-P284 (ie deletion of field 3).

The second mutant lacks the amino acid residues P2 84-R451 (ie deletion of fields 4 and 5). The third mutant lacks the amino acid residues after Y476 (ie deletion of the cytoplasmic domain). The results of such investigations show that fields 1, 2 and 3 are mainly involved in interactions between ICAM-1 and anti-ICAM-1 antibody and LFA-1.

EXAMPLE 34

Effect of mutations in ICAM-1 on LFA binding

The ability of ICAM-1 to interact with, and bind to, LFA-1 is mediated by ICAM-1 amino acid residues present in field 1 of the ICAM-1 molecule (Figures 8, 9 and 10). However, such interactions are supported by contributions from amino acids present in fields 2 and 3 of ICAM-1. Among the preferred functional derivatives produced according to the present invention are thus soluble fragments of the ICAM-1 molecule containing fields 1, 2 and 3 of ICAM-1. More preferred are soluble fragments of the ICAM-1 molecule containing domains 1 and 2 of ICAM-1. Most preferred are soluble fragments of the ICAM-1 molecule containing field 1 of ICAM-1. Several amino acid residues in the first ICAM-1 field are involved in the interaction between ICAM-1 and LFA-1. Substitutions of these amino acids with other amino acids change the ability of ICAM-1 to bind LFA-1. These amino acid residues and substitutions are shown in Figure 25. Figure 25 shows the effects of such mutations on the ability of the resulting mutant ICAM-1 molecule to bind to LFA-1. In Figures 23-25, residues are described with reference to the one-letter code for amino acids, followed by the position of the residue in the ICAM-1 molecule. Thus, e.g. "E90" the glutamic acid residue at position 90 in ICAM-1. Likewise, "E90V" denotes the dipeptide consisting of the glutamic acid residue in position 90 and the valine residue in position 91. The substitution sequence is indicated to the right of the slash ("/"). The V4, R13, Q27, Q58 and D60S61 residues in ICAM-1 are involved in LFA-1 binding.

Substitution of these amino acids altered the capacity of ICAM-1 to bind to LFA-1. E.g. replacement of V4 with G results in the formation of a mutant ICAM-1 molecule that is less able to bind to LFA-1 (Figure 25). Replacement of the Rl3 residue in ICAM-1 with E leads to the formation of a mutant molecule with substantially less capacity to bind LFA-1 (Figure 25). Replacement of the Q58 residue in ICAM-1 with H leads to the formation of a mutant molecule with a largely normal capacity to bind LFA-1 (Figure 25). Replacement of the D60S residues in ICAM-1 with KL leads to the formation of a mutant molecule with substantially less capacity to bind LFA-1 (Figure 25).

Glycosylation sites in the second field are also included in LFA-1 binding (figure 23). Replacement of N103 with K, or A155N with SV, results in the formation of a mutant ICAM-1 molecule that is essentially unable to bind LFA-1. In contrast, replacement of the glycosylation site N175 with A did not appear to substantially alter the capacity of the mutant ICAM-1 to bind LFA-1.

Mutations in the third ICAM-1 domain did not alter ICAM-1-LFA-1 binding to any significant extent (Figure 24).

EXAMPLE 35

Multiple forms of ICAM-1 with increased

biological half-life, affinity and clearance

Chimeric molecules are constructed where fields 1 and 2 are attached to the hinge region of the immunoglobulin heavy chain. Preferred constructs attach the C-terminal end of ICAM-1 domain 2 to a segment of the immunoglobulin heavy chain gene just N-terminal to the hinge region, allowing for the segment flexibility afforded by the hinge region. ICAM-1 fields 1 and 2 will thus replace the Fab fragment in an antibody. Attachment to heavy chains in the IgG class and production of animal cells will result in the formation of a chimeric molecule. Formation of molecules containing heavy chains derived from IgA or IgM will result in the formation of molecules of higher multimerism containing from 2 to 12 ICAM-1 molecules. Co-expression of the J-chain gene in the animal cells that form the chimeric ICAM-1 heavy chain molecules will enable the correct assembly of IgA and IgM multimers, resulting mainly in IgA molecules containing 4-6 ICAM-1 molecules , and in the case of IgM, which contains approx. 10 ICAM-1 molecules. These chimeric molecules may have several advantages. First, Ig molecules are designed to last a long time in the circulation, and this can improve the biological half-life.

Furthermore, the multimeric nature of these engineered molecules will enable them to interact with greater intensity with rhinovirus as well as with cell surface LFA-1, depending on the therapeutic context, thus greatly reducing the amount of recombinant protein that needs to be administered to achieve an effective dose. IgA and IgM are highly glycosylated molecules that are usually present in secretions in mucous membranes such as the nose. Their highly hydrophilic nature helps keep bacteria and viruses to which they bind out of the mucosa, preventing attachment to cells and preventing crossing of the epithelial cell membrane barrier. Thus, they may have increased therapeutic effect. IgM and especially IgA are stable in mucosal environments, and they will be able to increase the stability of the constructed ICAM-1 molecules. If such a functional ICAM-1 derivative is administered in the blood stream, it will also be able to increase the biological half-life. IgA does not fix complement, and would thus be ideal for applications where this would be harmful. If chimeric IgG-H chain molecules are desired, it would be possible to mutate areas involved in attachment to complement as well as in interactions with Fc receptors.

EXAMPLE 36

Generation of ICAM-1 mutants

Oligonucleotide-directed mutaqenesis

The coding region for an ICAM-1 cDNA in a 1.8 kb SalI-KpnI fragment was subcloned into the expression vector CDMB (B. Seed et al., Proe. Nati. Acad. Sei. ( U. S. A.), 84=3365-3369 (1987)). Based on the procedure according to T. Kunkel, (Proe. Nati. Acad." Sei.

( U. S. A. ) , ,82:488-492 (1985)) and modifications according to D. Staunton et al., (D.E. Staunton et al., Cell, 52:925-933 (1988)), this construct (pCD1.8 ) used to generate a single-chain uracil-containing template for use in oligonucleotide-directed mutagenesis.

Briefly, E. coli strain XS127 was transformed with pCD1.8. Single colonies were grown in 1 ml Luria broth (LB) medium (Difco) with 13 µg/ml ampicillin and 8 µg/ml tetracycline until near saturation point was reached. 100 µl of the culture was infected with R408 helper phage (Strategene) at a multiplicity of infection (MOI) of 10, and 10 ml of LB medium with ampicillin and tetracycline was added for a 16 hour culture at 37°C. After centrifugation at 10,000 rpm for 1 minute, and 0.22 µm filtration of the supernatant, the phage suspension was used to infect E. coli BW313/P3, which was then plated on LG agar (Difco) plates supplemented with ampicillin and tetracycline. Colonies were picked, grown in 1 ml of LB medium with ampicillin and tetracycline until near saturation was reached, and infected with helper phage at an MOI of 10. The culture volume was then increased to 250 ml, and the cells were grown overnight. Single-stranded DNA was isolated by standard phage extraction.

Mutant oligonucleotides were phosphorylated and used with the pCD1.8 template in a second-strand synthesis reaction (D.E. Staunton et al., Cell, 52:925-933 (1988)).

Transfection

COS cells were seeded in 10 cm tissue culture plates such that they would be 50% confluent after 16-24 hours. The COS cells were then washed once with TBS and incubated for 4 hours with 4 ml RPMI containing 10% Nu-sera (Collaborative) 5 µg/ml chloroquine, 3 µg mutant plasmid and 200 µg/ml DEAE-dextran sulf that. The cells were then washed with 10% DMSO/PBS, followed by PBS, and cultured for 16 hours in culture medium. The culture medium was replaced with fresh medium, and 48 hours after transfection (OS cells were suspended by trypsin/EDTA (Gibco) treatment and split into 2.10 cm plates as well as 24-well tissue culture plates for HRV binding. After 72 h, the cells were harvested from 10 cm plates with 5 mM EDTA/HBSS and processed for adhesion to LFA-1-coated plastic and immunofluorescence.

LFA-1 and HRV binding

LFA-1 was purified from SKW-3 lysates by immunoaffinity chromatography on TS2/4 LFA-1 mAb Sepharose and eluted at pH 11.5 in the presence of 2 mM MgCl 2 and 1% octylglucoside. LFA-1 (10 µg per 200 µl per 6 cm plate) was bound to bacteriological Petri dishes by diluting octylglucoside to 0.1% in PBS (phosphate buffered saline) with 2 mM MgCl2 and incubating overnight at 4° C. The plates were blocked with 1% BSA (bovine serum albumin) and stored in PBS containing 2 mM MgCl 2 , 0.2% BSA, 0.025% azide and 50 µg/ml gentamycin.

<51>Cr-labeled COS cells in PBS containing 5% FCS (fetal calf serum), 2 mM MgCl2 and 0.025% azide (buffer) were incubated with or without 5 µg/ml RRl/1 and R6.5 in LFA-l -coated microtiter plates at 25°C for 1 hour. Non-adherent cells were removed by 3 washes with buffer. Adherent cells were eluted by addition of EDTA to 10 mM and the γ counted.

Results

Anti-ICAM-1 antibodies such as RR1/1, R6.5, LB-2 and CL203 have been identified. If these antibodies can inhibit ICAM-1 function, they must be able to bind to a special place in the ICAM-1 molecule that is also important for ICAM-1 function. By producing the above-described deletion mutants of ICAM-1, and determining the extent to which the anti-ICAM-1 antibodies can bind to the deletion site, it is thus possible to determine whether the deleted fields are important for the function.

ICAM-1 is a membrane protein, where the extracellular field is predicted to consist of 5 Ig-like C fields. For the identification of fields involved in the binding of LFA-1, field 3 and fields 4 and 5 (carboxyl-terminal) were deleted by oligonucleotide-directed mutagenesis and examined functionally after expression in COS cells. Additionally, the entire cytoplasmic domain was deleted to ascertain its potential impact on ICAM-1 interactions. As expected, deletion of the cytoplasmic field, Y476/<*>, showed no loss of RRl/1, R6.5, LB-2, and CL203 reactivity, while deletion of field 3, F185-R451, resulted in a decrease and loss of CL 2 O 3 reactivity (Figure 20). Thus, the CL203 epitope appears to be found in field 4, while RR1/1, R6.5 and LB-2 appear to be found in the two amino-terminal fields.

All 3 deletion mutants show wild-type levels of attachment to LFA-1 (Figure 21). Amino acid substitutions in predicted S-turns in fields 1, 2 and 3 were also generated and functionally tested after expression in COS cells. The R6.5 epitope was thus localized to the sequence E111GGA in field 2, and may also include E3 9 in field 1, while RR1/1 and LB-2 both depend on R13 in field 1 (figure 22). Moreover, RR1/1 binding is reduced by mutations in the sequence D71GQS. Mutations that delete N-linked glycosylation sites at N103 and N165 result in decreased RR1/1, R6.5 and LB-2, LFA-1 HRV binding. These mutations appear to cause processing to form ICAM-1 dimers.

Other mutations in fields 2 or 3 did not result in altered LFA-1 adhesion (Figures 23 and 24). The amino acids in field 1, R13 and D6 0, are both involved in the binding of LFA-1 (figure 25).

Thus, it appears that LFA-1 and HRV binding is a function of the amino-terminal Ig-like domain of ICAM-1. Figure 26 shows an arrangement of amino-terminal ICAM fields.

Table 24 is data where different mutations are shown in relation to their effect on rhinoviral, CL203 and LFA-1 binding to ICAM-1. The mutant molecules produced according to the present invention can be divided into four groups and have advantageous properties in the treatment of diseases and other disorders.

The first group consists of mutants that have an improved binding capacity to LFA-1 (and to HRV as well), e.g. Ql/E (175%,149%), S3/T (149%, 196%), or Y52/FA (138%, 125%) of region 1 (Dl), or K128/R (109%, 137% ) of area 2(D2), or N24ODS/KNA (14 7%) of area 3 (D3). From this it can be deduced that there is an improved binding of these mutants to LFA-1 (and HRV) which constitutes a particular advantage of these mutants compared to wild-type ICAM-1 protein.

The second group of mutants shows significantly reduced binding capacity to LFA-1, but an improved binding to human rhinovirus, e.g. K8/E (84%, 132%), Y52/F (72%, 174%) or Y66/T (135%, 204%) of Dl, G101/KN (97%, 140%), E127/R ( 82%, 131%) of D2 and A189/SI (91%, 168%) of D3. In some cases, the improved binding capacity of these mutants is approx. 200%, or even over 200% of the wild type. This group of mutants is also clearly advantageous over wild-type ICAM-1 protein. It should be noted that the disclosure also teaches other uses of ICAM-1 independent of LFA-1 binding. E.g. can ICAM-1 also bind to human rhinovirus (HRV), see e.g. example 36 which describes that LFA-1 and HRV can bind to an overlapping region on ICAM-1 and that the region on the cell adhesion molecule chosen for binding by the virus is similar to the region adapted for binding to the cell adhesion receptor. Accordingly, such mutants that have reduced binding to LFA-1 and enhanced or unenhanced binding capacity may have advantages over wild-type ICAM-1 in that they have a full capacity to bind HRV and a reduced capacity to bind LFA-1, since the capacity to bind LFA-1 may be suboptimal in certain situations, such as when using soluble ICAM-1 derivatives to treat infections/inflammations caused by rhinoviruses.

The smaller third group of mutants has essentially the same binding capacity for LFA-1 (and for HRV) as wild-type ICAM-1, e.g. D213GL/HGV (90%, 99%) of D3. Because of this similarity in binding capacity, there is at first glance no particular advantage of these mutants compared to the wild-type protein. However, such mutants enable one skilled in the art to perform binding studies for scientific and diagnostic purposes, or to vary doses for treatment.

The fourth group of mutants have reduced, attenuated or abolished capacity to bind both LFA-1 and HRV. Nevertheless, these mutants may still be valuable, e.g. in the production of anti-ICAM-1 antibodies and in binding studies.

While the invention has been described in connection with specific embodiments thereof, it will be understood that it may be further modified, and this patent application is intended to cover all variations, applications or adaptations of the invention which generally follow the principles of the invention, and includes such deviations from the present description which is within known or common practice within the area to which the invention relates and which may be suitable for the basic features set out in the foregoing, which follow within the scope of the associated claims.

Claims (7)

1. Process for the preparation of a therapeutic, active water-soluble, functional derivative of ICAM-1 which is a fragment of ICAM-1, which lacks the cytoplasmic domain and the transmembrane domain of ICAM-1, and which contains domain 1 or domains 1 and 2, or domains 1, 2 and 3 selected from the group containing at least one of the mutations referred to in figures 23, 24 and 25, possibly multimeric and chimeric forms of these with corresponding biological activity, characterized in that it comprises the cultivation of a host microorganism that has been transformed with a recombinant vector encoding a water-soluble, functional derivative of ICAM-1 as defined above in combination with transcriptional and translational regulatory sequences, as well as a selection signal coding sequence, under conditions whereby the derivative is expressed and secreted in the culture medium, and then isolation of the derivative. 2. Method according to claim 1, characterized in that the multimeric form comprises up to 10 ICAM-1 derivatives. 3. Method according to claim 1, characterized in that the chimeric form of the ICAM-1 derivative is linked to an immunoglobulin or functional part thereof. 4. Method according to any one of claims 1-3, characterized in that the multimeric form comprises a chimeric molecule where regions 1 and 2 of the ICAM-1 derivative are linked to a hinge region of the heavy chain of the immunoglobulin. 5. Method according to any one of claims 1-4, characterized in that the multimeric form comprises a chimeric molecule in which the C-terminal end of ICAM-1 in region 2 is linked to a segment of the heavy chain of the immunoglobulin, immediately N-terminally to the hinge region. 6. Method according to any one of claims 3-5, characterized in that the heavy chain of the immunoglobulin is selected from the group IgA, IgG or IgM. 7. Recombinant DNA molecule capable of expressing a water-soluble functional derivative of ICAM-1, which lacks the cytoplasmic domain and the transmembrane domain of ICAM-1, and which contains field 1, or fields 1 and 2, or fields 1, 2 and 3 selected, from the group containing at least one of the mutations referred to in Figures 23, 24 and 25 and optionally chimeric forms thereof, characterized in that the molecule contains a sequence that codes for said protein in combination with a selection marker and transcription and translation regulatory sequences that enable expression in a suitable host cell.