AU638210B2 - Protein micelles - Google Patents

Description

AsOPI DATE 29/11/89 AOJP DATE 04/01/90

PCT

APPLN. ID 37310 89 PCT NUMBER PCT/US89/01858 INTERNATIONAL APPLICATION PUBLISHED UNDER THE PATENT COOPERATION TREATY (PCT) (51) International Patent Classification 5 -lnternational Publication Number: WO 89/10938 C07K 9/00, 15/14, A61K 37/02 A l (43) International Publication Date: 16 November 1989 (16.11.89) (21) International Application Number: PCT/US89/01858 (81) Designated States: AT (European patent), AU, BE (European patent), CH (European patent), DE (European pa- (22) International Filing Date: 2 May 1989 (02.05.89) tent), FR (European patent), GB (European patent), IT (European patent), JP, LU (European patent), NL (European patent), SE (European patent).

Priority data: 189,971 4 May 1988 (04.05.88) US 238,926 31 August 1988 (31.08.88) US Published With international search report.

(71)Applmant: DANA-FARBER CANCER INSTITUTE, INC. [US/US]; 44 Binney Street, Boston, MA 02115

(US).

(72) Inventors: DUSTIN, Michael, L. 231 Park Drive,,23, Bos- 6 3 8 2 1 0 ton, MA 02215 SPRINGER, Timothy, 28 Manadnock Road, Chestnut Hill, MA 02167 (US).

(74) Agent: CLARK, Paul, Fish Richardson, One Financial Center, Suite 2500, Boston, MA 02111-2658 (US).

(54)Title: PROTEIN MICELLES Abstract A micelle of an adhesion protein which naturally includes a phosphatidylinositol lipid anchor, the micelle being capable of binding multivalently to a plurality of target molecules on a cell surface.

WO 89/10938 PC/US89/0O1858 1 PROTEIN MICELLES BACKGROUND OF THE INVENTION This invention relates to adhesion proteins (as defined below), and in particular to lymphocyte function associated antigen-3 (LFA-3), a glycoprotein which is widely distributed on the surfaces of almost all human cells, including monocytes, granulocytes, cytolytic T lymphocytes, B lymphoblastoid cells, platelets, and fibroblasts.

LFA-3 is a ligand for cluster of differentiation 2 (CD2), another glycoprotein, found on the surfaces of T lymphocytes. These two glycoproteins interact to mediate T lymphocyte adhesion to target cells. Similarly, binding of thymocyte to thymic epithelial cells requires CD2 on the thymocyte and LFA-3 on the thymic epithelial cell. LFA-3 in purified form inhibits intercellular adhesion between T lymphocytes and erythrocytes and mediates aggregation of T lymphocytes.

SUMMARY OF THE INVENTION The invention provides a micelle of an adhesion protein, LFA-3, which naturally includes a phosphatidylinositol lipid anchor; the micelle WOFS9/10938 PCT/US89/01858 2 (which preferably contains less than about ten molecules of the protein) is capable of binding multivalently to a plurality of target molecules on a cell surface. Thus the invention provides multimeric, purified LFA-3 which has an avidity for CD2-bearing cells such that the Kd of the LFA-3 for the cells is lower: than the Kd of monomeric LFA-3, which has been determined to be about 400 nM. More preferably, the Kd is <50 nM, and even more preferably <20 nM. An adhesion protein is defined herein as any protein mediating the contact and union of two or more human cells, and preferably, is a protein present on the surface of a cell. A target molecule is defined herein as a molecule to which an adhesion protein binds selectively, binds to a degree greater than the degree to which it binds to any other molecule, preferably, this binding is exclusive.

The micelles of the invention can be used to inhibit the adhesion of a first cell to a second cell where the first cell bears on its surface a PI anchor-bearing protein capable of binding to a molecule present in multiple form on the second cell. The method includes contacting the second cell with a micelle, preferably of less than about 10 such protein molecules (or micelle-forming, binding fragments thereof), which is capable of binding multivalently to the multiple molecules on the second cell.

The adhesion protein micelles of the invention can further be used for the treatment of medical conditions characterized by the presence of an excess of activated T-cells by administering LFA-3 micelles in a physiologically compatible buffer to a patient to achieve a bloodstream concentration of LFA-3 effective to inhibit the binding of the activated T-cells to other cells in the patient; the resultant bloodstream concentration of LFA-3 is preferably 0.4 to 40 nM of WO 89/10938 PCF/US89/01 858 3 LFA-3. Disease states characterized by the presence of an excess of activated T-cells include multiple sclerosis, sarcoidosis, juvenile type diabetes mellitus, systemic lupus erythmatosis, thyroiditis, rheumatoid arthritis, ankyloses spondylitis, primary biliary cirrhosis, autoimmune hemolytic anemia, immune thrombocytopenia purpura, myesthenia gravis, allograft rejection and graft-versus-host disease.

Another aspect of the invention features a 101 method of- increasing the avidity and decreasing the dissociation constant of any adhesion protein which naturally includes a PI anchor. The method includes solubilizing the adhesion protein in a detergent solution and diluting the detergent below its critical micelle concentration in such a way as to form adhesion protein micelles. A detergent is defined herein as any small amphipathic molecule having hydrophobic and hydrophilic regions that are preferably spatially separated on the molecule, one end is charged or polar and the other end is apolar. The critical misclle concentration (CMC) is the concentration of a specific detergent in solution at which micelles just begin to form in the solution. This CMC is different for different detergents.

A further aspect of the invention features a method of effecting activation or proliferation of peripheral blood mononuclear cells or T lymphocytes by contacting such cells with LFA-3 micelles together with an antibody to CD2.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The drawings will first be briefly described.

Drawings Figure 1 is a graph depicting the results of gel filtration of two forms of LFA-3; WO 89/10938 PCr/US89/01 858 4 Figure 2 is a schematic of the transmembrane (TM) form of LFA-3; Figure 3 is a schematic of an adhesion protein micelle; Figure 4 is the DNA sequence and deduced amino acid sequence of human TM LFA-3 cDNA.

Figure 5 is a graph of the dissociation time-course for LFA-3 micelles from Jurkat cells; Eigures 6a and 6b are graphs depicting 10.7 equi:i:riunr binding of LFA-3 micelles to Jurkat and resting T cells and the same data analyzed by the Scatchard method, respectively; Figures 7a and 7b are graphs showing the time-course.and dose-response for LFA-3 and CD2 monoclonal antibody (MAb)-induced proliferation; Figure 8 is an emission spectra graph which +2 shows the cytoplasmic mobilization of Ca+ 2 over time in Jurkat cells induced by LFA-3 and CD2 MAb; Figure 9 is an emission spectra graph which shows the cytoplasmic mobilization of Ca 2 over time in resting T cells induced by LFA-3 and CD2 MAb; Figures 10a and 10b are an emission spectra graph and a set of histograms, respectively, that show the concentration dependence of Ca+ 2 influx increase in Jurkat. cells on the LFA-3 occupation of CD2.

Purification of LFA-3 The following cell lines were used in research ralating to the present invention. Peripheral blood mononuclear cells (PBMC) were isolated by dextran sedimentation and Ficoll-Hypaque (Sigma, St. Louis, MO) density gradient centrifugation. Peripheral blood T lymphocytes (PBL-T) were enriched by nylon wool (Polyscience, Warrington, PA) filtration and plastic adherence. The Jurkat cell line was obtained from Dr.

M.K. Ho (Dupont, Boston, MA). Cells were maintained in WO 89/10938 PC1/US89/01858 RPMI-1640, 10% fetal bovine serum (Gibco, Grand Island, NY), 5mM L-gutamine, 50 pg/mi gentamycin (Complete media) JY B lymn.%ioblastoid cells and Jurkat cells for LFA-3 and CD2 puridication, respectively, were grown by the MIT Cell Culture Center (Cambridge, MA).

LFA-3 was purified from Triton X-100 lysates of human erythroclytes (Dustin, M. L. et al. J. Exp Med. 165:677 or JY B lymphoblastoid cells (Waliner, et al., J. Exp. Med., 166:923 (1987)), or from supernatants of phosphatidylinositol-specific phospholipase C (PIPLC) treated JY B lymphoblastoid cells by immunoaf finity chromatography on TS2/9 MAb-Sepharose CL-4B (Pharmacia, Piscataway, 0) In the latter case, 50 g of viable JY cells were washed with Hank's buffered saline solution (HBSS) and treated with Bacillus thuringiensis PIPLC (froal Dz. Martin Low, Columbia University, NY, NY) in a total volume of 100 ml for 1 hr. at 37 0 C. The concentration of enzyme used could hydrolyze 300 nmol/rnin,/ml of 3 H-phosphatidylinositol at pH 7 in the presence of 0.1% deoxycholate detergent (Low, Methods Enzymol., 71:741 (1981)). The cells were pelleted at 1000 x g and cell debris was pelleted at 100,000 x g for 1 hr, Lysates o~r supernatants containing LFA-3 were passed over the affi-nity column and the column washed (Dustin, MLet al. supra). LFA-3 from erythrocytes was eluted from the irmmuaoaffinity column at pH 3 in the presence of 1% octylglucoside (OG) detergent. LFA-3 from PIPLC supernatants was eluted at pH Z in the absence of detergent, fractions containing LFA-3 were pooled (4-6 ml), passed over a 1 ml phenyl-Sepharose column (Pharmacia): equiilibrated with TSA. Some prepnarations were passed over a 1 ml protein A-Sepharose CL-4B colimn aft4, dilutiing 1:1 with MAPS binding buffer WO 89/10938 PCf~/US89/01858 6 (with 1% OG for LFA-3 from erythrocytes) (Biorad, Richmond, CA).

The LFA-3 purified as above has two forms: the mLFA-3 form, isolated from erythrocytes or lymphoblastoid cells, and the sLFA-3 form, isolated from lymphoblastoid cells. mLFA-3 has an intact phosphatidylinositol (PI) membrane anchor and is also referred to as the lipid-linked form. sLFA-3 has had the PI membrane anchor cleaved by PIPLC. Of these two, only'mLFA-3 can form micelles.

There is a third form of LFA-3, transmembrane LFA-3, or TM LFA-3, which does not have a lipid membrane anchor but remains attached to the cell membrane by a transmembrane domain that is a part of the complete LFA-3 protein. TM LFA-3, like mLFA-3, can form micelles and will be described in more detail below.

First, sLFA-3 will be described.

Our recent finding that about half of the LFA-3 on the surface of B lymphoblastoid cells is attached to the membrane by a PI glycan moiety which is cleavable by PIPLC (Dustin, et al., Nature, 329:846 (1987)) allowed us to obtain a non-aggregated, non micelle-forming, form of LFA-3 for binding studies.

LFA-3 released from the surface of live JY cells by EIPLC was purified by immunoaffinity chromatography, hydrophobic interaction chromatography on phenyl Sepharose, and protein A affinity chromatography. The middle step was added to remove traces of membranous material or LFA-3 micelles which might be released during enzyme treatment. The yield from 50g of cells was about O10og. When this material was treated with N-glycanase, a single band of 25.5 kD was obtained which corresponds to the lipid-linked forms from JY cells and erythrocytes.

WO 89/10938 PCT/US89/0 i858 7 The structure of mLFA-3 in the form of micelles, as used in subsequent activation experiments, was studied by gel filtration. The results are shown graphically in Figure 1. The size of mLFA-3 in the presence and absence of 1% OG detergent was examined by HPLC gel filtration. In the absence of detergent (solid line), mLFA-3 eluted slightly before thyroglobulin (600,000 kD) but after blue dextran on TSK GS 4000 or Zorbax GF 250 (Dupont) columns and had a diffusion coefficient similar to a globular protein of 700 kD.

The dotted line shows the elution profile in the presence of 1% OG. In the presence of 1% OG all of the LFA-3 coeluted with IgG (150 kD). LFA-3 migrates at 55-70 kD in SDS-PAGE (Dustin, et al., supra) The molecular weight of the mLFA-3 with and without detergent cannot be estimated without sedimentation data; however, these data show that mLFA-3 forms protein micelles of approximately 5 monomers or dimers in the absence of detergent. mLFA-3 does not form disulfide linked dimers based on SDS-PAGE, but the possibility that it forms non-covalent dimers cannot be ruled out.

Furthermore, sLFA-3 in the absence of detergent exists as a soluble molecule with a diffusion coefficient consistent with a monomer or dimer.

The second form of micelle-forming LFA-3 is the transmembrane form, TM LFA-3. The mRNA encoding this form differs from that of the mLFA-3 form, but only in the transmembrane and cytoplasmic domains of the molecule. The CD2 binding domain (the extracellular portion) of both forms of LFIA-3 is identical.

Although TM LFA-3 has no lipid anchor, the cytoplasmic domain includes a hydrophobic amino acid sequence which remains attached to the purified protein. Micelle formation occurs by protein-protein interactions rather than lipid- "~id interactions but WO 89/10938 PCI/US9/01858

-I

8 the result, micelle formation, is the same, TM LFA-3 and mLFA-3 micelles have comparable stability and target binding avidity.

It has been found that the TM form of LFA-3 occurs naturally in most nucleated cells, which often simultaneously also produce the lipid-linked form.

Formation of Micelles Micelles of mLFA-3 and TM LFA-3 were prepared hy-thriee cycles of ultrafiltration using a Centricon apparatus. (Amicon, Danvers, adding 2 ml of phosphate buffered saline (PBS) and reducing the volume to 50 pl with each cycle. Because the final ultrafiltrate of the last step contained the same low molecular weight components as the final retentate but not any LFA-3, this ultrafiltrate was used as a control for effects of buffer components in subsequent experiments.

Formation of LFA-3 micelles occurs spontaneously when detergent is removed or diluted below its critical micelle concentration. Detergent removal can be achieved by several techniques. The major requirement is that detergent is removed or diluted such that protein aggregation is the most favorable event rather than protein adsorbtion to other materials--i.e., test tube surfaces, etc. One technique which is successful with mLFA-3 is ultrafiltration in which an mLFA-3 solution in 1% octylglucoside detergent is concentrated from 2ml to 501 and diluted with detergent-free solution. This results in dilution of octylglucoside to below its critical micelle concentration (25mM). Other techniques for detergent removal/dilution include dialysis and sedimentation through detergent free sucrose gradients. The protein micelles generated by any technique will be functionally equivalent since the ?tructure of the protein micelles WO 89/10938 PCr/US89/01 858 9 is determined by the properties of the hydrophobic and hydrophilic portions of the protein and not on the means of detergent removal.

Whenever proteins with hydrophobic components, such as LFA-3, are purified, there is a tendency for these proteins to adhere to the glass or plastic containers used in the purification process. The rapid removal of the detergent counteracts this tendency, and promotes protein-protein rather than protein-plastic a protein-glass binding, and can result in approximately 100 percent recovery of the purified LFA-3 in micelle form. The micelles can be maintained in the PBS for at least two months without loss of activity.

Theoretically, if the PBS is completely free of proteases, the storage life of micelles should be much longer.

Micelles created according to the invention have their hydrophobic lipid or hydrophobic proteinaceous "tails" sequestered within the micelle and the globular hydrophilic domains at the perimeter facing outwards. Figure 3 is a schematic of such a micelle containing five monomer units.

Formation of Micelles From Recombinant LFA-3 Micelles can also be formed from recombinantly produced LFA-3. The preferred method of producing recombinant LFA-3 is in a eukaryotic expression system that enzymatically attaches the PI anchor, Chinese Hamster Ovary (CHO) cells.

The DNA sequence for human TM LFA-3 cDNA and the deduced amino acid sequence are known (Wallner, et al., J. Exp. Med., 166:923-32 (1987)), and are given in Fig. 4 in which potential N-glycosylation sites are indicated by arrows and the potential transmembrane domain is underlined by a broken line. These sequence WO 89/10938 PCr/US89/01858 10 data are available from the EMBL/GenBank Data Libraries under accession number Y00636. The cDNA sequence for the PI form of LFA-3 is identical to that of the TM form except that it is slightly shorter. The cloned cDNA encoding the PI glycan linked form of LFA-3 is expressed in a CHO cell system by standard techniques. These cells recognize a transmembrane domain sequence and covalently bind the PI anchor to the remainder of the pr.otein after cranslation. Alternatively, the TM form of LFA-3' can be produced in CHO cells or other cell systems that are not capable of attaching the PI anchor, such as murine L cells (Flavell et al. 1987).

The recombinant TM LFA-3 or mLFA-3 is isolated and caused to form micelles in the same manner as naturally occurring LEA-3, as described above.

Competition between LFA-3 Micelles and CD2 MAb.

A number of CD2 MAb's were tested for their ability to block binding of mLFA-3 to Jurkat cells or peripheral T cells at saturating concentrations of MAb (Table CD2 MAb's which strongly inhibit E-rosetting were able to completely block mLFA-3 binding.

MAbs used in this study are listed in Table 2.

9-1 was a gift of Dr. Dupont (Memorial Sloan-Kettering Cancer Institute, NY). 9.6 was a gift of Dr. John Hansen (Fred Hutchinson Cancer Research Center, Seattle,

WA).

Binding of LFA-3 Micelles to Cell Surface CD2 It has been shown that iodinated LFA-3 binds to CD2 positive cells when Triton X-100 is absorbed with 15% bovine serum albumin, and that this binding is inhibited by CD2 Mab and LFA-3 MAb (Dustin, et al., supra) According to the present invention, mLFA-3 micelles were found to bind efficiently to Jurkat T leukemic cells or resting PBL-T (Table mLFA-3 i. I WO 89/10938 PCT/US89/01858 11 micelle binding was inhibited by the TS2/18 MAb, the LFA-3 MAb, and cold (unlabeled) mLFA-3 micelles.

Figure 5 is a graph of the dissociation time-course for mLFA-3 from Jurkat cells. When cells which had bound mLFA-3 micelles at 4°C were resuspended in a large volume of medium containing no mLFA-3, the dissociation of bound mLFA-3 micelles was very slow (t 1 /2>600 min.; open squares), while in the presence of 10 ig/ml TS2/18 MAb (filled square) the dissociation rate was much more rapid (t 1 /2=60 When cells were resuspended in media containing 100 pg/ml TS2/18 MAb (filled triangle) the tl/2 was less than 2 min. and dissociation was complete within 1 hour. Because TS2/18 MAb and mT'A-3 appear to bind CD2 in a strictly competitive manner, the increased rate of dissociation in the presence of CD2 MAb indicates that the mLFA-3 micelles bind multivalently to cell surface CD2.

The binding domains of these proteins in the micelles participate in multiple interactions when the micelle contacts the surface of a cell, thereby greatly increasing the avidity of the proteins for their targets. A single protein-target interaction has a tendency to rapidly dissociate whereas the micelle can cause up to approximately 5 to 10 such interactions to occur simultaneously. Thus the overall dissociation constant is greatly decreased for the micelle form of an adhesion protein due to this multivalent binding. To verify this, we compared the ability of mLFA-3 micelles and non-micelle sLFA-3 to bind to cell surface CD2.

LFA-3 from JY cells and erythrocytes exhibit different degrees of glycosylation. Therefore, both mLFA-3 and sLFA-3 were isolated from JY cells to rule out differences in binding arising from differences in WO 89/10938 PC'/US89/01858 12 glycoslyation. mLFA-3 micelles from JY cells were of similar size to mLFA-3 micelles from erythrocytes despite being a mixture of PI and TM forms of LFA-3 (Dustin et al., Nature, supra). sLFA-3 was iodinated to the same specific activity as mLFA-3, and after incubation with Jurkat leukemic T--cells for 90 min at 4°C cells were spun through an oil cushion. The level of JY mLFA-3 micelle binding was found to be 150 fold higher than JY sLFA-3 binding (Table Consistent ID:. with. this, cold sLFA-3 did not appreciably inhibit the binding of mLFA-3 to Jurkat cells when present at a 100 fold weight excess. Therefore, multivalency appears to be an important factor in high avidity mLFA-3 binding to cell surface CD2.

The complete reversal of mLFA-3 binding by CD2 MAb described above shows that LFA-3 is not endocytosed after 90 min. at 4 0 C and that it is not inserting into membranes through its PI glycan anchor during this period. To further examine endocytosis of mLFA-3, LFA-3 was bound to Jurkat cells or PBL for 30 min. at 37 0 C or min. at 4"C. The cells were then washed with cold media and then incubated for 6 hrs. with 100 pg/ml TS2/18 MAb (Table LFA-3 bound at 4°C was almost completely released by TS2/18 MAb, but 50% of the Jurkat cells and 20% of the peripheral blood T lymphocytes (PBL-T) associated counts bound at 37 0 C remained cell associated in the presence of TS2/18 MAb (Table suggesting that this proportion of the LFA-3 was not on the cell surfata, but had been internalized, When TS2/18 MAb wi@ added with mLFA-3 at 37 0 C, no association of LFA-3 with the cells was detected, demonstrating that the persistent association is not due to non-specific lipid mediated insertion of mLFA-3 into the plasma membrane or to fluid phase pinocytosis.

WO 89/10938 PCT/US89/0 1858 13 The purified mLFA-3 was tested for binding to purified CD2 reconstituted into liposomes bound to glass coverslips. Iodinated mLFA-3 micelles were found to bind to this solid phase (Table The binding was inhibited by LFA-3 MAb and TS2/18 MAb, but not by the CD2.1 MAb, demonstrating that mLFA-3 binds to CD2 in the absence of other cell surface components.

Equilibrium binding experiments were done to determine the number of mLFA-3 molecules bound per cell and the avidity of the interaction with CD2 on Jurkat cells and resting PBL-T cells. The results are shown graphically in Figures 6a and 6b. The percentage of active mLFA-3 protein micelles after iodination was determined by measuring the proportion of mLFA-3 which could be bound to three sequential aliquots of Jurkat cells. This was between 85 and 95% for different preparations. Specific mLFA-3 binding was defined as any binding inhibited by excess TS2/18 MAb. Binding of mLFA-3 was done for 1 hr. at 4°C in the presence of control MAb (squares) or TS2/18 MAb (triangles). mLFA-3 bound saturably to Jurkat cells (filled symbols) and PBL-T (open symbols) (Fig. 6a). Scatchard plots showed that the Kd for Jurkat CD2 was 1.7-2.2 nM, while the Kd for PBL-T was 12-16 nM (Fig. 6b). This Kd value is a measure of the avidity of the LFA-3 micelles for each of these cell types. Approximately 420,000 mLFA-3 monomers bound per Jurkat cell, while about 50-80,000 mLFA-3 molecules! bound to PBL-T measured at saturation.

These values are consistent with the relative numbers of CD2 molecules on Jurkat cells and PBL-T measured using equilibrium MAb binding (Martin, et al., J.

Immunol., 131:180 (1983), Plunkett, et al., J.

Immunol., 136:4181 (1986)). The presence of saturating concentrations of the non-blocking CD2.1 !MAb did not WO 89/10938 PCY/ US8901 858 14 significantly affect either binding parameter for Jurkat cells- or PBL-T.

Use of Micelles to Increase Proliferation of Cells mLFA-3 micelles alone have no effect on PBL-T proliferation, but they can induce proliferation of PBL-T in the presence of submitogenic concentrations of anti-CD2 MAb's. The combination of mLFA-3 micelles nM) with CD2.1 MAb was found to be strongly mitogenic for. pgri-heral blood mononuclear cells (PBMC) from all donors tested (Table This response was usually (9 of 10 donors) seen in the absence of exogenous IL-2, although phorbol myristate acetate (PMA) was still required for a maximal response. Proliferation induced by the combination of mLFA-3 micelles and CD2.1 MAb was generally lower than that obtained with phytohemagglutanin (PHA), while the combination of mLFA-3, CD2.1 and PMA resulted in greater thymidine incorporation than PHA alone by up to two fold in some donors. The combination of mLFA-3 and CD2.1 MAb was also mitogenic for nylon wool enriched 'i cells (Table 8).

The combination of sLFA-3 at concentrations up to 800 nM and CD2.1 MAb, with or without PMA, was not mitogenic for PBMC in two experiments for which sufficient amount of material could be obtained.

Experiments with the same donors showed strong responses to mLFA-3 plus CD2.1 MAb with or without PMA.

Stimulation of proliferation by a combination of: 9- and 9.6/35.1 MAbs was seen with nylon wool enriched PBL-T or PMBC in the presence of 15% human serum (Table Under these conditons the combination of 9-1 MAb with 40 nM mLFA-3 was without effect except in one of two experiments done in round bottom wells with PBMC, in which case relatively weak stimulation augmented by exogenous IL-2 was seen (stimulation index WO 89/10938 PCrUS8/01 858 15 16.8 vs 110 for 9-1 9.6 MAb's). With nylon wool enriched T cells on flat bottom wells, mLFA-3 cooperated with 9-1 MAb only in the presence of 1 nM PMA. mLFA-3 appears to cooperate with 9-1 MAb for resting T cell activation under some conditions, but this activation is very weak compared to that achieved by a combination of mLFA-3 and CD2.1 MAb.

Because the mLFA-3 with CD2.1 MAb response was stronger and more reproducible than the 9-1 MAb with mLFA-3 response, the former was further characterized.

Figures 7a and 7b are graphs showing the time-course and dose-response for LFA-3 and CD2.1 MAb induced proliferation. PBMC were treated with mLFA-3 (40 nM) CD2.1 (104g/ml) (open squares); mLFA-3, CD2.1, and IL-2 (filled squares); or mLFA-3, CD2.1, and PMA (open circles) from day zero until the 16 hour pulse was initiated on day 3. Stimulation of DNA synthesis by mIFA-3 and CD2.1 MAb peaked on days 3 to 4. (Figure 7a). In the presence of exogenously added IL-2 and 1 nM PMA the optimal periods were days 3 to 4 and days 4 to respectively.

The dose-responses for mLFA-3 and CD2,1 MAb were determined in the presence of saturating concentrations of the other reagent. PBMC were treated with the indicated concentration of mLFA-3 with no addition (open squares), PMA (1 nM) (filled squares), CD2.1 (1Og/ml), or CD2.1and PMA (open circles).

Wells were pulsed on day 3 for 16 hr. In the presence of saturating CD2.1 MAb (67 nM), maximal PBMC or PBL-T proliferation was obtained with 4 nM mLFA-3 (Figure 7b). This is in the range of the Kd for mLFA-3 binding to PBL-T of 12 nM, The presence of exogenous IL-2 did not alter the dose-response, but, addition of I nM PMA decreased the mLFA-3 concentration required for WO 89/10938 PCIr/ISS89/01858 16 maximal response 10 fold. Similarly, in the presence of nM mLFA-3, the CD2.1 MAb concentration giving a maximal response was 6.7 nM, but only required for 0.67 nM in the presence of 1 nM PMA.

Inhibition of LFA-3-Dependent Activation.

The proliferative response to mLFA-3 with CD2.1 MAb was completely blocked by both anti-LFA-3 MA's and the 9.6 MAb. The 9.6 MAb blocks mLFA-3 binding to CD2 but- is usually not comitogenic with CD2.1 MAb (Table 9) The: resonse to mLFA-3 with CD2.1 MAb was also blocked by CD25 (anti-IL 2 receptor p55 chain) MAb suggesting that it is IL-2 dependent. The mitogenic response of PBMC or PBL-T to mLFA-3 with CD2.1 MAb was accompanied by aggregation of cells within 16 hours.

These clusters could not have been the result of passive agglutination of cells by mLFA-3 micelles, because clusters were not seen with mLFA-3 in the absence of CD2.1 MAb.

+2 Effect of LFA-3 Micelles on Intracellular Free Ca' There are several early indicators of T cell activation useful in evaluating the activating potential of mLFA-3. Several studies have shown that combinations of CD2 MAb can stimulate Ca 2 fluxes in T cell tumor lines. We have examined the ability of mLFA-3, alone and in combination with CD2.1 MAb, to stimulate Ca+ 2 mobilization in Jurkat cells and PBL-T. Because +2 [Ca +2 responses can occur within minutes and can be measured on single cells, their detection demoSstrates that the binding of the mLFA-3 micelles to cell surface CD2 leads to an activation signal without a chance for LFA-3 to insert into membranes or cross-link CD2 molecules on different cells. PBL-T or Jurkat cells were loaded with the fluorophore Indo-1 by incubation with its precursor, Indo-1 acetoxymethyl ester, and WO 89/10938 PCY/US89/018588 17 relative [Ca+ 2 was determined by flow microfluorimetry using the ratio of Indo-1 emissions at different wavelengths (410 nm/480 nm) as an indication of (Ca+2i When mLFA-3 (up to 240 nM) was added to Jurkat cells (at 30-60 sec.) there was no change in thi ratio within 15 min. (Figure Similarly, CD2.1 MAb (67 nM) alone had no effect on [Ca+2] i (Figure 9).

However, when mLFA-3 and CD2.1 MAb were added simultaneously, there was a rapid and sustained increase in: [Ca+2 i (Figure 9) which was similar in magnitude to that seen with CD3 MAb, OKT3 (Figure 8).

Similar results were obtained with PBL-T using 120 nM mLFA-3. There was no evidence of cell-cell interaction during these experiments, as there was no change in forward angle light scatter.

To allow direct comparison of Ca 2 mobilization to equilibrium mLFA-3 binding data, Jurkat cells were equilibrated with different concentrations of mLFA-3 by incubation for 30 min at 37 0 C, and then CD2.1 +2 MAb was added to initiate Ca mobilization (Figure Under these conditions, a near half-maximal increase in [Ca+ 2 was seen at about 1.2 nM mLFA-3, which is very close to the Kd for mLFA-3 binding to Jurkat CD2 determined as described above. The addition of ionomycin to Jurkat cells after 240 nM mLFA-3 and CD2.1 resulted in a further increase in [Ca +2]i suggesting that the Indo-l was not at saturation for Ca 2 in'these experiments. Therefore, it appears that the signal delivered through CD2 by a combination of mLFA-3 and CD2.1 which leads to an increase in [Ca+2 is proportional to the degree of saturation of CD2 with mLFA-3. Figure lOb is a graph showing histograms of sections through activation time plots at 3 minutes: 0.04 nM (solid line), 0.4 nM (dashed line), WO 89/10938 PCT/US89/018?i 18 1,2 nM (dot-dash line), 4 nM (dot-dot-dash line) and nM (dotted line).

Use of Adhesion Protein Micelles in Therapy Protein micelles of the invention can be used therapeutically to competitively inhibit reactivity of specific surface antigens on target molecules. In particular, LFA-3 micelles bind to cell surface CD2 on T-lymphocytes with high avidity, rendering them therapeutically useful by inhibiting binding of the T-lymphocytes to other cells.

Furthermore, LFA-3 micelles can be cointernalized with CD2, causing loss of CD2 from the cell surface. This irreversible internalization allows LFA-3 micelles to be used therapeutically below the Kd of monomeric LFA-3 Kd to down-regulate CD2.

LFA-3 micelles of the invention in a physiologically compatible carrier such as saline, are preferably administered to yield a therapeutically effective concentration of LFA-3 in the blood of 0.04 to 4 nM. Depending upon blood volume of the patient and clearance of the micelles from the bloodstream, 0.5 to pg/kg of micelle/patient/day are administered intravenously to achieve this concentration, Thae micelles function by effectively saturating CD2 receptor sites on the surfaces of T-cells and thereby inhibit binding of those T-cells to other cells. This binding inhibition can ameliorate the effects of disease states in which binding of T-cells to other cells is a continuing factor, autoimmune diseases such as rheumatoid arthritis; allograft rejection; and graft-versus-host disease.

Other embodiments are within the following claims.

WO 89/10938 PTU8/15 PCT/US89/01858 -19 Table 1 Inhibition of mLFA-3 binding by CD2 MAb Cold licgand 125 1-rnLFA-3 bound/10 6 cells_(CPM) control E4Ab 20,531 596 9.6 173 17 TS 2/18 153 57 9-2 154 8 CD2.8 173 29 35.1 7,171 98 CD2.9 164 CD2.1 22,817 362 9-1 24,092 878 D66 15,073 72 OKT3 21,242 487 Binding to Jurkat cells was done for 1 hr. at 4 0

C.

6 Input counts were 100,00/10 cells. Bound and free were separated by centrifugation through a 15% BSA cushion, All the MAb except control and OKT3 are against CD2, Results are average of quadruplicates with standard deviation and are representative of two experiments.

WO 89/10938 PTU8/15 PL"T/US89/01858 20 ?4Ab TS2/ 9-1 CD2.

CD2.

CD2.

9.6 D66 9.2 35.1 TS 1/ TS2 OKT3 1 2.

3.

4.

6.

7, 8.

18 1 .Specificity CD 2 CD2 CD2 TABLE 2 Isotype mouse IgG1 mouse igG3 mouse IgG1 Form purif ied purif ied purif ied Reference 1 3 or ascites 2 8Cfl2 rat IgG2a ascites 2 9 CD2 rat IgG2a ascites 2 CD2 mouse IgG2a ascites 4 CD2 mouse IgG2b ascites CD2 mouse IgM ascites 6 CD2 mouse IgG2a ascites 7 '18 LFA-11B mouse IgG1 ascites 1 '9 LFA-3 mouse IgG1 purified 1 CD3 mouse IgG2a culture sup 8 Sanchez-Madrid, et al., Proc. Natl. Acad. Sci.

USA, 79: 7489 (198,9) Olive D. et al.. Euro.- J. I-mmunol. 16:1063 (1986), Bernard et al,, Human Immunology, 17:388 (1986).

Kamour M. et al., J. Exp. Med. 153:207 (1981).

Bernard et al., J. Exp. Med. 155:1317 (1982) Bc -nard et al. Summary of the CD2 Workshop: In Leukocyte typing III, McMichael, ed,, Oxford Univ. Press, Oxford, p. 106 (1987)).

Martin et al., J. Immunol., 131:180 (1983).

Chang, T,W. et al., Proc. Nat. Acad. Sci. USA, 78: 1805 (1981).

WO 89/10938 PCT/ US89/0 1858 -21 Table 3 Binding of 1 2 5 -mLFA-3 to Jurkat cells and resting T cells.

125 I-mLFA-3 bound/l 6 cells (CPM~) Cold ligands Jurkat Cells resting T cells control MAb 2- _.513 33,262 mLFA-3 4,414 1,152 TS2/18.CD2 ?4Ab 440 S3 TS2/9 LFA-3 MAb 510 192 Binding was done for 1 hr. at 4 0 C. Input 125 -LFA-3 was 200,000 cprn. Bound and free radioactivity were separated by centrifugation through a BSA cushion. Control IgGl, TS2/18 and TS2/9 were added at l0pg/znl and cold rnLFA-3 was added at 5p~g/rnl. Results are averages of' duplicates and are representative of three experimtents.

WO 89/10938 PL-r/US89/01858 22 Table 4 Binding of PIPLC released LFA-3 to Jurkat Additions sLFA-3 mLFA-3 control MAb 5,553 387 148, 760 4003 TS2/18 4,030 14 5,023 23 Jurkat cells (10 6) were incubated with 500,000 CPM of either sLFA-3 or rnLFA-3, both prepared from JY cells, for 2 hrs. at 4 0 C and aliquots were centrifuged through oil cushions.

WO 89/1093$ WO 8910938PCT/US89/O 1858 23 Table Accessibility of bound mLFA-3 to extracellular TS2/18 Additions Cell Associated CPM Pre-incubation Post-incubation Jurkat PBL-T 4WC control control 11,441 500 1,543 225 TS2/18 control 86 5 95 14 control TS2/18 299 26 111 13 37WC control control 10,990 941 1,326 86 TS2/18 control 67 12 89 19 3,Q control TS2/18 4,975 400 390 43 Jurkat cells of PBL-T (5 x 10 5) were incubated with 125 ImLFA-3 (100,000 CPM) for 60 min. at 4 0 C or 30 min. at 37 0 C, Jurkat cell and PBL-T CD2 was initially at approximately occupation with LFA-3. The post-incubation at 40C without or with 7-2/18 was carried out !~or 6 hrs. to insure complete distociation of accessible LFA-3. Data are means of quadruplicate determinations with standard deviations.

WO 89/10938 P~ S9o15 PCT/US89101858 Table 6 Binding of mLFA-3 to purified CD2 in plastic bound vesicles Cold ligand 125 1-mLFA-3 bound/well control I4Ab 8,040 5-TS2/18 MAb Cfl2.1 MAb 8,122 TS2,9MAb 612 Input counts were 50,000. Bound and free were separated by five washes over 15 min. Results are averages of 1D' duplicates and are representative of re-5uits obtained in three S-x-Pdriments- WO 89/10938 PCr/1JS89/O1 858 Table 7 Induction of T cell proliferation b.y rmLFA-3 and CD2.1 Additions 3H-thyrnidine incoKrora b,!'-c31O PBMC (CPM4) buffer 593 392 4 17 mLFA-3 411 64554 CD2.1 437 839 646 IL-2 1,451 2,145 1-,034 mLFA-3 IL-2 1,345 1,399 1,284 S CD2.l+IL-2 1,609 2,039 1,545 PMA 1,713 1,382 834 mLFA-3 PMA 1,549 978 828 CD2.1 PMA 1,704 1,252 1,009 mLFA-3 CD2.1 22,294 35,439 47,478 j~mLFA-3 CD2.l IL-2 22,497 37,065 45,384 rnLFA-3 CD2. 1 PMA 77,784 1.14,313 153,240 Additions were as indicated on day zero: rIL-2 (o.lng/ml), PMA CD2.l (1:500 dilution of ascites or lOpg/ml), and mILFA-3 (40n.M). PHA (1l±g/ml) gave approximnately 80,000 100,000 cprn on day 3 to 4 in these experiments. Wells were pulsed with (3HI-thymidine on day 3 for 16 hr. Results are means of triplicates and are represerititive of 14 experiments.

WO 89/10938 PTTS9O15 PCr/US89/01858 26 1ID, Table 8 Induction of T cell proliferation by mLFA-3 and 9-i Additions 3 H-thymidine incorporation/5xlO 4 cells(CPM) PBMC T cells buffer 2,029 262 CD2.l 3,066 565 9-1 2,045 510 mLFA-3 1,145 342 CD2. 1 PMA 4,748 1,159 9-1 PMA 10,520 mLFA-3 1I4A 2,463 2,052 CD2.l rLGA-3 263,326 21,476 CD2.l mLFA-3 PMA 330,866 216,933 9-1 IL-2 9-1 mLFA-3 9-1 MLFA-3 PM A *4 ,078 *7,193 1,074 77 ,582 9-1 MLFA-3 IL--S *34,074 9-1 9o6 *181,570 40,764 PRMC and nylon wool enrich~ed T cells were from different donors. Experiment with PBMC was done in the presence of 15% heat inactivated human serum. Points with() were done in round bottom wells. Corresponding conditions in flat wells gave lower stimulation. Additions were made on day zero and wells were pulsed on day 3 for 16 h~r, 9-1 MAb was added at 1lig/ml; 9.6 MAb at 1:1000 dilution of as.iites.

Other concentrations as in Table 4. Restilts are means of triplicate determinations and are representative of two experiments.

WO 89/10938 PT S9015 PCT/US89/01858 27 Table 9 Inhibition of proliferation by MAb to LFA-3, CD2 and LFA-l Additions 3 H-thymidine incorporation/5xl10 cells(CPM) CD2.1 LFA-3 CD2.l mLFA-3 PMA control MAb 35,439 114,313 TS2/9 1,684 1, 116 9.6* 776 6,375 PBMC were used. All additions were m'ade on day zero and wells were pulsed on day 3 for 16 hrs. *9.6 CD2.l gave 549 cpm and 9.6 CD2.1 PMA gave 1663 cpm with this donor.

TS2/9 MAb was added at l0~pg/ml. 9.6 was added at 1:5000 dilution of ascites. All results are means of triplicate determinations from a single experiment and are representative of three experiments.

Claims (12)

1. A micelle of an adhesion protein which naturally includes a phosphatidylinositol lipid anchor, said micelle being capable of binding multivalently to a plurality of target molecules on a cell surface.

2. The micelle of claim 1, comprising less than about ten molecules of said protein.

3. Multimeric, purified LFA-3 which has an avidity for CD2-bearing cells such that the Kd of said LFA-3 for said cells is lower than the Kd of monomeric LFA-3.

4. The micelle of claim 1, wherein said adhesion protein is mLFA-3. The micelle of claim 1, wherein said i hesion protein is synthesized by recombinant DNA techniques.

6. A method of inhibiting tne adhesion of a first cell to a second cell, said first cell bearing on its surface a protein which is capable of binding to a molecule present in multiple form on said second cell and. which naturally includes a phosphatidylinositol lipid anchor, said method comprising contacting said second cell with a micelle of said protein comprising less than about 10 protein molecules and being capable of binding multivalently to said multiple molecules on said second cell. WO.89/10938 PC/US9/01858 29

7. A method for the treatment of a patient having a medical condition characterized by the pesence of an excess of activated T-cells, comprising administering to said patient LFA-3 micelles in a physiologically compatible buffer to achieve a bloodstream concentration of said LFA-3 effective to inhibit the binding of said activated T-cells in said patient to other cells of said patient.

8. The method of claim 7, wherein said bloodstream concentration of LFA-3 micelles following administration is 0.04 to 4 nM.

9. A method of increasing the avidity and decreasing the dissociation constant of an adhesion protein which naturally includes a phosphatidylinositol lipid anchor, comprising solubilizing said adhesion protein in a detergent solution, concentrating said protein in said detergent solution, and removing substaatially all detergent from said solution to form adhesion protein micelles having said hydrophobic protein ends sequestered therein. The method of claim 9, wherein said protein is LFA-3.

11. The method of claim 9, wherein said protein is produced by recombinant DNA techniques.

12. A method of effecting activation or proliferation of peripheral blood mononuclear cells or T-celis, comprising contacting said cells with LFA-3 micelles together with an antibody to CD2. WO 89/10938 WO 8910938PT/US89/O1 858 1/7 0 U N I' it ii I' I' i I It I' I' 6 7 'Volume (ml) FIG. I '0 U 0 In) fA 100 150 600 Time (min.) SUBITOSTT SHEET WO 89/10938 PCT/US89/o 1858 2/7. FIG. 2 TM FIG. 3 -D +R -0 T N S N SUM G3IITUTE SHEET (GA CIAGCC ATGGTTGCTGGGAGCGACGCGGGGCGGGCCCTGGGGTCCTCAGCGrGGTCTGCCTGCTGCACTGCTTT Mv A G SO0 A G A A L G V L S V V C L L HC F 'SC F So0 -28 AACA&&rATAGTGTTGTTATGGGAAT GTAACTTTCCTrrACCAAGCAATTAAAGGTCCTAT.I G G T v f" rFNVP S m V P L K E V L w K Q K 0 K V A 36 AGAAC TGGAAAATTCTGAATTCAGAGCTTTCTCATCTTTTAAAAATAGGGTTTATTTAGACACTGTGTCAGGTAGGCTCACTATCTACAACTTAACATCA 300 E L. f N S f F R A f S S F K N R V V L 0 T V S G S L T I Y N L T j 69

172-73 riA TCAGA rGAAGATGAGTATGAAATGGAATCGCCAAATATTACTGATACCArGAAGTTCTTTCTrTATGTGCTrGAGTCTCTTCCArCTCCCACACTAACTT 400 S0 E 0 E y E '4 E S P N I T 0 T U K F F L V V L E S L P S P T L' T C 103 A TIOS A L T t4 G S I E V Q C M I P C H~ Y N S. R G L f M v S W 0 C P M E 136 A T68 GCAA rGrAAACGTAACTCAACCAGTATATArTTTAAGATGGAAAATGATCTTCCACAAAAAATACAGTGTACTCTTAGCAATCCATTATTTAATACAACA 600 0 C N S T S I V F K M E N 0 L P I K c r C L S N P L F N T T 169 A A TCArCAA TCATTTTGACAACCTGTATCCCAAGCAGCGGTCATTCAAGACACAGATATGCACTTATACCCATACCATTAGCAGTAATTACAACATGTATTG 700 S S I I L T T C I S S G H S R H1 R VALE- P -I P -LA V I T T CI _V 203 TGCTGTATA rGAATGGTA TTGAAATGTGACAGAAAAC-CAGACAGAACCAACTCCAATTGATTGGTAACAGAAGATGAAGACAACAGCArAACTAAATT 800 L M G L K_ 222 AT TTTkAAAAACTAAAAAGCCATCTGATTTCTCATTTGAGTArTACA/&TTTTTGAACAACTGTTGGAAATGTAACTTGAAGCAGCTGCTTTAAGAAGAAAT 900 ACCCA CrAACAAAGAACAAGCATTAGTTTTGGCTGrCATCAACTTA rTATATGACTAGGrGCTTGCTT.TTTTTGrCAGTAAATTGTTTTTACrGATGATG 1000 TAGATACrTTTGTAAATAAATGTAAATAIC-rACACAAGTG(AAA.AAAAAAA) 1040 FIG. 4 WO 89/10938 PT1S9O PCr/US89/01858 4/7 FIG. 6A Input (CPM x 10O 6) FIG. 6B3 600 700 Bound (pM) S~1EET PCT/US89/O1 858 WO 89/10938 5/17 FIG. 7A C 0 0 I.- 0 C 4., C 0 E H C 0 0 0~ 0 C., C C E H 300 O 200 X Cj 100 0 1 23 4 Day of pulse FIG. 7B 5 6 150 S100 0~ 0 0.04 0.4 mLFA-3 4 (nM) J F I Relative raiio (410 nm/480 nm) RELATIVE RATIO (4 10 nm/480 nm) (o (.0 A A ,q..l~I.n RELATIVE CELL RELATIVE RATIO (4 I10 nm/480 nm) N~UMBER L. rn0 r' 'El 1 0 03) -n 0 L IL SE601/68 OM 99910/68SflhI3d INTERNATION1AL SEARCH REPORT International Application NoPCT/US 89 /01858 1. CLASSIFICATION OF SUBJECT MATTER !it several classification symbols apply, indicate all) 6 According to International Patlit Classification (IFC oI both NeUQna CjsIfion and IPC IPC C07K 9/00 ;CO7K 1%174; AbLKa i l US CIL. 5301322;530/395;530/396;530/39-4-514/8;514/885;424/450 If. FIELDS SEARCHED Minimum Documentation Searched 7 Documentation Searched other than Minimum Documentation to the Extent that such Documents are Included In the Fields Searched I Chemical Abstracts and BioSiS III. .DOCUMENTS CONSIDERED TO BE RELEVANT 9 Category Citation at Document, 'I with Indication, where appropriate, at the relevant Passages 12 Relevant to Claim No, 13 YKUS 4,169,090 (MURRAY) 25 September 1979 1-5,9,10 (25-09-1979) "Protein 'Product and Pro- F cess for Preparing the Same" See the Abstract. Y US 4,578,269 (MOREIN) 25 March 1986 (03- 1-10 25-1986) "Immunogenic Protein or Peptide Complex, Method tor Producing Said Complex and the Us "e Thereof as ran im- mune Stimulant and as a Vaccine" See the Abstract, column 1, lines 49-50 and column 2, lines 13-24, column 4, lines 5-35 and column 15, E.iperiment 1. Y Stephen Shaw, Nature, 323,262, 1986 6-8 "Two Antigen-Independent Pathway Used by Ruman Cytotoxic T-Cell Clones" See the paragraph bridging pages 263 to 264. Special categories of cited documents; to rater document Published alter the international filing date document defining the general state of the art, whlc',,Is not orpioyd at n o ncnlc ihtea~l~o u considered to be of particular relevance cite to understand the principle or theory underlying the invention I.El earlier document but published on or alter the International document of particular relevance: the claimed Invention fing date cann I -,onsidered novel or cannot be considered to 11L" document which may throw doubts on priority claim(s) or invo'e an invenrtive step which isocited to establish the publication date at another c'ccu mean t a atcular relevance; the claimed invention citation or other special reason (as specified) :annot be cofn ,;did to Involve an inventive step when the 110" document referring to an oral disclosure, use, exhibition or document is combined with one or more other such docu- other means ments, such combinatiorn heing obvious to a person skilled IT" document pub lshed prior to the International filung date bk.c in the amt later than the irlority date claimed document member of the same pste-it lamily IV E ITIFICATOI~l Dat* of the Actual qoampletion of the International Search Date of Mailing at this International Search Report JULY 1989 16AUJG 1989 Ifltenational Serchthg Authority Sgnatur.ol A tw d Officer,Q_ VIFatemahi Moezie IrtIernalionaI AcoiPc.1gn No PC'm/r7 89/nlpqQ MI OOCUMENTS CONSIOEREO TO 81 RELEVANT (CONTINUEO FROM THE SEC~ON SI4EMr Category Cqrji~on of Ooczme"r -sinr indhi'tion. -hr~e sfoorooriate. of the reley irt oass.qes Rel,anlr 0o Claim No Sten r4ammarstrom, Scand. J. Immunol., 9-iLl Sa-56, 1973 "Binding of Helix. Po- mat.a, A Wemagglutinin to Human Eryth- rocytes arnd Their Cells" See the abstract. X L.S. Wolf, Clinical Research, 34, No. 6-8 2, 1986 "Monoclonal Antibodies tEo Til, TjFA-2 and LFA-3 Antigens Inhibit Binding of Human Thymocytes to Autolo- gous Thymic Epitheliall Cells" See 647A, first paragraph. k Barbosa, J. of Immun., 36, No.8, 3085, 1-12 L9876' "The Map~pinEg-and Somatic C--_11 Hy- brid Analysis of the Role of Human Lymphecyte Function-Associated Antigen-3 (LFA-3) in CTL-Target Cell Interaction". A Paul J. Martin, J. of Immun. 131, No. 1 1-12

180-185, 1983 "Identification and Func- tional Characterization of Two Distinct Epitopes on the Human T-cell Surface Protein Tp A Francisco Sanchez-Madrid, Proc.s Natl. Acad. Sci. 79, 7489-7493, 1982 "Three itin-ct Antigens Associated with Human -T-Lymphoyte-Mediated Cytosis: LFA-l, LFA-2, and LFA-3. F-XM.V~wW"(MlIjM