0041-1337/04/7708-1288/0 TRANSPLANTATION Copyright © 2004 by Lippincott Williams & Wilkins, Inc. Vol. 77, 1288–1294, No. 8, April 27, 2004 Printed in U.S.A. A NOVEL MONOCLONAL ANTIBODY INHIBITS THE IMMUNE RESPONSE OF HUMAN CELLS AGAINST PORCINE CELLS: IDENTIFICATION OF A PORCINE ANTIGEN HOMOLOGOUS TO CD58 KATHERINE CROSBY, CHRIS YATKO, HAROUT DERSIMONIAN, LUYING PAN, Background. Human CD58 is an adhesion molecule that interacts with CD2 on lymphocytes. We describe here an antibody that blocks responses of human peripheral blood mononuclear cells (PBMCs) to porcine cells and reacts with a porcine protein with homology to CD58. Methods. Antibodies were isolated with a screen for inhibition of the human antiporcine response. One of these antibodies was used for immunoaffinity purification of a protein that was identified by molecular weight determination, endoglycosidase sensitivity, and microsequencing analysis as a porcine homologue of CD58. Results. The antigen recognized by this antibody was a cell surface protein of relative molecular mass (Mr)ⴝ45,000 containing N-linked carbohydrate chains. Immunoaffinity purification of this protein and microsequencing revealed homology to sheep CD58 as well as sequences that were common to this protein and both sheep and human CD58. The protein was widely distributed on porcine cells, including lymphocytes, endothelial cells, muscle cells, and neuronal cells. This antibody efficiently inhibited lysis of porcine targets by human PBMCs in addition to preventing proliferation of the human PBMCs in response to the porcine cells. Conclusions. The CD2 interaction with porcine cells is important for the efficient recognition of porcine tissue, and inhibition of the human antiporcine immune response with the antibody is likely to be caused by the disruption of the human CD2 interaction with this porcine homologue of CD58. The antibody may prove to be useful for the blocking of this interaction without interfering with other functions of T cells. Xenogeneic transplantation could provide an alternative that would overcome the limited availability of human cells and organs for allogeneic transplantation. In the attempt to intervene in the immune response to porcine cells and organs, investigators have identified molecules on the pig tissue that are important in the development of the human immune response. Development of reagents that can block the immune response to porcine tissues will be advanced by knowledge of the molecular targets for the humoral and cellular immune response. Thus, the understanding of the importance of alpha-linked galactose to the humoral response has resulted in intensive efforts to overcome antibody (Ab)-mediated rejection by removing this antigen from the AND ALBERT S. B. EDGE pig tissue or inhibiting the complement-mediated rejection of the xenografts (1–3). The identification of the key molecules in the T-cell and natural killer (NK)– cell-mediated responses to grafts will lead to the development of agents that can interfere with rejection of grafts and increase our ability to transplant tissues and organs as replacement therapy for a variety of diseases. The important mediators of the cellular response to xenografts appear to be similar to the cellular mediators of the alloresponse. The rejection of porcine xenografts by the mouse is clearly mediated in part by T cells, and the restriction by the major histocompatibility complex (MHC) occurs with porcine tissue in a similar manner as the mouse, both using direct and indirect (through host MHC restriction) elements (1). The human response to porcine tissue is similarly mediated by T cells (1, 4). In addition to this T– cellmediated rejection, NK cells play a role in the rejection of xenografts (5–7). Few reagents are available to interfere with NK-mediated rejection. Thus, although some interactions between adhesion molecules and costimulatory molecules and their ligands, as well as between MHC with bound peptides and T-cell receptors, are of a lower affinity in xenotransplant pairs, the adhesion molecules and costimulatory molecules do appear to be recognized by T cells and NK cells across species (8, 9). We have initiated studies to identify reagents that will recognize cell-surface molecules on porcine cells and interfere with recognition by human T cells and NK cells. We have previously shown that the T-cell response to porcine tissue is inhibited in vitro by MHC class I antibodies (Abs) that shift the cytokine profile to a Th2 dominated response (10). In the present study, we identify a mouse monoclonal Ab that recognizes a porcine homologue of CD58 and blocks the human antiporcine response in vitro. MATERIALS AND METHODS Immunization and Production of Monoclonal Antibodies Porcine erythrocytes and lymphocytes were used for immunization of Balb/c mice. Cells were injected biweekly for 6 weeks and 3 days before splenectomy. On the day of fusion, the spleen was removed aseptically and transferred to a conical vial containing Hanks’ balanced salt solution (HBSS). Splenocytes were collected and washed in HBSS, and erythrocytes were lysed by addition of red blood cell (RBC) lysing solution (Sigma, St. Louis, MO). The P3X63Ag8.653 cell line was obtained from American Type Culture Collection and maintained in Dulbecco’s minimum essential medium Diacrin, Inc. Charlestown, MA. (DMEM, Gibco) supplemented with 10% heat-inactivated fetal boAddress Correspondence to: Albert Edge, Eaton-Peabody Labora- vine serum (FBS, Hyclone). On the day of fusion, 2.5⫻107 splenotory, Massachusetts Eye and Ear Infirmary, 243 Charles Street, cytes from Balb/c mice immunized with porcine RBC or peripheral blood mononuclear cells (PBMCs) were mixed with 2.5⫻107 Boston, MA 02114. E-mail: albert–edge@MEEI.harvard.edu. Received 11 September 2003. Accepted 2 November 2003. P3X63Ag8.653 cells and were pelleted in a 50 mL tube. Then, 1 mL DOI: 10.1097/01.TP.0000120377.57543.D8 1288 1289 CROSBY ET AL. April 27, 2004 of 50% polyethylene glycol (Sigma) was added to the cell pellet drop-wise with gentle stirring over a 1-minute period. After gently mixing cells for an additional 1 minute, 5 mL of serum-free DMEM was added drop-wise over 5 minutes with gentle stirring. The fusion reaction was stopped by addition of 40 mL serum-free DMEM. Cells were centrifuged at 800 rpm for 5 minutes. The pellet containing fused cells was resuspended in 10 mL hypoxanthine, aminopterin, thymidine (HAT) selection medium (DMEM plus 10% FBS, 10% hybridoma enhancing supplement, 1X HAT, 1% sodium pyruvate, and 1% penicillin/streptomycin) at 2⫻106 cells/mL and plated at 100 ␮L/ well in 96-well plates. HAT selection medium was changed every 3 to 4 days for 2 weeks. On day 17, 50 ␮L of culture supernatants were collected for initial screening of Ab activity by fluorescenceactivated cell sorter (FACS) analysis. Hybridoma cultures with strong Ab binding activity were further expanded in a 24-well plate in HT medium (DMEM plus 10% FBS, 10% hybridoma enhancing supplement, 1X HT, 1% sodium pyruvate, and 1% penicillin–streptomycin). Cloning and Establishment of Hybridoma Cell Lines Ab-producing hybridoma cells were cloned by limiting dilution. Hybridoma cells were plated at 0.3 to 1 cell/100 ␮L per well in DMEM plus 10% FBS and 10% hybridoma enhancing supplement in a 96-well plate for 4 weeks and were rescreened for Ab-binding activity by FACS analysis. FACS Analysis Hybridoma supernatants were screened for their binding activity to porcine cells by FACS analysis. Briefly, 50 ␮L of hybridoma supernatant was incubated with 50 ␮L of porcine cells (erythrocyte or peripheral blood mononuclear cells [PBMC]) for 1 hour at 4°C. After washing, the cells were stained with donkey fluorescein isothiocyanate-F(ab’)2 anti-mouse immunoglobulin (Ig) (H⫹L) for an additional 30 minutes at 4°C. Cells were washed again and analyzed using a FACScan (Becton Dickson). Proliferation Assays To measure the proliferation of human PBMC in response to porcine cells, the human PBMC (2⫻105 cells/100 ␮L per well) were added to 96-well plates containing porcine smooth muscle cells (2⫻104 cells/100 ␮L per well) and coincubated for 7 days at 37°C. Porcine cells were preincubated with or without hybridoma supernatant or purified Ab (10 ␮g/mL) for 1 hour at 4°C. Ab was present during the entire coincubation period. 3H-thymidine (1 ␮Ci/well) was added for the last 20 hours of incubation. All experiments were performed in AIM V medium (Gibco) supplemented with 5% heat inactivated FBS (Hyclone). Cultures were harvested using a cell harvester (Packard Instruments), and thymidine incorporation was determined by counting the plate with a microplate scintillation counter (Model B9906, Packard). Cytotoxicity Assays Cytotoxicity was determined by 51Cr release assay as previously described (6). Human PBMC were used as effector cells, and porcine PBMC treated with concanavalin A at 5 ␮g/mL for 3 days were used as targets. Porcine PBMC blasts (2⫻106) were labeled with 100 ␮Ci 51 Cr for 2 hours at 37°C and washed three times afterwards. Effector cells were incubated with 10,000 labeled target cells at various effector:target ratios in the presence or absence of Ab for 4 hours at 37°C in a U-bottom 96-well plate. At the end of the assay, 100 ␮L of supernatants were transferred to a LumaPlate, and the plate was dried in the hood. 51Cr activity was determined by counting the plate with a microplate scintillation counter (Model B9906, Packard). Maximal and background release of 51Cr were determined by incubation of the target cells with 2% Triton-X (Sigma Chemical Co, St Louis, MO) or medium, respectively. The percentage of specific lysis is calculated as (%)⫽100⫻(sample count⫺background count)/(maximal count⫺background count). Immunopreprecipitation and Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis Porcine erythrocytes were diluted to 2.5⫻107/mL in 0.5 mg/mL NHS-LC-biotin (Pierce) in phosphate-buffered saline (PBS) and incubated at room temperature for 30 minutes with rocking. After washing with 10 mM glycine in PBS three times, cells were lysed with 1% Igepal (Sigma) in PBS plus aprotinin at 4°C for 60 minutes. Lysed cells were centrifuged for 15 minutes to pellet nuclei, and the lysate was transferred to a new tube. The lysate was precleared at 4°C with washed protein A-Sepharose (Repligen) and immunoprecipitated with hybridoma supernatants, rabbit anti-mIgG, and protein A Sepharose. After washing four times with 1% Igepal, samples were boiled for 5 minutes in 100 ␮L 2X LSB, supernatants were transferred to fresh tubes, and proteins were separated by 10% sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis in the system of Laemmli (11) and transferred to a nitrocellulose membrane. Detection of the biotinylated-proteins was accomplished by incubation of membranes with extravidin peroxidase (Sigma) followed by development with ECL components (Amersham). The wet membranes were then exposed to the radiographic film (Kodak, Rochester, NY). Affinity Purification of 12–99 Antigen The 12–99 Ab was coupled to CNBr-Sepharose (Pharmacia) according to the manufacturer’s instructions. An irrelevant Ab was coupled to a separate batch of CNBr-Sepharose. Lysate from porcine erythrocytes was passed over the irrelevant Ab column first, and the flow through from this preclearing column was then applied to the column of 12–99. The column was washed with 10 volumes of wash buffer (10 mM Tris-HCl, 0.15 M NaCl, pH 7.4) and eluted with 10 mL of elution buffer (10 mM Tris-HCl containing 50 mM diethylamine, pH 11.0). Fractions (0.5–1.0 mL) were collected into tubes containing 50 ␮L 1 M Tris, pH 6.2, and tested for the antigen by using enzyme-linked immunoadsorbent assay. The eluate from each tube was coated onto a well of a 96-well plate, and detection was carried out with 12–99 Ab followed by horseradish peroxidase– conjugated-anti-mouse IgG Ab and developed with the substrate o-phenylenediamine. The optical density was measured at 490 nm. Eluates with detectable 12–99 antigen were pooled and analyzed on a Phastgel apparatus (Pharmacia) with silver staining. N-Glycanase Digestion Biotinylated porcine erythrocytes were lysed in 1% Igepal at 2x108/mL and then immunoprecipitated with 5 ␮g of Ab. After washing four times with 1% Igepal, 50 ␮L of N-glycanase buffer (0.75 M Tris/0.2% SDS/0.5% mercaptoethanol) was added to each sample. The samples were boiled, and the supernatant after centrifugation was digested with 1U N-glycanase (Oxford Glycosciences) for 18 hours at 37°C. Proteolytic Cleavage and Peptide Sequencing The purified protein from the immunoaffinity column was run on a Phastgel and transferred to a nitrocellulose membrane. The protein eluted from the nitrocellulose membrane was used for microsequencing, both intact and after tryptic digestion. The reduced and alkylated protein was digested with trypsin, and tryptic peptides were resolved by high-performance liquid chromatography (HPLC) on a Zorbax C18 column. Individual fractions were subjected to MALDI mass spectrometry. The MALDI analysis was performed on a Finnigan Lasermat 2,000 instrument at the Harvard Microchem Laboratory. The intact protein and the peptides separated by HPLC were sequenced by Edman degradation using a PerkinElmer Procise 494 HT Protein Sequencing System. 1290 TRANSPLANTATION RESULTS Inhibition of the Human Antiporcine Response by Monoclonal Antibodies We immunized mice with porcine PBMCs or erythrocytes because both cell types express molecules of potential interest in the human antiporcine immune response. Hybridoma clones with Ab binding activity toward porcine PBMCs and erythrocytes determined by FACS analysis were further screened by inhibition of the human lymphocyte response to the porcine cells. Proliferation of human PBMCs can be stimulated by exposure to porcine-cell antigens as measured in an in vitro mixed culture system. We screened panels of murine hybridomas for their ability to inhibit this response. Supernatants from several hybridoma clones from the immunization with PBMCs (15–12, 19 –22, 19 –55) or erythrocytes (12–99) reduced the incorporation of thymidine into human cells that were responding to porcine cells in the mixed-culture system. In this assay, the porcine smooth muscle cells were incubated with human PBMCs in the presence of the supernatants for 7 days. Control Abs had no effect on proliferation, and the other Abs in the panel were unable to reduce proliferation. Proliferation was blocked 48% by 12–99 and 77% by 15–12 (data not shown). Immunoprecipitation of 12–99 Antigen Two of the Abs that reduced proliferation by human PBMCs (19 –22 and 19 –55, data not shown) were identified as MHC class I specific by immunoprecipitation studies (Fig. 1A) and were not further pursued in the present study because the inhibition of the human antiporcine response by Vol. 77, No. 8 such Abs has been described (10). Immunoprecipitation revealed that the 12–99 antigen had a relative molecular mass (Mr) of 45,000 (Fig. 1A). The 12–39 Ab, used as a control that did not reduce proliferation, immunoprecipitated several bands. The broad band for the 12–99 Ab was indicative of glycosylation. Treatment of the antigens precipitated by 12–99 and that corresponding to 15–12, which comigrated with 12–99, with N-glycanase demonstrated that the protein had an Mr of 26,000 after removal of the N-linked carbohydrate (Fig. 1B). The protein precipitated by Ab 15–19, an Ab used as a control that had no effect on proliferation, was more modestly affected by N-glycanase digestion (Fig. 1B). Several of the Abs that reduced the proliferation of human PBMCs (12–99 and 15–12) were purified and were further characterized. The proliferative responses against porcine cells were retested with the purified Abs. Abs 12–99 and 15–12 decreased the response 50% and 65%, respectively (Fig. 2). In a similar experiment, Ab against porcine MHC class I (PT-85) reduced the proliferation by 55%, whereas an Ab against human MHC class I (W6 –32) had no effect (data not shown). Inhibition of Porcine-Cell Lysis by Antibody 12–99 Porcine PBMC blasts are targets for the killing activity of human NK cells in an unprimed response of human PBMC mixed with porcine blasts in vitro (6). The killing could be inhibited by Ab to human CD2, which blocks the interaction of human CD2 on the NK cells with CD58-like molecules on the porcine targets (Fig. 3). Killing was not blocked by Abs to MHC class I. Ab 12–99 was a potent inhibitor of the NKmediated killing of porcine PBMC blasts (Fig. 3). FIGURE 1. Biotin labeled proteins immunoprecipitated by monoclonal antibodies. (A) Immunoprecipitation of cell lysate from cell surface biotin labeled cells was performed to screen for the relative molecular mass (Mr) of the proteins recognized by antibodies 12–39, 12–99, 19 –22, and 19 –55. After immunoprecipitation and sodium dodecyl sulfate (SDS) gel electrophoresis, the proteins were transferred to a nitrocellulose membrane and detected with streptavidin-horseradish peroxidase (HRP) followed by chemiluminescence. (B) Molecular weight and endoglycosidase sensitivity of the antigens immunoprecipitated by 12–99 and other monoclonal antibodies. Red blood cells were biotinylated, solubilized with 1% Igepal, and immunoprecipitated (as described in Methods) with 12–99, 15–12, and 15–19. The immunoprecipitates were then subjected to digestion with N-glycanase (N-glycanaseⴙ) or left undigested (-). Products were analyzed by SDS polyacrylamide gel electrophoresis. After transfer to nitrocellulose, the blot was developed with streptavidin-HRP, and detection was performed by chemiluminescence. The migration of the intact protein (top arrow) and N-glycanase digestion product are indicated (bottom arrow). Migrations of standard proteins are shown to the left of the gels. April 27, 2004 1291 CROSBY ET AL. FIGURE 2. Effect of purified 12–99 and 15–12 on the proliferation of human peripheral blood mononuclear cells (PBMC) in response to porcine cells. Porcine smooth muscle cells were preincubated with the purified antibody (Ab) at increasing concentrations for 1 hour, and cells were then cultured with human PBMCs for 7 days in the presence or absence of hybridoma supernatants. The proliferation of the human PBMC for the last 20 hours of incubation was measured as described. The 3H-thymidine incorporation for human PBMCs cultured alone was less than 6,000 cpm in this assay. Ig, immunoglobulin. Purification and Microsequencing of 12–99 Antigen Cell Distribution of 12–99 Antigen An immunoaffinity matrix was prepared with immobilized 12–99, and cell lysate from RBCs was passed through the column. Elution of bound antigen yielded a protein of Mr⫽45,000 that migrated as a single band detected on a Phastgel by silver staining (data not shown). The lyophilized protein sample was subjected to tryptic digestion. Both the intact protein and the tryptic peptides obtained by HPLC of the digested protein were analyzed by microsequencing. Tryptic peptides separated by HPLC were evaluated by MALDI mass spectrometry, and a major peak yielded a peptide of Mr⫽1,525 (Fig. 4). The sequence of the intact protein was obtained for the N-terminal 17 amino acids, and the tryptic peptides subjected to microsequencing were examined for overlap with the intact protein. The Cterminal position of the N-terminal peptide and the first position of the Mr⫽1,525 peptide were occupied by alanine (Fig. 5). Overlap of the two peptides at the alanine resulted in a sequence of 29 amino acids that was homologous to human and sheep CD58 and had a conserved Ile at position 5 and a conserved Ser-Gln-Xaa-Phe-Xaa-Glu-Ile-Xaa-Trp-Lys at positions 21 to 29. A total of 45% of the amino acids showed identity with the sheep sequence, and 28% were identical to the human (Fig. 5). Comparison of this sequence with that of rat CD48 indicated 3 of 29 positions had the same amino acid. The 12–99 antigen was widely distributed on porcine cells, with expression apparent in lymphocytes, RBCs, hepatocytes, cardiac cells, and fetal neuronal cells from the lateral ganglionic eminence and ventral mesencephalon (Fig. 6). Human cells were not reactive with 12–99 (data not shown). DISCUSSION By screening Abs for their inhibitory capacity in an assay that detects the response of human peripheral lymphocytes to porcine targets, we have found an Ab that interfered with the antiporcine response. Both the proliferative response of T cells to porcine cells and the antiporcine cytotoxicity manifested by human NK cells were inhibited by the Ab. The inhibition was dose dependent in the in vitro assays. The Ab interfered with the interaction between human effector cells and the antigen recognized by 12–99, which was identified as a glycoprotein homologous to CD58 on the surface of porcine cells. Inhibition of human NK– cell-mediated lysis of porcine cells and of T-cell expansion by the Ab could prove important in the prevention of the immune response to porcine tissue and should provide a new tool for preventing rejection of xenografts. In our previous work, we and others have demonstrated the importance of NK cells in the human response to porcine tissue (5–7). The sequence information on the porcine MHC 1292 TRANSPLANTATION Vol. 77, No. 8 FIGURE 3. Inhibition of cytotoxicity by antibody (Ab) 12–99. The killing of porcine cells by human peripheral blood lymphocytes (PBL) is attributable to natural killer (NK) cells. In the cytotoxicity assay, human PBL from donor 1 (A) or donor 2 (B) are mixed with 51Cr-labeled porcine PBL blasts for 4 hours at 37°C in the presence of 12–99, anti-hCD2 or no Ab as control. At the end of the assay, 51Cr released into culture supernatants was determined by counting the plate with a microplate scintillation counter. class I alleles indicated that the signals for inhibition of human NK cells were lacking in the porcine MHC (12). As a result, NK cells were the major effectors of porcine-cell lysis in vitro (6). However, recent studies have shown that these signals could be supplied by expression in porcine cells of human MHC (human leukocyte antigen [HLA]-G) (13, 14) or April 27, 2004 1293 CROSBY ET AL. FIGURE 5. Sequence of peptides derived from 12–99 antigen. The sequence obtained from the tryptic peptides is compared with the sequences of human and sheep CD58. The first 17 amino acids are from sequencing of the intact protein, and this sequence overlaps with the amino acids (positions 17–29) obtained from the tryptic peptide shown in Figure 4. Amino acid homologies with the human or sheep CD58 are in bold. FIGURE 4. Mass spectrometry of tryptic peptide from 12–99 antigen. The products of trypsin digestion were resolved by high-performance liquid chromatography (HPLC), and individual peaks were subjected to MALDI mass spectrometry. An HPLC peak (166) with an Mr of 1,525 was subjected to sequencing. of mutated HLA molecules (15) which human MHC molecules protected the porcine cells from NK cells. Survival of bone-marrow xenografts and induction of tolerance by mixed chimerism has been shown to be partially blocked by NK cells in the recipient (16). Improved methods for inhibiting the NK-mediated rejection of xenografts could include the use of Abs that inhibit the interaction of these effector cells with porcine targets, and the 12–99 Ab is a good candidate for such studies. The isolation of the antigen recognized by 12–99 allowed the protein to be partially characterized. Using the Ab, we obtained the antigen in quantities that could be subjected to microsequencing. Because the 12–99 monoclonal Ab was found to inhibit the response to porcine cells by human PBMCs, this antigen on porcine cells appears to play a role in the recognition of porcine tissues by human immune cells. Comparison of the amino acid sequence of this antigen to known sequences in a protein database revealed homology to human CD58 (lymphocyte function-associated antigen-3). The homology of this protein to human CD58 and its profile of expression and apparent function indicate that the antigen belongs to this protein family. It is likely that the CD58 homologue that we have identified is the molecule recognized by human CD2 (17) and that it plays an important role as an adhesion molecule on porcine targets of the human immune system. In contrast with the stimulation of human T cells with human endothelial cells, which was dependent on CD28, the activation of T cells by porcine aortic endothelial cells required both the CD28 pathway and the interaction of human CD2 with the surface of the porcine tissue (17). The level of inhibition observed with 12–99 was similar to that by a CD2 Ab (Fig. 3). The sequence homologies of the antigen identified here with human CD58 suggest that the protein is a ligand for CD2 and, in fact, there is extensive homology among the proteins that bind CD2, particularly in the regions involved in the binding surface of the molecule (18). This protein could be a CD58 homologue, as in humans, or a CD48 homologue, as in rats, where the latter is the major ligand for CD2. The extent of identity of the 12–99 antigen with the rat CD48 sequence was less (3 of 29 residues) than that with CD58, indicating a more likely homology with CD58. The importance of CD2 in the interaction of the human NK and T cells with porcine targets is consistent with reports that CD2 interactions are critical in the response to xenografts. Work by others on the human anti-porcine response (19) showed that anti-CD2 Abs prevented T-cell proliferation in response to porcine cells and that this interaction was critical for cytotoxicity, although the molecules involved were not elucidated, and reagents to interfere with T-cell binding to the porcine ligands were not made. The fact that the 12–99 Ab interferes with this interaction indicates that porcine CD58-like molecules do interact with human CD2, so that binding in this case is conserved across species. Moreover, the interference with killing is evidence that the CD2 interaction with CD58 is important for NK-mediated cytotoxicity in this system. Results with anti-CD2 reagents have also proved promising in vivo, in the case of allo- and xenografts that could be prolonged with the use of an anti-CD2 reagent (17). An Ab that prevented binding of these two proteins by interfering with the target side of the pair would be preferable, in theory, because the other effector activities for which CD2 is required, such as antitumor responses and viral immunity, would not be affected by such a reagent. Further work will be required to establish the identity of this molecule and to determine whether it is one of a family of CD2 ligands in pig. Molecular cloning of CD58 homologues is in progress. ADDENDUM After submission of this manuscript, a paper on porcine CD58 reported the sequence of the cDNA (20). The structure of the protein indicated that the porcine CD58 had the potential to interact with human CD2. The sequence deduced from the cDNA (GenBank accession number AY303628) is in 1294 TRANSPLANTATION Vol. 77, No. 8 FIGURE 6. Distribution of 12–99 antigen on various cell types from the pig. The cells were incubated with the indicated antibody (Ab) followed by fluorescein conjugated antimurine IgG. Cells were subjected to flow cytometry as described in Methods. Cells stained with isotype control IgG1 or purified 12–99 Ab followed by fluorescein isothiocyanate-secondary antibody. VM, ventral mesencephalon; LGE, lateral ganglionic eminence. close agreement with the amino acid sequence of the tryptic peptides obtained here (amino acid identity at 25 of 28 positions). Acknowledgments. The authors thank the Process Development Group for providing cells for FACS analysis, Eric Johnson for assistance with the figures, and Doug Jacoby for critical reading of the manuscript. REFERENCES 1. Sachs DH, Sykes M, Robson SC, et al. Xenotransplantation. Adv Immunol 2001; 79: 129. 2. LaVecchio J, Dunne, AD, Edge AS. Enzymatic removal of alpha-galactosyl epitopes from porcine endothelial cells diminishes the cytotoxic effect of natural antibodies. Transplantation 1995; 60: 841. 3. Teranishi K, Gollackner B, Buhler L, et al. Depletion of anti-gal antibodies in baboons by intravenous therapy with bovine serum albumin conjugated to gal oligosaccharides. Transplantation 2002; 73: 129. 4. Yamada K, Sachs D, DerSimonian H. Human anti-porcine xenogeneic T cell response. Evidence for allelic specificity of mixed leukocyte reaction and for both direct and indirect pathways of recognition. J Immunol 1995; 155: 5249. 5. Schneider MK, Forte P, Seebach JD. Adhesive interactions between human NK cells and porcine endothelial cells. Scand J Immunol 2001; 54: 70. 6. Donnelly CE, Yatko C, Johnson E, et al. Human natural killer cells account for non-MHC class I restricted cytolysis of porcine cells. Cell Immunol 1997; 175: 171. 7. Holgersson J, Ehrnfelt C, Hauzenberger E, et al. Leukocyte endothelial cell interactions in pig to human organ xenograft rejection. Vet Immunol Immunopathol 2002, 10, 407. 8. Simon AR, Warrens AN, Yazzie NP, et al. Cross-species interaction of porcine and human integrins with their respective ligands: implications for xenogeneic tolerance induction. Transplantation 1998; 66: 385. 9. Warrens AN, Simon AR, Theodore PR, et al. Human-porcine receptor- 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. ligand compatibility within the immune system: relevance for xenotransplantation. Xenotransplantation 1999; 6: 75. DerSimonian H, Pan L, Yatko C, et al. Human anti-porcine T cell response: blocking with anti-class I antibody leads to hyporesponsiveness and a switch in cytokine production. J Immunol 1999; 162: 6993. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227: 680. Sullivan JA, Oettinger HF, Sachs DH, et al. Analysis of polymorphism in porcine MHC class I genes: alterations in signals recognized by human cytotoxic lymphocytes. J Immunol 1997; 159: 2318. Forte P, Pazmany L, Matter-Reissmann UB, et al. HLA-G inhibits rolling adhesion of activated human NK cells on porcine endothelial cells. J Immunol 2001; 167: 6002. Matsunami K, Miyagawa S, Nakai R, et al. The possible use of HLA-G1 and G3 in the inhibition of NK cell-mediated swine endothelial cell lysis. Clin Exp Immunol 2001; 126: 165. Sharland A, Patel A, Lee JH, et al. Genetically modified HLA class I molecules able to inhibit human NK cells without provoking alloreactive CD8⫹ CTLs. J Immunol 2002; 168: 3266. Nikolic B, Cooke DT, Zhao G, et al. Both gamma delta T cells and NK cells inhibit the engraftment of xenogeneic rat bone marrow cells and the induction of xenograft tolerance in mice. J Immunol 2001; 166: 1398. Murray AG, Khodadoust MM, Pober JS, et al. Porcine aortic endothelial cells activate human T cells: direct presentation of MHC antigens and costimulation by ligands for human CD2 and CD28. Immunity 1994; 1: 57. Arulanandam AR, Withka JM, Wyss DF, et al. The CD58 (LFA-3) binding site is a localized and highly charged surface area on the AGFCC’C face of the human CD2 adhesion domain. Proc Natl Acad Sci U S A 1993; 90: 11613. Rollins SA, Kennedy SP, Chodera AJ, et al. Evidence that activation of human T cells by porcine endothelium involves direct recognition of porcine SLA and costimulation by porcine ligands for LFA-1 and CD2. Transplantation 1994; 57: 1709. Brossay A, Hube F, Moreau T, et al. Porcine CD58: cDNA cloning and molecular dissection of the porcine CD58-human CD2 interface. Biochem Biophys Res Commun 2003; 309: 992.