Molecular cloning and functional characterization of the pig homologue of integrin-associated protein (IAP/CD47)

Abstract

We report the cloning of cDNA encoding the pig homologue of human integrin-associated protein (IAP or CD47). A pig CD47-specific probe was generated by polymerase chain reaction (PCR) amplification of pig leucocyte cDNA, using primers based on consensus regions among the known sequences of CD47 from different species. Screening of a pig aorta smooth muscle cDNA library identified seven clones, all containing identical sequences. The clones contained an open reading frame (ORF) that encoded an 18 amino acid putative signal peptide, a 122 amino acid sequence consisting of a single extracellular immunoglobulin variable (IgV)-like domain followed by a 147 amino acid region containing five membrane-spanning domains and a 16 amino acid cytoplasmic tail. The amino acid sequence of the clones was 73% homologous to human IAP and therefore it was termed pig IAP or CD47. Reverse transcription–polymerase chain reaction (RT–PCR) showed that pig CD47 was expressed in a wide range of tissues and detected different alternatively spliced forms. The monoclonal antibody (mAb) BRIC 126, anti-human CD47, was shown, by flow cytometry, to stain pig platelets as well as Chinese hamster ovary (CHO) cells transfected with the cDNA encoding pig CD47. Western blot analysis of pig erythocytes and platelets showed a molecular weight (MW) of 43 000–50 000 and of 55 000–65 000, respectively, under non-reducing conditions. Pig CD47 was stably expressed on CHO cells and shown to bind human thrombospondin (TSP). BRIC126 antibody inhibited the binding of platelets and of CD47-transfected cells to human TSP and to pig fibrinogen, whereas no effect was observed on control CHO cells.

Introduction

Cell adhesion is critical for the genesis and maintenance of both three-dimensional structure and normal function in tissues. Adhesion is required for cell growth, differentiation, survival and function. Integrins are heterodimeric molecules, consisting of α and β subunits which interact non-covalently at the cell surface. They are involved in the adhesion of cells to extracellular matrix proteins, in cell–cell interactions and also serve as signal-transducing receptors.¹

Integrin-associated protein (IAP), also known as CD47 (reviewed in ref. 2), is a human cell-surface glycoprotein of 50 000–55 000 molecular weight (MW) that is associated physically and functionally with β₃ (CD61) integrins.³ Structurally, CD47 is an unique member of the immunoglobulin family, with an N-terminal immunoglobulin variable (IgV) domain, five membrane-spanning domains, and a short, alternatively spliced C-terminal cytoplasmic tail.^3,4 CD47 is identical to OA-3, an ovarian carcinoma antigen, and to an erythrocyte membrane protein decreased in Rh_null disease.^3,5,6

In humans, CD47 has a broad tissue distribution, including haemopoietic cells (such as T cells,⁷ neutrophils,⁸ mast cells,⁹ bone marrow stromal cells and spleen,¹⁰ and red blood cells¹¹) and non-haemopoietic cells (such as mesenchymal cells,¹² epithelial and endothelial cells,⁸ fibroblasts and other cells with particularly strong expression in the brain¹³).

Within the plasma membrane of platelets and most cell types, CD47 can associate with and modulate the activity of several families of integrins. It is reported that ligand engagement of the integrin/CD47 complex can activate heterotrimeric G protein signal transduction.¹⁴ However, CD47 shows ubiquitous tissue and haematopoietic cell distribution, including expression on mature erythrocytes, cells that do not express integrins. This has increased interest in the possibility that CD47 might have integrin-independent functions.

Independent of its association with integrins, the signal regulatory protein of α subtype (SIRPα) has been shown to be a CD47 ligand.¹³ CD47–SIRPα interactions can mediate cell–cell adhesion. SIRPα is a membrane protein highly expressed on macrophages and dendritic cells¹⁵ and it is possible that CD47 has a physiological role in T-cell stimulation.

It has been demonstrated that thrombospondin (TSP) is the biologically relevant ligand for CD47.⁷ On platelets, CD47 is a TSP receptor that activates the fibrinogen-binding integrin αIIbβ₃ and thus plays a special role in platelet stimulation.¹⁶

CD47 knockout mice show a defect in host defence. In particular, granulocytes are deficient in β₃ integrin-dependent ligand binding, cell migration and activation.¹⁷

Interest in the pig as a model for immunological research has arisen mainly owing to its relevance as a worldwide food resource, but also as a result of its physiological similarity with humans. Swine are being used as large animal models for biomedical research and, currently, there is considerable interest of pigs as organ donors in xenotransplantation.^18,19 Recently, some concern has been raised with respect to cross-species effects of adhesion molecules, which might play an important role during the cell-mediated rejection of porcine xenografts.^20,21 The efficacy of many adhesive interactions relevant to xenogeneic organ transplantation still remains to be determined.²² An understanding of the adhesive interactions of these molecules in the pig-to-human context will be critical for the development of genetically engineered porcine donors appropriate for xenotransplantation to humans. We have undertaken the molecular and structural characterization of pig adhesion molecules in the expectation that the knowledge gained will be of relevance to xenotransplantation. We here report the identification and molecular characterization of the pig analogue of human IAP (or CD47). We have used a monoclonal antibody (mAb) to human CD47 to examine CD47 expression in pig blood cells and in transfected Chinese hamster ovary (CHO) cells. We have tested the effect of this mAb on the in vitro adhesion of CD47-expressing cells to either fibrinogen or TSP, from pig and human origin, respectively.

Materials and methods

Molecular biology

All general reagents were from Panreac (Barcelona, Spain) or Sigma Aldrich (Barcelona, Spain) unless otherwise stated. The porcine aorta smooth muscle Uni-ZAP™XR cDNA library was from Stratagene (Madrid, Spain). Taq polymerase, dNTPs and polymerase chain reaction (PCR) buffer were from Biotools (Madrid, Spain). RNA isolation kits, restriction enzymes and buffers were from Roche (Barcelona, Spain). The Geneclean III DNA purification kit was from Q-BIOgene (Heidelberg, Germany). The plasmid purification kit was from Qiagen (Hilden, Germany). The expression vector pDR2ΔEF1α was a gift from Dr I. Anegon (INSERM U437, Nantes, France) and contains the hygromycin resistance gene (allowing the selection of stable colonies) and the powerful chain elongation factor 1α promoter.

Tissues and cells

All tissues used for cDNA preparations were obtained fresh from the local abattoir. Fresh pig blood was obtained from the local abattoir, collected into 0·38% sodium citrate as anticoagulant, and used as a source of pig leucocytes, granulocytes, platelets and erythrocytes. Peripheral blood mononuclear cells (PBMC) were isolated from pig blood by density-gradient centrifugation on Ficoll–Paque (Pharmacia, Uppsala, Sweden). Granulocytes were recovered from the lower Ficoll phase and residual erythrocytes were lysed by hypotonic treatment. Platelet-rich plasma (PRP) was obtained from pig blood by centrifugation at 750 g for 10 min and replaced with an equal volume of phosphate-buffered saline (PBS). Porcine alveolar macrophages were collected by bronchoalveolar lavage as described previously.²³ Porcine platelets were obtained by centrifugation of the PRP at 1500 g for 10 min. CHO cells were a kind gift of Dr F. Almazán (Centro Nacional de Biotecnologı´a, Madrid, Spain).

Proteins and antibodies

Human TSP and fibrinogen from pig, human, sheep and dog plasma were from Sigma. BRIC-126 mAb (IgG2b) to human CD47 was purchased from Serotec (Oxford, UK). Goat anti-mouse immunoglobulin G (IgG)–horseradish peroxidase (HRP) and goat anti-mouse IgG–fluorescein isothiocyanate (FITC) conjugates were purchased from Sigma. Mouse anti-trout IgG INIA3B10 (IgG2b) was purchased from BV-UCO (Córdoba, Spain) and was used as an isotype-matched control to BRIC126.

Screening of a pig cDNA library

A 610-bp probe was generated by PCR from pig PBMC cDNA using two oligonucleotides: F20 (5′-CGGCGGGCGCGGAGATGT-3′) and R20 (5′-TCACCTGGGACGAAAAGAATGG-3′) based on the consensus regions among the sequences of human, mouse and rat CD47 (GenBank accession numbers: NM001777, NM010581 and D87659, respectively). The 610-bp PCR product was labelled with the Digoxigenin (Dig) system (Roche) according to the manufacturer’s protocol. PCR reactions were performed in a final volume of 50 µl containing 75 mm Tris–HCl (pH 9·0), 50 mm KCl, 20 mm (NH₄)₂SO₄, 0·001% bovine serum albumin (BSA), 0·05 mm each of Dig-labelled dATP, dCTP, dGTP and dTTP, 2 mm MgCl₂, 1 µm oligonucleotide primer and 1·25 U of Taq DNA polymerase (Biotools); 2 µl of cDNA from pig PBMC was subjected to 35 amplification cycles of 30 seconds at 95°, 30 seconds at 56° and 30 seconds at 72°. The labelled probe was purified and used to screen 500 000 plaque colonies of the porcine aortic smooth muscle cDNA library plated at 50 000 plaque forming units (PFU) per plate and grown on a lawn of XL1-Blue Escherichia coli for 8 hr. Lifts were taken onto Nytran-nylon membranes (Schleicher & Schuell, Dassel W, Germany), denatured in 1·5 m NaCl/0·5 m NaOH, neutralized in 1·5 m NaCl/0·5 m Tris (pH 8·0) and fixed by baking for 2 hr at 80°. Hybridization of the membranes with Dig-labelled probe and detection were carried out using the Dig detection kit (Roche) following the recommendations of the manufacturer. Positive plaques on membranes were identified, isolated in agar plugs, eluted in 1 ml of SM buffer [5.3g NaCl, 2g MgSO₄, 7H₂O, 50 ml 1 m Tris–Hcl (pH 7.5), 0.01% (w/v) gelatin, H₂O to 1 l] for 24 hr at 4° and replated. The above screening protocol was then repeated. Individual positive plaques from the secondary screening were isolated in agar plugs and eluted in SM buffer. The cDNA inserts were recovered from PCR screen-positive colonies using the Exassist/SOLR system (Stratagene). Individual bacterial colonies containing recombinant phagemid were cultured in LB broth (10g tryptone, 5g yeast extract, 5g NaCl, 1 ml 1 n NaOH, H₂O to 1 liter containing 50 µg/ml ampicillin, and phagemid DNA was purified using a QIAprep spin plasmid mini-prep kit (Qiagen) and sequenced. Sequences were analysed using the analysis software from the expasy web site (http://www.expasy.ch/).

DNA sequencing

Sequencing was performed using ABI PRISM Terminator Cycle Sequencing Kit (Applied Biosystems PE Hispania SA, Madrid, Spain). Sequencing reactions were carried out on a thermal DNA cycler GeneAmp PCR System 2400 (Applied Biosystems PE Hispania SA, Madrid, Spain), according to the instructions of the manufacturer, and analysed on an ABI PRISM 310 Sequencer (PE Applied Biosystems).

Reverse transcription–PCR analysis

Total RNA (10 µg) was extracted from lymphoid (spleen, thymus, PBMC, platelets, bone marrow, alveolar macrophages) and non-lymphoid (kidney and liver) pig tissues, using the Tri-Pure isolation reagent (Roche). The RNA was reverse transcribed using random hexamer DNA primers by incubation with 200 U of RNAse H⁻ reverse transcriptase (Gibco BRL, Barcelona, Spain) at 25° for 10 min, then at 42° for 90 min in the presence of 50 mm Tris–HCl, 75 mm KCl, 3 mm MgCl₂, 10 mm dithiothreitol (DTT), 30 U of RNAse inhibitor and 1 mm dNTPs, in a total volume of 30 µl.

The presence of cDNA encoding the pig CD47 was determined by PCR using primers F20 and R20 to amplify the most conserved coding sequences. The detection of alternatively spliced forms of porcine CD47 was performed by PCR using primers F18 (5′-CAATTTTGGCTATACTTCTGTTC-3′) and R18 (5′-GTCAAGAAACCCCAGGAT-3′) to amplify the 3′ terminal region of CD47. Cycling conditions were: 95° for 30 seconds, 56° for 30 seconds and 72° for 30 seconds, for 35 cycles. The PCR products were run on a 1% agarose gel. As an internal control, reverse transcription (RT)–PCR was performed on the same cDNAs using the primers P1 (5′-AAAGGATCCGACTCAACACGGGAAACCTCAC-3′) and P2 (5′-AAAGGATCCGCTTATGACCCGCACTTACTGG-3′), which amplify a fragment of 420 bp from the conserved home gene 18S RNA. The annealing temperature was 55°.

Construction of eukaryotic expression vector for pig CD47

The eukaryotic expression vector pDR2ΔEF1α carries the SV40 replication origin and contains the hygromycin resistance gene (allowing the selection of stable colonies) and the powerful chain elongation factor 1α promoter.²⁴

From the sequence of the pig CD47 clone, two primers, FBamHCHO (5′-GGTCGGATCCGGCGGACGCGGAGATGA-3′) and REcoVCHO (5′-GGCGATATCGTCAAGAAACCCCAGGAT-3′), were designed for PCR amplification of the full-length open reading frame (ORF) of pig CD47. The primers FBamHCHO and REcoVCHO contained BamHI and EcoRV restriction sites, respectively. These sites were also present as unique sites in the cloning region of the pDR2ΔEF1α expression vector, ensuring correct orientation of the insert. To ensure fidelity, PCR was performed using platinum pfx-DNA polymerase (Gibco BRL) that has proofreading capacity. PCR product and vector were digested with BamHI and EcoRV before ligation. The ligated construct was transformed into DH5α and colonies were picked and the plasmids purified using the QIAprep spin plasmid kit (Qiagen). Before transfection, the fidelity and orientation of pig CD47 cDNA in the vector was confirmed by sequencing. CHO cells were lipofectamine transfected with the empty expression vector as a negative control. Surface expression of pig CD47 was demonstrated by flow cytometry as described below. Expression was also confirmed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and Western blotting of cell lysates of both pig erythrocytes and platelets using the anti-CD47 mAb BRIC-126.

Western blotting

Pig platelets and erythrocytes were obtained as described above. Membranes were solubilized in lysis buffer, consisting of 10 mm Tris–HCl, 150 mm NaCl, 1 mm EDTA, 1% Nonidet P-40 (NP-40), 1 mm phenylmethylsulphonyl fluoride (PMSF), for 1 hr at room temperature. After centrifugation at 12 000×g for 15 min, supernatants were mixed with sample buffer and run on a 10% SDS–PAGE gel under non-reducing conditions. The separated proteins were transferred to nitrocellulose membrane (Schleicher & Schuell). Nitrocellulose was blocked with PBS containing 3% BSA and incubated with the mAb BRIC126, or with INIA3B10, as negative control, for 1 hr at room temperature, followed by a 1-hr incubation with a peroxidase-labelled rabbit anti-mouse immunoglobulin (Sigma). Peroxidase activity was visualized using SuperSignal (Pierce & Warriner, Chester, UK) as a substrate, following the recommendations of the manufacturer.

Transfection of pig CD47 into CHO cells

CHO cells were grown and maintained as a monolayer in Dulbecco’s modified Eagle’s minimal essential medium (DMEM) F-12 medium (Gibco BRL) supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin and 10% fetal calf serum (FCS) (Gibco BRL). Cultures were incubated at 37° in a CO₂ incubator containing 5% CO₂/95% air at an initial density of 1×10⁶ cells/flask, in 25-cm² tissue culture flasks. CHO cells were grown to 50% confluence and transfected with 8 µg of plasmid DNA using the Lipofectamine plus reagent (Gibco BRL). Cells were grown for 24 hr in non-selective medium. CHO cells that had been transfected with pig CD47 construct were selected by adding hygromycin B (Gibco BRL), at a concentration of 400 µg/ml in DMEM F-12 medium. Selection medium was changed every 4 days for ≈2 weeks, by which time all of the non-transfected control cells had died. The pool of survivors was analysed using flow cytometry.

Flow cytometry

Pig cells or harvested CHO cells (10⁶ cells/ml in PBS) were incubated on ice with mAb BRIC126 (5 µg/ml) or INIA3B10, as negative control, for 30 min. Cells were washed three times in flow cytometry buffer [FCB: PBS, 0·1% bovine serum albumin (BSA) and 0·1% NaN₃]. Cells were incubated on ice for 30 min with 50 µl of FITC-conjugated goat anti-mouse IgG (Sigma), diluted 1 : 500 in FCB. Cells were washed three times in FCB buffer. Fluorescence was analysed using a Becton-Dickinson FACSort™ (San Jose, CA, USA) equipped with the CellQuest™ software. Cells were identified and gated by their characteristic forward and side scatter. Platelet acquisition was performed on a logarithmic scale and with the pressure selector in the low position.

Adhesion assay

Adhesion of pig platelets and of transfected CHO cells to ligand-coated surfaces was performed as described previously.²⁵ Briefly, wells of a 96-well microtitre plate were coated with up to 10 µg/ml of human TSP, or with 4 µg/ml of fibrinogen from pig, human, sheep and dog plasma, in 100 µl of PBS and incubated at 4° overnight. Non-adherent protein was removed by washing three times with PBS. Wells were blocked for 90 min at room temperature with PBS containing 1% BSA (Roche). Control wells were coated with 1% BSA to determine background adhesion. CHO cells and CHO pig CD47-transfected cells (CHO-pCD47) were washed twice with PBS and resuspended in DMEM F-12 medium at a concentration of 1×10⁶ cells/ml. Platelets were purified as described previously, washed in PBS and resuspended in PBS at a concentration of 1×10⁸ cells/ml. After washing, 0·5 mm MnCl₂ was added to the cell suspension to induce a high-affinity state of fibrinogen ligands.²⁵ Aliquots (100 µl) of CHO, CHO-CD47, and platelet suspension were added to wells in triplicate. The plate was incubated in a humidified 37° incubator for 90 min. After washing with PBS, the adherent cells were checked by visual inspection, and adhesion was quantified with a colorimetric reaction using endogenous cellular acid phosphatase (AP-ase) activity, as previously described.²⁶ Briefly, 100 µl of the substrate/lysis solution [1% Triton-X-100, 6 mg/ml p-nitrophenyl-phosphate (Roche) in 50 mm sodium acetate buffer, pH 5·0] was added to each well. After 1 hr of incubation at 37°, the reaction was terminated by the addition of 50 µl of 1 m NaOH, and AP-ase activiy was determined in an enzyme-linked immunosorbent assay (ELISA) plate reader (model 550; Bio-Rad Laboratories SA, Madrid, Spain) with a 415-nm filter. Background values were subtracted from the results obtained for each test well.

For the antibody-inhibition studies, the adhesion assay was performed on surfaces coated with either TSP or with pig fibrinogen (4 µg/ml), and with cells and platelets previously incubated with 20 µg/ml of the mAb BRIC-126, or INIA3B10 as a negative control, in PBS for 30 min at room temperature.

RESULTS

Isolation and characterization of pig CD47 cDNA

Based on the known sequences of CD47 from different species, two oligonucleotides (F20 and R20) were synthesized and used for PCR amplification of a CD47-specific probe using pig PBMC cDNA as template. Screening of a porcine smooth muscle cDNA library identified seven positive clones, all from separate plates, from a total of 4·2×10⁵ PFU screened. All clones were isolated and sequenced in their entirety. All contained identical sequences of 2759 base pairs (bp) within which was a region of 121 bp 5′ untranslated (UTR) flanking region, a single ORF of 912 bp encoding a polypeptide of 303 amino acids, and a 1726-bp 3′-UTR flanking region (Fig. 1).

Figure 1.

Nucleotide and deduced amino acid sequence of pig CD47. The 2759-bp sequence obtained from library screening is shown. The initiation codon (atg) is underlined. Primers used in reverse transcription–polymerase chain reaction (RT–PCR) are below arrows and named. The polyadenylation site in the 3′ untranslated sequence (aataaa) is underlined. The first amino acid of the mature protein (Q) is in bold. Numbering is for the mature protein sequence. Potential N-glycosylation (NXS/T) and O-glycosylation sites are indicated by Ψ and Ω, respectively. The predicted transmembrane segments are underlined. GenBank accession number: AF332698.

The clone encoded an 18 amino acid putative signal peptide, a 122 amino acid sequence consisting of a single extracellular IgV-like domain, followed by a 147 amino acid region containing five membrane-spanning domains (Fig. 1) and a 16 amino acid cytoplasmic tail (Fig. 1). The sequence of the entire clone was submitted to GenBank (Fig. 1; accession number: AF332698). Comparison of the deduced amino acid sequence of the clone with the sequence of human CD47 revealed an amino acid identity of 73% (Fig. 2). The highest homology (83%) was found for Bos taurusCD47. A homology of 71% was found for rat CD47, whereas mouse CD47 showed a homology of 66% with the pig sequence. These data strongly suggested that the cDNA isolated was a pig homologue of IAP or CD47 and it was hereafter termed pig CD47.

Figure 2.

Alignment of amino acid sequences of pig CD47 from different species. The amino acid sequence of pig CD47 was aligned with the homologous sequence in human (GenBank accession number: NP 001768), mouse (GenBank accession number: NP 034711), rat (GenBank accession number: BAA 13420) and Bos taurus (GenBank accession number: CAB 76905). Numbering is from the mature protein sequence. Aligned sequences are in capital letters. Non-aligned sequences are in lower case. Potential N-glycosylation sites are in bold.

The pig homologue of CD47 contained four potential N-glycosylation sites, all of which were located at the IgV domain (N5, N16, N32 and N91) (Fig. 1). The glycosylation site located at N5 is conserved in human CD47, but not in mouse, rat or B. taurus CD47, whereas the N32 glycosylation site is present in human and Bos taurus CD47, but is absent from mouse and rat CD47 (Fig. 2). The disulphide linkage suggested by the homology with human, mouse, B. taurus and rat CD47 is located at residues C23 and C94 of pig CD47 (Fig. 2).

Analysis of the pig CD47 sequence by using the NetOGlyc 2·0 software (at expasy web site) revealed one O-glycosylation site at S84. However, although conserved, this site was not predicted for human, mouse, rat and B. taurus CD47.

Analysis of pig CD47 expression in tissues

To ascertain whether pig CD47 mRNA was expressed in other tissues, RT–PCR was performed on mRNA from various tissues using primers specific for pig CD47. The predicted length of the product for pig CD47 was 610 bp. Products of the expected size were generated in each of the eight tissues studied (PBMC, thymus, macrophages, platelets, spleen, bone marrow, liver and kidney). RT–PCR (Fig. 3a) demonstrated that mRNA encoding pig CD47 was widely distributed, present in all tissues tested, albeit weakly in non-lymphoid tissues, such as liver (Fig. 3a). RT–PCR using 18S RNA-specific primers generated specific bands of similar intensity in all tissues (Fig. 3a). All controls for non-specific amplification were negative (results not shown).

Figure 3.

Reverse transcription–polymerase chain reaction (RT–PCR) analysis of pig CD47 in tissues. Primer pair F20 and R20 were chosen to specifically amplify the extracellular portion of pig CD47. Primers P1 and P2 were chosen to specifically amplify a fragment of 420 bp from the conserved home gene 18S RNA. Primer pair F18 and R18 were chosen to amplify the C-terminal portion of pig CD47. Equal volumes of all PCR products were loaded onto an agarose gel and separated as described in the Materials and methods. (a) The products of 610 bp (pig CD47) and 420 bp (18S RNA) obtained are apparent in the gel images. (b) When using primers F18 and R18, three different bands are evident, suggesting the existence of at least three different isoforms of pig CD47 in the cells and tissues tested. PBMC, peripheral blood mononuclear cells. Size markers are shown on the right.

Detection of pig CD47 isoforms

To determine the presence of alternatively spliced forms of pig CD47, RT–PCR was performed on mRNA from various tissues using primers which bracketed the region of the mRNA encoding the region of the mRNA encoding the alternatively spliced cytoplasmic tails. Using this PCR strategy, three different forms could be distinguished on the basis of the size of the PCR product (Fig. 3b). All tissues tested expressed CD47 mRNA. However, the different forms were expressed at varying levels in different tissues (Fig. 3b). A major band of 800 bp, corresponding to form 2 of human CD47, was detected in all tissues tested. Three visible bands, corresponding to forms 1 and 3 of human and mouse CD47 were detected in kidney, platelets, thymus and PBMC, whereas in liver and bone marrow a single band was obtained, corresponding to form 2 of human and mouse CD47 (Fig. 3b).

Surface expression of pig CD47 on blood cells

The mAb BRIC 126 (anti-human CD47) was shown, by flow cytometry, to stain pig erythrocytes as well as CHO cells transfected with cDNA encoding pig CD47, thus confirming the cross-reactivity of the antibody with the pig homologue of human IAP. Therefore, we used this mAb to analyse the expression of pig CD47 on circulating blood cells. Flow cytometry analysis revealed the presence of pig CD47 on all circulating cells (Fig. 4). Erythrocytes, granulocytes and platelets all strongly and uniformly expressed pig CD47. The highest expression level was found for platelets (Fig. 4). However, unfractionated PBMC gave a broad staining distribution for pig CD47 (Fig. 4), probably as a result of the different level of expression of CD47 on all leucocyte populations. Isotype-matched control antibody INIA3B10 was negative on all cell types.

Figure 4.

Flow cytometric analysis of pig CD47 expression on blood cells. Surface expression on purified erythrocytes, granulocytes, platelets or unfractionated peripheral blood mononuclear cells (PBMC) was measured by flow cytometry, using the monoclonal antibody (mAb) BRIC126. Cells were stained as described in the Materials and methods. Controls in which primary antibody was omitted are indicated by dotted lines.

Expression of pig CD47 on CHO cells

Stable populations of CHO cells expressing pig CD47 or the empty pDR2ΔEF1α vector were generated. The vector pDR2ΔEF1α was chosen because it is reported to give reproducibly high levels of expression in a given cell type.²⁴ Expression was assessed by flow cytometry, as described above, using the mAb BRIC126 (anti-human CD47). Uniformly high levels of pig CD47 expression were obtained for transfected CHO cells (Fig. 5). The mAb BRIC126 was negative on control CHO cells transfected with the empty vector. The level of expression of pig CD47 in the transfected CHO cells was similar to that obtained for endogenous CD47 on pig platelets, as assessed by flow cytometry (Fig. 4).

Figure 5.

Flow cytometric analysis of pig CD47-transfected cells. Surface expression of pig CD47 on transfected Chinese hamster ovary (CHO) cells was measured by flow cytometry, using the monoclonal antibody (mAb) BRIC126 or INIA3B10, as a negative control. Cells were stained as described in the Materials and methods. CHO cells transfected with the empty expression vector were not stained with the mAb BRIC126.

Pig erythrocytes and platelets were also analysed for CD47 expression by SDS–PAGE under non-reducing conditions and Western blotting (Fig. 6). From erythrocyte lysates, pig CD47 ran as a broad band of approximately 43 000–50 000 MW. The mAb BRIC126 was unable to recognize pig CD47 following reduction (data not shown). Western blotting of cell lysates from platelets under non-reducing conditions identified a band of ≈55 000–65 000 MW. Blots using INIA3B10, an isotype-matched control antibody, were negative (not shown).

Figure 6.

Western blotting of pig CD47. Cell lysates were prepared from erythrocytes and platelets. Lysates were run on sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) under non-reducing conditions, blotted to nitrocellulose and probed with the monoclonal antibody (mAb) BRIC126. Blots were developed using a chemiluminescence detection system. Molecular weight markers are shown on the right. Control blots with an isotype-matched irrelevant antibody showed no binding.

Adhesion of pig CD47 expressing cells to TSP

In order to test the functional properties of the cloned pig IAP/CD47, we examined the adhesion of pig CD47-transfected CHO cells to surfaces coated with its ligand TSP. Adhesion of pig platelets, which express endogenous CD47, was also examined as an internal positive control.

Adhesion of CHO cells and platelets was assessed by measurement of the AP-ase activity following adhesion to the TSP-coated surface. AP-ase activity was determined as a colorimetric reaction and displayed in absorbance units at 415 nm (A₄₁₅). At increasing concentrations of TSP, a high level of AP-ase (from adherent cells) was detected on pig platelets and on CHO cells expressing pig CD47 cDNA, whereas no modification in the AP-ase activity was detected on vector control CHO cells, indicating that the concentration of TSP bound to the surface had little or no effect on the adhesion of vector control cells (Fig. 7). By visual inspection, no significant adhesion of cells to BSA-coated surfaces was observed (data not shown) and the A₄₁₅ from these samples was considered as background. This result suggests that the observed adhesion of platelets and of CD47-expressing cells to the surface is specifically mediated by the concentration of TSP. In this experiment, cells were considered to be adherent to the surface when the absorbance was >1 unit. When the absorbance of cells was ≤0·5 units, cells were considered non-adherent (Fig. 7). Therefore, the minimum concentration of TSP necessary for the adhesion of the cells and platelets was experimentally found to be 4 µg/ml (Fig. 7) and this concentration was used thereafter for the study of the effect of the mAb BRIC126 on the adhesion of pig CD47-expressing cells. In this assay, adhesion of pig platelets expressing endogenous CD47 did not differ significantly from pig CD47-transfected cells, which is consistent with the level of expression observed by flow cytometry (Fig. 4, Fig. 5).

Figure 7.

Adhesion of CD47-expressing cells to thrombospondin (TSP)-coated surfaces. Pig platelets (▴), Chinese hamster ovary (CHO) cells expressing pig CD47 (▪), or vector control cells (•), were added to wells coated with different concentrations of TSP (x-axis). After washing, adhesion was quantified by endogenous acid phosphatase (AP-ase) activity of remaining cells using p-nitrophenyl phosphate as a substrate. AP-ase was determined as a colorimetric reaction, measured in a plate reader, and displayed in absorbance units at 415 nm (A₄₁₅) (y-axis). Results represent the mean value of triplicate determinations ±SE. Cells were considered adherent when the A₄₁₅ was approximately 1 unit or higher.

Adhesion of pig CD47-expressing cells to fibrinogen

We next examined the adhesion of pig CD47-transfected CHO cells to surfaces coated with fibrinogen from different species. Adhesion of pig platelets was also examined as an internal positive control. Cells were considered to be adherent to the surface when the absorbance was higher than 1 unit. When the absorbance of cells was ≤0·5 units, cells were considered non-adherent. Fibrinogen from pig, human, sheep or dog plasma when bound to the surface had little or no effect on the adhesion of vector control cells (Fig. 8). However, both pig platelets and CHO cells expressing pig CD47 showed adhesion to surfaces coated with fibrinogen from pig and dog plasma, whereas no effect was found when the fibrinogen was from human or sheep origin (Fig. 8). These results suggest the conservation of integrin-binding regions between pig and dog fibrinogen.

Figure 8.

Adhesion of CD47-expressing cells to fibrinogen from different species. Pig platelets, Chinese hamster ovary (CHO) cells expressing pig CD47 and CHO cells transfected with the empty expression vector were added to wells coated with fibrinogen from pig, human, sheep and dog plasma. After removal of the non-adherent cells, adhesion was quantified by endogenous acid phosphatase (AP-ase) activity of remaining cells using p-nitrophenyl phosphate as a substrate. AP-ase was determined as a colorimetric reaction, measured in a plate reader, and displayed in absorbance units at 415 nm (A₄₁₅) (y-axis). Results represent the mean value of triplicate determinations ±SE. Cells were considered adherent when the A₄₁₅ was approximately 1 unit or higher.

Effect of BRIC126 on the adhesion of pig CD47-expressing cells

To further explore the functional properties of the cloned pig CD47, we next investigated whether the anti-CD47 mAb might affect the adhesion of pig CD47-expressing cells to surfaces coated with either TSP or fibrinogen from pig plasma. In CHO cells, it is reported that hamster CD47 is detected with a polyclonal anti-CD47 antibody.³ Western blot and flow cytometry analysis revealed that BRIC126 mAb neither recognized untransfected cells nor control CHO cells transfected with the empty expression vector (Fig. 5). Therefore, in principle, the function of the endogenous hamster CD47 from CHO cells should not be affected by the mAb BRIC126. The effect of BRIC126 on transfected CHO cells containing pig CD47 was also conducted in order to confirm that the observed adhesion of CHO cells and platelets to coated surfaces was pig CD47-mediated. The adhesion of pig platelets was also examined as an internal control. Based on preliminary results, cells were considered adherent to the surface when the AP-ase activity of remaining cells was close to 1 unit of A₄₁₅ or higher.

Incubation of pig platelets and of CHO cells with a saturating concentration of the mAb BRIC126 markedly reversed the adhesion of pig platelets and of CD47-transfected cells to the coated surfaces, whereas no effect was observed for the adhesion of control CHO cells transfected with the empty expression vector (Fig. 9a, 9b). Incubation of cells with similar concentrations of INIA3B10, an irrelevant isotype-matched mAb, did not affect the adhesion of cells (Fig. 9a, 9b). Thus, the mAb BRIC126 efficiently inhibited the adhesion of pig CD47-expressing cells to TSP and to fibrinogen. These data support the results obtained previously and indicate that the observed adhesion of platelets and CHO cells to the coated surfaces was specifically mediated by the expression of pig CD47 on the surface of pig platelets and of CHO cells.

Figure 9.

Effect of BRIC126 on the adhesion of CD47-expressing cells. Pig platelets, Chinese hamster ovary (CHO) cells expressing pig CD47 and CHO cells transfected with the empty expression vector, previously incubated with either monoclonal antibody (mAb) BRIC126 or mAb INIA3B10, as a negative control, were added to wells coated with human thrombospondin (TSP) (a) or with fibrinogen from pig plasma (b). After removal of the non-adherent cells, adhesion was quantified by endogenous acid phosphatase (AP-ase) activity of remaining cells using p-nitrophenyl phosphate as a substrate. AP-ase was determined as a colorimetric reaction, measured in a plate reader, and displayed in absorbance units at 415 nm (A₄₁₅) (y-axis). Results represent the mean value of triplicate determinations ±SE. Cells were considered adherent when the A₄₁₅ was approximately 1 unit or higher.

Discussion

Adhesion molecules in the pig has become a subject of interest because of the planned use of pig organs for transplantation to humans (xenotransplantation). The importance of the adhesive interactions in xenotransplantation is clear considering the possible consequences of their failure.²² It is well known that an adhesion molecule interaction might not only play a role in adhesion, but might also actively transmit signals to stimulate the host immune response to xenogeneic organs in vivo.²⁷ Blocking of these interactions, by antibodies, might help to treat and prevent cellular rejection in pig-to-human xenotransplantation.²⁸ However, many of the interactions of porcine homologues with human molecules have not yet been identified and the efficacy of their adhesive interactions still remains to be determined.²² The recent discovery of a porcine cell-surface receptor as a member of the SIRP molecules in pigs,²⁹ together with the widespread potential for CD47 modulation of integrin function, led us to search for the pig homologue of human IAP (CD47).

Human CD47 is expressed ubiquitously in all cells and tissues. In order to clone the pig homologue of CD47, we used a pig smooth-muscle cDNA library. The CD47 probe was generated by PCR amplification of cDNA from pig PBMC using primers based on consensus regions among the sequences of human, mouse and rat CD47. Seven clones were identified from different plates all containing identical sequences. All clones contained an ORF of 912 bp encoding a polypeptide of 303 amino acids. The Kozak sequence (^A/_GNNATG), recognized by ribosomes as the translational start site and thus required for protein expression, conformed poorly to the sequence found within the CD47 5′-UTR. However, this is probably the initiation site, based on the absence of a preceding initiation codon in any of the clones. In fact, it is reported that human CD47 also conforms poorly to the Kozak consensus.³ The sequence of the predicted polypeptide was highly homologous (about 70% overall identity) with the published sequence of human, mouse, rat and B. taurus CD47, suggesting that it was the pig homologue of CD47. In addition, this sequence was predicted to contain an IgV-like domain, five membrane-spanning regions and a cytoplasmic tail, resembling the structure of a CD47 molecule.

RT–PCR analysis of RNA extracted from pig tissues was performed using primers F20 and R20, capable of amplifying the possible conserved sequence of pig CD47 and previous to the common exon for splicing.⁴ These studies demonstrated that CD47 was broadly expressed in all tissues examined, although the level of expression was lower in non-lymphoid tissues, such as in liver. Southern blot analysis of RT–PCR products confirmed that the band of 616 bp was CD47-specific (results not shown).

CD47 exists as four isoforms, differing in the tail coding region.² These alternatively spliced forms are generated by inclusion or exclusion of three short exons within 5 kb in the genome and are highly conserved between human and mouse CD47. There is tissue specificity of expression of the alternatively spliced forms of CD47 mRNA transcribed from a single gene. The sequence isolated from the pig cDNA clone corresponded with the isoform 2 of human CD47; however, no other form was found in any of the seven clones sequenced, suggesting that this particular single form must be predominant in smooth muscle cells of the pig. Moreover, form 2 of human CD47 is the most abundant form in bone marrow-derived cells and endothelia, particularly thymus and spleen, while form 4 is highly expressed in cells of the brain and peripheral nervous system.⁴ All four forms of CD47 mRNA are found in both murine and human cells, with a very high interspecies conservation with respect to both amino acid sequence of the alternatively spliced regions and gene structure.⁴ Evidence for the existence of similar isoforms in the pig was also obtained by RT–PCR amplification of cDNA from different pig tissues. RT–PCR amplification of pig CD47 cDNA showed three bands of different sizes, corresponding to different isoforms (Fig. 3). Our results are consistent with the in vivo expression of the alternatively spliced forms of CD47 in human and rodents. However, in the tissues tested, we were unable to detect the band corresponding to the form 4 mRNA of human CD47, which is predominantly expressed by neural tissue.⁴ The preservation of the various isoforms and their similar tissue-restricted distribution suggests that these cytoplasmic tails play an essential role in CD47 function. Probably, the reason for this tissue-specific alternative splicing is that CD47 might have different functions in different tissues, depending on the nature of its cytoplasmic tail. However, to date, there are no data on the function of the different CD47 cytoplasmic tails.

Flow cytometric analyses using the mAb BRIC126 confirmed the abundant expression of pig CD47 on circulating cells (Fig. 4). As in humans, pig CD47 was broadly expressed on all circulating cells, and was particularly abundant on platelets. CHO cells stably expressing pig CD47 showed a homogeneous population of high-expressing cells, as shown by flow cytometry using the mAb BRIC126 (Fig. 5). This homogeneity of expression is mediated by the elongation factor 1α promoter in the pDRΔEF1α expression vector, which has previously been shown to vary little in its expression levels in a given cell.³⁰

Human CD47 has five potential N-glycosylation sites in the extracellular domain; three of these sites are glycosylated in erythrocytes, leading to its broad migration at 45 000–60 000 MW on SDS–PAGE. However, in leucocytes and platelets, CD47 ran as a band of 47 000–55 000 MW.³¹ To examine CD47 expression on pig erythrocytes and platelets, membranes were solubilized and analysed by Western blot using the mAb BRIC126. Consistent with the established MW of CD47, BRIC126 readily detected a band of ≈43 000–50 000 on erythrocytes (Fig. 6). However, on platelets, pig CD47 ran as a band of 55 000–65 000 MW (Fig. 6). These results suggest that, as in humans, type-specific cell glycosylation may be responsible for the differences in the apparent molecular weight of pig CD47.

CD47 binds plasma and extracellular matrix glycoprotein TSP.³² TSP is a matrix glycoprotein, reviewed in ref. 33, which was first described as an α-granular protein secreted by platelets upon thrombin activation.³⁴ We utilized CHO cells stably expressing pig CD47 to examine the adhesion of CHO cells and platelets to surfaces coated with different concentrations of TSP. Pig platelets and CD47-transfected cells adhered to the coated surfaces in a TSP concentration-dependent manner (Fig. 7). We next tested the hypothesis that pig CD47 is an adhesion receptor for TSP on platelets and on CD47-transfected CHO cells. Using the mAb BRIC126 to selectively block any CD47–TSP interaction, we analysed platelet and CHO cell adhesion to TSP. Preincubation of platelets and CHO cells with saturating concentrations of BRIC126 efficiently blocked the adhesion of pig CD47-transfected cells and platelets (Fig. 9a), whereas no effect was observed on cells and platelets preincubated with an isotype-matched control mAb. Usually, CD47-inhibition studies have been performed with the blocking mAb B6H12. The mAb B6H12 specifically interrupts a CD47–TSP interaction.³⁵ However, BRIC126 and B6H12 probably recognize similar or overlapping epitopes on CD47.³⁵ It has been shown that the function-blocking effect of B6H12 and BRIC126 antibodies is not a general consequence of an antibody binding to the surface of the platelets, or even to the CD47 molecule, because 2D3, a non-functional blocking antibody that binds to CD47, failed to block platelet adhesion to TSP at any concentration tested.³⁵ Our results suggest that BRIC126 also recognizes functionally important regions of pig CD47 and that CD47 mediates platelet adhesion to TSP, thus implicating CD47 as an adhesion receptor on pig platelets.

CD47 was first discovered as a protein physically and functionally associated with the integrin α_vβ₃. Human CD47 co-precipitates with α_vβ₃ from leucocytes and placenta.³⁶ Further studies demonstrated that CD47 also associated, in the same plasma membrane, with additional integrins. On platelets, CD47 interacts with α_vβ₃, α_IIbβ₃ and α₂β₁ to form a complex possessing signalling properties.² Ligation of the integrin–CD47 complex on a variety of cells can induce adhesion, chemotaxis, spreading, secretion and other biological effects of cell activation.^35,37 We have previously reported the characterization of the porcine homologue of α_IIbβ₃ integrin using a mAb³⁸ and have developed a mAb to its ligand, fibrinogen.³⁹ In the third International Workshop on Swine Leucocyte Differentiation Antigens, we observed that an antibody, submitted as anti-CD47, was able to recognize the fraction of α_IIbβ₃, purified using our specific mAb JM2E5. This suggested that CD47 was co-purified with α_ΙΙββ₃ and that a physical interaction of CD47 and α_IIbβ₃ must exist in pig platelets.⁴⁰

Antibodies to CD47 have been shown to inhibit ligand binding to α_IIb/β₃ on several cell lines and to block the increase in [Ca²⁺], which occurs upon endothelial cell adhesion to fibronectin- or vitronectin-coated surfaces, without disturbing cell adhesion to these surfaces.⁴¹ In order to elucidate the role of pig CD47 on its interaction with the porcine integrins α_IIbβ₃ and/or α_vβ₃, we analysed the effect of the function-blocking mAb BRIC126 on the adhesion of pig platelets and on transfected cells to fibrinogen-coated surfaces. We first tested the adhesion of transfected CHO cells and platelets to coated surfaces with fibrinogen from different species. Only fibrinogen from pig or dog plasma efficiently bound to these cells (Fig. 8). Therefore, pig fibrinogen was used thereafter to analyse the effect of the mAb BRIC126.

The adhesion of transfected CHO cells and platelets was specifically mediated by CD47, as judged by the effects of the function-blocking mAb BRIC126 (Fig. 9b). However, although CD47 is the primary TSP receptor on resting platelets, there is no evidence of fibrinogen as a ligand for CD47. The observed adhesion to the fibrinogen-coated surfaces would then be a consequence of the CD47-mediated enhancement of the avidity of the integrins for fibrinogen. Despite the fact that we found similar levels of surface expression of pig CD47 on platelets and on transfected cells (Figs 4 and 5), the fibrinogen adhesion and the inhibition effect of the mAb BRIC126 were more dramatic on pig platelets than on transfected CHO cells. We propose that, upon platelet interaction with the fibrinogen-coated surface, TSP might be secreted from activated platelets and strengthen ligand binding of α_IIbβ₃ or α_vβ₃, resulting in the increased adhesion of platelets to fibrinogen. This adhesion does not occur when CD47–TSP interaction is blocked by BRIC126. Actually, CD47 ligation to TSP has a role in modulating α_vβ₃ and α_IIbβ₃ function, such as fibrinogen binding and platelet cross-linking.^32,42

Recently, it has been demonstrated that, independently of its association with integrins, CD47 is a ligand for SIRPα.¹³ SIRPα are ubiquitous molecules of the immunoglobulin superfamily that negatively regulate protein tyrosine kinase receptor-dependent cell proliferation. The SIRP/CD47 interaction was first described in mouse,¹³ but was shown to be conserved in rats and humans.⁴³ SIRPα is a member of the immunoglobulin superfamily with three immunoglobulin-like domains in its extracellular component. SIRPα seems to act primarily as a site for the recruitment of tyrosine phosphatase activity to the membrane, leading to inhibition of signalling from growth factor receptors. Recently, a porcine cell-surface receptor, SWC3, has been recently identified as a member of the SIRP family and associates with protein-tyrosine phosphatase SHP-1.²⁹ Further studies are underway to determine whether SWC3 associates with pig CD47 and mediates cell-to-cell adhesion.

The adhesion experiments reported here are indirect measures of CD47 function in platelets and have not addressed whether the porcine CD47 has a higher affinity for human or pig TSP, or whether pig CD47 may be mediating an event subsequent to the TSP adhesion. Moreover, the study of the cell-to-cell adhesive interaction of CD47 with its ligands, from either pig or human origin, and the capacity of signalling, should be of relevance for xenotransplantation. However, the porcine homologue of SIRPα has not yet been cloned, and the significance and efficacy of its possible interaction with pig or human CD47 remains to be determined.

To conclude, we here report the cloning and functional analysis of the pig homologue of CD47. Our results suggest that pig CD47 is widely expressed on porcine cells and tissues and that pig CD47 has a conserved role as a TSP ligand and in the modulation of integrins, as evidenced by CD47-mediated adhesion of cells and platelets to surfaces coated with fibrinogen.

Abbreviations

A

absorbance

AP–ase

acid phosphatase

BSA

bovine serum albumin

bp

base pair

CHO

Chinese hamster ovary

Dig

digoxigenin

dNTP

deoxynucleoside triphosphase

ELISA

enzyme–linked immunosorbent assay

FITC

fluorescein isothiocyanate cyanate

IAP

integrin–associated protein

IgG

immunoglobulin G

IgV

immunoglobulin variable

kDa

kilodalton

mAb

monoclonal antibody

MFR

macrophage fusion receptor

NP–40

Nonidet P–40

ORF

open reading frame

PBMC

peripheral blood mononuclear cells

PBS

phosphate–buffered saline

PCR

polymerase chain reaction

PFU

plaque–forming unit

PMSF

phenylmethyl sulphonyl fluoride

RT–PCR

reverse transcription–polymerase chain reaction

SDS–PAGE

sodium dodecyl sulphate–polyacrylamide gel electrophoresis

SE

standard error

SIRPα

signal regulatory protein of α subtype

TSP

thrombospondin

UTR

untranslated region