Search

In pig-to-human discordant xenotransplantation, human complement (C) is a major barrier to long survival of xenografts. The current idea on how to cope with this barrier is that human complement regulatory proteins are forcibly expressed on xenografts to serve as safeguards against host C-induced hyperacute rejection of xenografts. Co-expression of decay-accelerating factor (DAF) (CD55) and membrane cofactor protein (MCP) (CD46) would be the first choice for this trial, because most of the human cells are protected from C-mediated damage by two different modes with these two kinds of C-regulators. Many problems have arisen, however, for MCP expression on grafts. (i) MCP acts as a measles virus receptor, which may function to render donor pigs measles virus (MV) sensitive. (ii) MCP signals immune suppression which causes devastation of the recipient's immune responses. (iii) MCP exerts relatively low self-protective activity against C compared with other cofactors; development of more efficient forms is desirable. (iv) Grafts with a high expression level of MCP are difficult to produce. In this study, we made a number of cDNA constructs of MCP, expressed them on swine endothelial cell lines, and tested cell-protective potency and MV susceptibility. The short consensus repeat 1 (SCR1)-deleted MCP with glycosyl phosphatidylinositol (GPI)-anchored form (Delta1MCP-PI) of MCP was found to be most suitable for the purpose of overcoming these problems. However, it was also found that MV induces two modes of cytopathic effect (CPE) on swine endothelial cells, either MCP-dependent or -independent. Here, we discuss these two points which will be raised through study of MCP-transgenic animals.

Human membrane cofactor protein (MCP, CD46) is a regulator of complement activation and also serves as a receptor for measles virus (MV). We recently isolated an MCP homolog from B95a, an Epstein-Barr virus-transformed marmoset B-lymphoblastoid cell line, which is 76% identical to human-MCP. B95a-MCP acts as an MV receptor for CAM, a vaccine strain of MV, but not for Nagahata, wild-type MV strain. The four residues in human-MCP (Asp27, Lys29, Arg69, and Asp70) are reportedly MV binding sites, and these are changed in B95a-MCP (Glu27, Asp29, Pro69, and Asn70). In the present study, we constructed B95a-MCP mutants by replacing the four residues with those in human-MCP, and tested whether the Chinese hamster ovary (CHO) transfectants expressing B95a-MCP mutants become susceptible to the Nagahata strain. The CHO transfectants expressing B95a-MCP mutants formed syncytium with the CAM strain but not with the Nagahata strain. The binding of the hemagglutinin (H) of MV with B95a-MCP mutants was observed with the CAM strain but not with the Nagahata strain. These results suggest that the failure of B95a-MCP as the MV receptor for the Nagahata strain is not due simply to the natural mutations at these four residues. The critical residues for MV binding in an MCP molecule seem to differ depending upon the structure of the MV H protein.

Down-regulation of CD46 secondary to stimulation with measles virus (MV) was investigated using CD46-positive cell lines and Japanese wild-type MV strains. The cells used were simian cell lines B95a and Vero in which MV strains have been adapted to be amplified, and Chinese hamster ovary (CHO) cell transfectants expressing human CD46 with (CHO(tail+)) or without (CHO(tail-)) the cytoplasmic tail. Of four Vero-adapted and three B95a-adapted MV strains, one Vero-adapted strain named Khono (KO), down-regulated CD46 within 60 min (early down-regulation) in all cell lines examined except Vero. No strains other than Toyoshima (TY), which induced early down-regulation only in CHO(tail+) cells, induced early down-regulation of CD46 in any combination. On the other hand, conventional down-regulation of CD46 was observed 24 h post-MV inoculation (late down-regulation) when cell lines used were adapted to MV strains. Thus, we concluded that there are two modes of CD46 down-regulation by MV and the unique strain KO markedly induces early down-regulation. Also, the CD46 homologue of B95a, which fails to act as a MV receptor, is down-regulated concomitantly with MV replication (>24 h) in cells principally by competent virus strains.

We have previously shown that the ligand-binding activity of type II Fc receptor for IgG (FcgammaRIIB) on guinea pig peripheral blood polymorphonuclear leukocytes is very low and dramatically increases after treatment of the cells with proteolytic enzymes. In the present study, we analyzed the mechanism of this augmentation. We found that the protease treatment failed to enhance the binding of monomeric IgG to FcgammaRIIB, increased the binding of small immune complexes (IC) prepared under antigen-excess conditions only modestly, but markedly enhanced the binding of large IC prepared under antibody-excess conditions. These results suggest that proteolysis increases the ligand-binding avidity but not the intrinsic affinity of FcgammaRIIB. Confocal laser scanning microscopy revealed that the mobility of FcgammaRIIB on the cell surface was increased after protease treatment. In addition, transfection experiments indicated that the effect of proteolysis on IC binding to CHO cells expressing guinea pig FcgammaRIIB was strongly dependent on the receptor density. Finally, we demonstrated that the transmembrane and cytoplasmic domains of FcgammaRIIB were not involved in the proteolysis-induced augmentation of IC binding. Together our results suggest that the mobility of FcgammaRIIB, which may be restricted due to the association of the ectodomain of the receptor with unknown membrane proteins, is enhanced by proteolysis, allowing the receptors to bind multivalent ligands more readily and hence with higher avidity.

Human CD46 was identified as a complement regulator and was later shown to be a measles virus receptor. The ubiquitous distribution profile of CD46 accounted for systemic measles infection and general protection of host tissue/organs from autologous complement. A similar ubiquitous distribution was observed for swine and simian CD46 homologues based upon subsequent cDNA cloning and Northern analysis, reinforcing the roles of CD46. In contrast, recent cDNA cloning and distribution analyses of murine and guinea-pig CD46 revealed the predominant expression of these rodent CD46 homologues in the testis, especially in mature testicular germ cells. These results do not support the established functions of human CD46 but support the hypothesis that CD46 on sperm serves as a fertilization-related adhesion molecule toward eggs. Here, we review the structure, function and distribution of human CD46 and discuss the possible differences between human CD46 and its homologues recently cloned from a variety of non-human primates and other animals.

We obtained a unique CD46 cDNA, STc/CY4, from the human testis, the predicted amino acid sequence of which suggested the presence of a novel isoform of CD46. This message was present predominantly in the testis, and the predicted isoform possessed a short (11 amino acids) transmembrane section (TM) and an unidentified cytoplasmic tail (CY). When expressed in Chinese hamster ovary (CHO) cells, this CD46 isoform underwent no O-glycosylation and was mostly retained in the endoplasmic reticulum. This unusual behaviour of the new isoform was due in part to the short TM and the unusual sequences of the CY. The molecular mass of this isoform was 42,000, approximately 20,000 smaller than conventional CD46. These properties of the STc/CY4 isoform were similar to those of sperm CD46. The only difference between sperm CD46 and the STc/CY4 isoform expressed on CHO cells was that only the latter possessed N-linked sugars of high mannose types. Since the STc/CY4 isoform may behave like sperm CD46 in cellular localization and post-translational modification, studies of sperm-egg interassociation were performed using hamster eggs and CHO cell clones expressing various isoforms including the STc/CY4. Rosette formation was seen most effectively between hamster eggs and STc/CY4-expressing CHO cells. These results infer that O-glycosylation perturbs CD46-mediated sperm-binding to eggs and thus sperm CD46 lacking O-linked sugars can serve as an adhesion molecule. The possible role of CD46 in fertilization and the structural differences between sperm and conventional CD46 are discussed.

Measles virus (MV) infects not only human beings but also some simian species. The MV receptor on Vero cells (a cell line established from African Green monkey kidney cells) and human cells has been shown to be the membrane cofactor protein MCP/CD46, which is an inhibitor of autologous complement (C) activation. B95a, an Epstein-Barr virus (EBV)-transformed marmoset B cell line, is a simian cell line used for MV selection and is much more susceptible to MV than Vero cells. In the present study, we isolated cDNAs encoding MCP homologues from B95a cDNA library and assessed whether B95a-MCP is responsible for the high susceptibility of B95a to MV. The deduced amino acid sequence of the cDNA of B95a-MCP was 76% identical to that of human-MCP, and the recombinant B95a-MCP exerts C inhibitor activity. Although CAM, a vaccine strain of MV, infected Chinese hamster ovary (CHO) cells expressing B95a-MCP, Nagahata strain, a wild type of MV, failed to infect the CHO transfectants, suggesting that additional membrane molecules of B95a are responsible for the high susceptibility of B95a to the Nagahata strain.

Two phosphatidylinositol (PI)-anchored versions of a measles virus (MV) receptor membrane cofactor protein (MCP; CD46) were generated by fusing the extracellular domain of MCP to the decay-accelerating factor (DAF; CD55) or its PI anchor. The PI-anchored forms of MCP expressed on Chinese hamster ovary cells, otherwise non-permissive to MV, conferred a smaller MV cytopathic effect than a wild-type MCP, a Ser/Thr-rich domain-deletion mutant and a cytoplasmic tail-deletion mutant of MCP. Therefore the differences in MV receptor properties between the two PI-anchored and three transmembrane forms were investigated. The PI-anchored forms were predominantly expressed on microvilli as in DAF, whereas the other transmembrane forms were found on intracellular membranes. The PI-anchored forms conferred high MV-binding capacity compared with the transmembrane versions. MV replication was, however, severely suppressed in cells expressing the PI-anchored forms, resulting in ineffective syncytium formation. In contrast, cell-to-cell fusion occurred efficiently after co-transfection of cDNA species encoding MV-H. MV-F and any version of MCP. Thus the PI-anchored forms, despite showing sufficient MV binding and cell-to-cell fusion competence together with MV-H and MV-F, mediate inefficient MV entry or replication, which causes severe suppression of the MV cytopathic effect. A biased receptor distribution on microvilli might participate in the selection of a low MV uptake pathway in the PI-anchored forms of MCP. Taken together, the transmembrane portion of MCP is a critical factor for effective virus-cell fusion and the subsequent MV replication.

We isolated a 1257-bp cDNA encoding a membrane cofactor protein (MCP, CD46)/measles virus (MV) receptor-like protein from a cDNA library of Vero cells, in which wild MV strains were established. Vero cells contain MCP mRNA splice products encoding different cytoplasmic tails like human cells. The deduced amino acid sequence of the cDNA was 86% identical to that of human MCP. Vero cell MCP expressed on CHO cells was recognized by monoclonal antibodies against human MCP, and served as a potent MV receptor. In addition, Vero MCP was as effective as human MCP in human factor I-mediated C3b cleavage. Thus, the high MV susceptibility of Vero cells can in part be attributed to an MCP-like molecule that is structurally and functionally similar to human MCP.

Monkey cells persistently infected by measles virus (MV) Biken strain (Biken-CV-1 cells) showed no cytopathic effects and lacked surface expression of a homolog of human cell receptor, membrane cofactor protein CD46. Transfection of a human CD46 gene into these cells induced extensive cell fusion, indicating that down regulation of the endogenous CD46 homolog was essential for the maintenance of a noncytopathic mode of infection. Surface expression of the exogenously introduced human CD46 was also drastically down regulated in the persistently infected cells compared with uninfected cells. The down regulation was specific for CD46 and did not affect surface expression of exogenously introduced CD4. Exogenous human CD46 was synthesized efficiently in the persistently infected cells, but it did not accumulate on the cell surface. Fusion of Biken-CV-1 cells required the extracellular hemagglutinin (H-protein)-binding domain but not the cytoplasmic domain. Replacing the transmembrane and cytoplasmic domains of CD46 with a glycosylphosphatidylinositol anchor did not prevent cell fusion but completely alleviated down regulation of the glycosylphosphatidylinositol-anchored CD46 in Biken-CV-1 cells. Deletion analyses revealed that the membrane-distal sequences of the CD46 cytoplasmic domain were not only unnecessary but also inhibitory for CD46 down regulation. By contrast, the six amino acid residues proximal to the membrane contained a sequence required for CD46 down regulation in the persistently infected cells. These results indicate that CD46 is down regulated in the persistently infected cells by a mechanism that recognizes a membrane-proximal sequence in the CD46 cytoplasmic domain.

Membrane cofactor protein (MCP, CD46) of the complement system is a measles virus (MV) receptor. Human lymphocytes express a heavily glycosylated (H) and a lightly glycosylated (L) form of MCP, which confers a two-band profile on SDS-PAGE the ratio of which is controlled genetically and organ-specifically. In contrast, granulocytes express a single heavily glycosylated form regardless of lymphocyte MCP phenotype. We investigated susceptibility to MV of granulocytes and lymphocytes from individuals with different lymphocyte MCP phenotypes. In any individual, granulocytes were > 10-fold less susceptible to MV than lymphocytes, and the lymphocytes with predominant H form were generally less susceptible to those with an increasing amount of L form. Thus, lymphocytes always exhibit high susceptibility to MV compared to granulocytes in all individuals. This finding may explain the lymphopenia and immunosuppression observed secondary to MV infection.

The complement system membrane cofactor protein (MCP) CD46 serves as a C3b/C4b inactivating factor for the protection of host cells from autologous complement attack and as a receptor for measles virus (MV). MCP consists of four short consensus repeats (SCR) which are the predominant extracellular structural motif. In the present study, we determined which of the four SCR of MCP contribute to its function using Chinese hamster ovary cell clones expressing each SCR deletion mutants. The results were as follows: 1) SCR1 and SCR2 are mainly involved in MV binding and infection; 2) SCR2, SCR3, and SCR4 contribute to protect Chinese hamster ovary cells from human alternative complement pathway-mediated cytolysis; and 3) SCR2 and SCR3 are essential for protection of host cells from the classical complement pathway. These results on cell protective activity of the mutants against the human classical and the alternative complement pathways were compatible with factor I-mediated inactivation profiles of C4b and C3b, respectively, in the fluid-phase assay using solubilized mutants and factor I; the results were mostly consistent with those reported by Adams et al. (Adams, E. M., Brown, M. C., Nunge, M., Krych, M., and Atkinson, J. P. (1991) J. Immunol. 147, 3005-3011). SCR2 and SCR3 were required for C3b and C4b inactivation, and SCR4-deleted MCP showed weak cofactor activity for C4b cleavage but virtually no cofactor activity for C3b cleavage. The functional domains of MCP for the three natural ligands C3b, C4b, and MV, therefore, map to different, although partly overlapping, SCR domains.

Human membrane cofactor protein (MCP, CD46) functions as an inhibitor of the complement (C) cascade to protect host cells from C attack, and as a receptor for measles virus (MV). Normal human sera contains 10-60 ng/ml of naturally produced soluble forms of MCP, which is also a cofactor for the factor I-mediated inactivation of C3b. We produced monoclonal antibodies (mAb) against MCP and a recombinant soluble form of MCP similar to the natural soluble forms, and tested their ability to block MV infection. Vero cells and CHO cells expressing human MCP were the targets. Of the antibodies tested, M75 and M177, which blocked the C regulatory activity of MCP, efficiently blocked MV infection. More than 50 micrograms/ml of the soluble form moderately blocked MV infection of CHO cells expressing MCP, but barely blocked that of Vero cells. The two mAb and the soluble form also inhibited MV H protein-mediated green monkey erythrocyte rosette formation. A quantitative analysis suggested that 30 micrograms/ml of the soluble form functionally corresponded to 0.2 microgram/ml of M177 or M75. These data established that the C regulatory function and the MV receptor function of MCP were blocked simultaneously by the individual mAb, and that soluble forms of MCP could inhibit MV infection in cells expressing human MCP, although doses far higher than the natural concentration of soluble MCP were required.

Three major membrane cofactor protein (MCP) phenotypes with different serine-threonine (ST)-rich regions, namely STc (L-phenotype), STBC (H or U phenotype) and STABC, and the MCP without the ST domain (delta ST) were expressed in Chinese hamster ovary (CHO) cells by transfecting the respective cDNAs. The expressed molecules migrated with a larger molecular mass on SDS/PAGE than those expected from their amino acid sequences. O-Glycanase digestion showed that this was due to O-linked sugar chains. The apparent sugar contents in each ST segment were compatible with their serine and threonine contents in the ST regions. The functional properties of these phenotypes as inhibitors of human complement (C) and receptors of measles virus (MV) were compared. The classical pathway-dependent CHO cell lysis by human C was more effectively suppressed by the expressed delta ST and STC than by the STABC and STBC phenotypes, although the difference was not so prominent. In contrast, alternative C pathway-dependent CHO-cell lysis was most effectively suppressed by the STABC phenotype and was only slightly blocked by the ST-deleted mutant. MV infection occurred with all of the phenotypes, but the infectious dose required to cause the same level of syncytium formation was 100-times higher in large ST (STABC and STBC) than in small ST (STC and delta ST) phenotypes. Thus, the ST domain serves as a functional modulator in MCP: MCP with a large ST domain having high O-linked sugar contents is favourable to the effective suppression of both the alternative C pathway-mediated cytolysis and MV infection, whereas MCP with a small ST domain is favourable to the suppression of the classical C pathway.

C activation on the cell surface is supposedly regulated by membrane cofactor protein (MCP) and decay accelerating factor (DAF). These are complementary in function: MCP acts as a cofactor in factor I-mediated C3b and C4b inactivation, thus preventing the assembly of C3 convertases, whereas DAF accelerates spontaneous decay of the assembled C3 convertase. In this report, a hybrid MCP-DAF was expressed on Chinese hamster ovary cells by transfecting cDNA, and its regulatory activity was compared with those of MCP and DAF transfectants and with a transfectant expressing both MCP and DAF (MCP + DAF). The C3 deposition on sensitized CHO cells through activation of the classical pathway was blocked to a different degree with these transfectants, the order being MCP + DAF > DAF > hybrid MCP-DAF > MCP. Likewise, the C3 deposition via the alternative pathway was blocked efficiently in the order hybrid > MCP + DAF > MCP. The C-mediated cytolysis of CHO cells virtually reflected the degree of C3 fragment deposition. The MCP-DAF transfectant acquired additive protective activity against alternative pathway-mediated C3 deposition and cytolysis but was less potent in circumventing classical pathway attack than cells that expressed DAF alone or DAF + MCP. Hybrid MCP-DAF may be useful for alleviating C-mediated cell damage, especially via the alternative pathway.

Decay-accelerating factor (DAF) is a membrane protein that protects host cells from attack by its own complement. Although DAF expression on endothelial cells is thought to increase pathophysiologically, little is known about the natural mediators that modulate DAF expression on endothelial cells. In this study, we evaluated the effect of histamine on DAF expression on human umbilical vein cells (HUVEC). HUVEC were cultured with histamine, and DAF expression on HUVEC was determined by flow cytometry after immunostaining with a mAb to DAF. DAF expression on HUVEC was increased at 10 microM histamine, and the final level was increased time-dependently by 150% to 200% after a 24-h incubation with 100 microM histamine. The histamine-induced DAF expression was inhibited by actinomycin D and cycloheximide and accompanied by an increase in the DAF mRNA level, indicating that both transcription and translation are required. In addition, the histamine-induced DAF expression was inhibited by pyrilamine, an H1 blocker, but not by cimetidine, an H2 blocker, indicating that histamine induces the DAF expression through H1 receptors. We also demonstrated that the turnover of DAF is faster than that of MCP and CD59, and DAF is released into the culture supernatant. DAF is a glycosylphosphatidylinositol-linked protein that is released from HUVEC by a phosphatidylinositol-specific phospholipase C. Although HUVEC also contain the glycosylphosphatidylinositol-anchored complement inhibitor CD59, this was not released during a 24-h incubation, suggesting that the shedding of DAF from HUVEC is not caused by PI-PLC but by other enzymes, possibly proteinases. These results suggest that histamine, which is released from mast cells and basophils by complement-derived anaphylatoxins, increases the complement defense ability of endothelial cells by increasing their levels of DAF expression.

Membrane cofactor protein (MCP) cDNA was transfected into Chinese hamster ovarian tumor (CHO) cells and the functional properties of the expressed protein were studied. Cells adherent to flasks were essential for continuous expression of MCP on CHO cells. If the cells were maintained in noncoated flasks, MCP expression was markedly reduced but upon being transferred to coated flasks reexpressed the protein. MCP expressed on CHO cells had the expected m.w., approximately 50 kDa, and possessed factor I-cofactor activity. By a propidium iodide incorporation assay and by 51Cr release assay, antibody-sensitized CHO cells expressing MCP were protected from C-mediated cytotoxicity. The inhibition of lytic activity correlated with a decrease in C3 deposition. This host cell protective activity was exerted efficiently for the alternative pathway. The classical pathway was not blocked by MCP unless the cells were presensitized with low concentrations of antibody. These results imply that MCP primarily protects host cells from alternative pathway-mediated C3 targeting. In a pathologic state such as autoimmune diseases, the binding of an autoantibody to a target may overcome this protective effect of MCP via the classical pathway.

Human granulocytes (polymorphonuclear leucocytes, PMN) possess a membrane cofactor protein (MCP, CD46), which is structurally and functionally distinct from the MCPs of other cell types: it shows a single broad band of 56-80 kDa (without the doublet pattern characteristic of MCP) on SDS/PAGE and has less affinity for complement component C3b. We purified PMN MCP using monoclonal antibodies in order to study the molecular differences between it and other MCPs. Several forms of PMN MCP with size heterogeneity were noted on SDS/PAGE and by immunoblotting. O-Glycanase treatment decreased this heterogeneity, yielding a fast-migrating component identical in position on SDS/PAGE to the O-glycanase-treated MCP of other cells. The cell-specific variation of MCP, therefore, arises from post-translational glycosylation and not from a difference in primary structure. The Factor I cofactor activity of PMN MCP was more efficient in cleaving the methylamine-treated complement components C4/C3 than was MCP from other cells, which shared a similar potency of cofactor activity on a weight basis. Two types of small-form PMN MCP were identified during purification. These were 42 kDa and 30 kDa in size; the former was recognized by M177 (a monoclonal antibody against the active site marker), possessed N-linked sugars [located on the short consensus repeats (SCRs)] but not O-linked ones (on the Ser/Thr-rich region), and retained cofactor activity for C3b/C4b cleavage, similar in potency to that of other MCPs. The functionally active soluble form of MCP was observed specifically in PMN. Protease inhibitors did not inhibit liberation of the fragments, although the generated fragments became susceptible to serine proteases. The findings show that the SCRs are the functional domain of MCP and that the MCP proteolysis found only in PMN may modulate the properties of PMN MCP. In conclusion, the structural features of PMN MCP largely reflect a variability in the O-linked sugars, and the decreased affinity for C3b may be in part attributable to proteolysis.

Membrane cofactor protein (MCP) of the C system is a widely distributed regulatory molecule with C3b/C4b binding and factor I-dependent cofactor activity. A rabbit polyclonal antibody was raised against purified human MCP, and it was found to also immunoprecipitate C4b-binding protein (C4bp). Other related complement regulatory proteins, factor H, C3b/C4b receptor, and decay-accelerating factor, were not recognized by this polyclonal antibody to MCP. The cross-reactive epitope was sensitive to reduction with 2-ME and about 3% of the anti-MCP antibody reacted with C4bp. The amino-terminal 48,000-Da, chymotryptic fragment of C4bp was recognized by the antibody to MCP. This fragment of C4bp contains a seven-amino acid peptide that is identical, in its sequence and its location in the third short consensus repeat, to one found in MCP. Two polyclonal antibodies to C4bp, one raised to native and the other to reduced C4bp, did not cross-react with MCP. In addition to this one-way cross-reaction with C4bp, a protein with a m.w. of approximately 60,000 (p60) was found in two of three C4bp preparations that also cross-reacted with antiserum to MCP. p60 was present in trace quantities in the C4bp preparation and was successfully isolated from plasma by C3b affinity chromatography. Its Mr was distinct from that of MCP and other known C3b/C4b binding proteins. Furthermore, p60 was isolated by two different procedures and such material possessed no detectable cofactor activity. Based on these results, p60 is a plasma C3b-binding protein that shares epitopes with C4bp and MCP, and is probably not a soluble form of MCP.