Binding Of Xenoreactive Natural Antibodies Research Articles

Gene transfer of the adaptor Lnk (SH2B3) prevents porcine endothelial cell activation and apoptosis: implication for xenograft’s cytoprotection

Targeting protective gene expression to porcine endothelium by genetic modification of the donor could improve xenograft survival by controlling cell activation and death. We previously found that, in endothelial cells (EC), the molecular adaptor Lnk (SH2B3) is a negative regulator of cytokine signaling. We also have shown that Lnk is upregulated in pig EC (PAEC) in response to tumor necrosis factor-α (TNF) and xenoreactive natural antibodies (XNA) binding. The present study investigated whether ectopic expression of human Lnk using gene transfer may be efficient to control signaling pathways associated with inflammation and apoptosis in porcine aortic endothelial cells (PAEC). Endothelial cells cultures were established from WT and Gal(-/-) pigs and transduced with a recombinant adenovirus encoding human Lnk. Phenotype and functions of transduced PAEC expressing Lnk were analyzed by flow cytometry, western blot and XNA and complement-dependent assays. The regulatory functions of Lnk toward inflammation were assessed in TNF-activated EC, and the protective functions were tested toward TNF-induced apoptosis and anoïkis. Apoptosis assays included DNA content analysis and caspase-3/7 activity. First, we found that as a result of adenoviral transduction, human Lnk was efficiently and similarly expressed in EC from WT or Gal(-/-) pigs. Lnk expression or EC transduction caused no significant change in the binding of XNA (IgG and IgM) to PAEC and has no effect on complement activation and C5b-9 formation. We demonstrated that expression of human Lnk efficiently inhibits TNF signaling in PAEC and decreases VCAM-1 induction by 46.3 ± 1.2% compared to controls (n = 6, **P < 0.01). Furthermore, expression of Lnk was associated with a significant decrease in the percentage of caspase-3/7-dependent apoptosis caused by TNF in the presence of actinomycin D and also reduces cell death by anoïkis by 25.0 ± 1.9% compared to controls (n = 5, **P < 0.01). Together, these findings indicate that the signaling adaptor Lnk is effective to reduce PAEC activation and apoptosis. Thus, Lnk is a potential candidate for the modulation of signaling pathways to protect vascular EC from inflammation in xenotransplantation.

Resistance to anti-xenogeneic response by combining α-Gal silencing with HO-1 upregulation

Background A major barrier to clinical xenotransplantation is preformed xenoreactive natural antibodies (XNA) found in higher primates which react to Galα(1,3)Gal (α-Gal) epitopes found on lower species. Accommodation of organs to xenogeneic recipients involves upregulation of cytoprotective genes and resistance to complement dependent cytotoxicity (CDC). Methods To develop methods of increasing these organ-protective effects, we established an in vitro CDC model utilizing human serum as the source of XNA and porcine endothelial cells (pEC) as targets. Results Using this system we demonstrated that downregulation of α-Gal epitopes by siRNA silencing of α1,3-galactosyltransferase (α-GT) led to marginal protection from CDC while α-Gal silencing combined with Griffonia simplicifolia isolectin B4 (GS-IB4), a lectin that specifically binds to α-Gal epitopes, led to complete protection. Interestingly, α-Gal silencing and GS-IB4 mediated effects were not associated with inhibition of XNA binding to cells, but with significant decreased E-selectin expression and cytoprotective gene HO-1 upregulation. PI3K inhibitor LY294002 could block the elevation of HO-1 protein expression and reverse the protective effect of α-Gal silencing and GS-IB4 against CDC. Conclusion These data support the use of combination approaches targeting independent accommodation mechanisms to synergistically enhance donor organ survival in a xenogeneic setting.

Non‐human to human organ transplantation:Its biologic basis and a potential role for radiation therapy

There is an inadequate supply of human donor organs for transplantation. Xenotransplantation, the transplantation of organs from non-human animals to humans, is one of the potential solutions to this problem. The pig appears to be the preferred donor. For xenotransplantation to be successful, researchers must deal with three fundamental problems: (1) Hyperacute rejection of porcine organs, related to binding of xenoreactive natural antibodies of the recipient to antigens on the graft's endothelial cells, must be overcome. (2) Transmission of animal pathogens to humans must be prevented. Concern about zoonosis is not only directed to the transplant recipient but also concerns the risk that an infectious agent will be transferred from the recipient to the general population. (3) The xenografted organ must be physiologically compatible with the recipient. The physiological function of a pig organ in a human and its ability to sustain a human are problematic. Total lymphoid irradiation (TLI) and thoraco-abdominal irradiation (TAI) as immunosuppressive modalities have been investigated in rodent-to-rodent, large mammals and non-human primates-to-primates, and pig-to-primate models. In certain clinical situations, TLI and TAI may prove to be important components for the preparation of the xenotransplant recipient. Progress in genetic engineering and cloning may soon lead to clinical trials in xenotransplantation.

Suppression of the ??-1,3-Galactosyl Epitope by Golgi Expression of a Coffee Bean ??-1,3-Galactosidase Transgene

725 The α-1,3-galactosyl epitope (α-1,3Gal) is thought to be the main target antigen for human natural antibodies to porcine xenografts. Several methods have been employed to modify α-1,3Gal expression and have been partially successful in reducing the levels of α-1,3Gal as measured by IB4 lectin binding. However, antibody/complement-mediated cytotoxicity remains unchanged or increased in these models. Discrepancies between α-1,3Gal expression and functional antibody activity may be attributed to the currently unpredictable, tissue specific changes in Golgi glycosylation. To help elucidate this problem, we have developed a model designed to suppress α-1,3Gal in the late stage of Golgi glycosylation. Our aim was to cleave the α-1,3Gal epitope at or distal to its production in the Golgi, allowing for increased sialic acid capping in the Trans Golgi Network (TGN). Sialic acid has been shown to confer resistance to complement-mediated lysis of pathogens and mammalian cells. A fusion protein composed of the enzymatic region of the coffee bean α-1,3-galactosidase gene linked to an amino terminal TGN targeting peptide derived from the rat 2,6-sialyltransferase was designed and produced. Next, stable NIH-3T3 transfectants positive for both expression and activity of the fusion protein were generated. Flow analysis of these cells using IB4 and L. polyphemus lectins demonstrated a >30% decrease in surfaceα-1,3Gal levels and a 100% increase in sialic acid expression. Although binding of xenoreactive natural antibodies to these cells remained unaltered, a >10% decrease in complement-mediated cytotoxicity as measure by propidium iodide exclusion was noted. From these data, we conclude that TGN expression of the galactosidase fusion protein was able to partially suppressα-1,3Gal, causing a shift in surface carbohydrate expression. More importantly, expression of the coffee bean α-1,3-Galactosidase transgene resulted in an increased resistance to complement mediated lysis not seen in previous transgenic models.

Suppression of the α-1,3-Galactosyl Epitope by Golgi Expression of a Coffee Bean α-1,3-Galactosidase Transgene.

725 The α-1,3-galactosyl epitope (α-1,3Gal) is thought to be the main target antigen for human natural antibodies to porcine xenografts. Several methods have been employed to modify α-1,3Gal expression and have been partially successful in reducing the levels of α-1,3Gal as measured by IB4 lectin binding. However, antibody/complement-mediated cytotoxicity remains unchanged or increased in these models. Discrepancies between α-1,3Gal expression and functional antibody activity may be attributed to the currently unpredictable, tissue specific changes in Golgi glycosylation. To help elucidate this problem, we have developed a model designed to suppress α-1,3Gal in the late stage of Golgi glycosylation. Our aim was to cleave the α-1,3Gal epitope at or distal to its production in the Golgi, allowing for increased sialic acid capping in the Trans Golgi Network (TGN). Sialic acid has been shown to confer resistance to complement-mediated lysis of pathogens and mammalian cells. A fusion protein composed of the enzymatic region of the coffee bean α-1,3-galactosidase gene linked to an amino terminal TGN targeting peptide derived from the rat 2,6-sialyltransferase was designed and produced. Next, stable NIH-3T3 transfectants positive for both expression and activity of the fusion protein were generated. Flow analysis of these cells using IB4 and L. polyphemus lectins demonstrated a >30% decrease in surfaceα-1,3Gal levels and a 100% increase in sialic acid expression. Although binding of xenoreactive natural antibodies to these cells remained unaltered, a >10% decrease in complement-mediated cytotoxicity as measure by propidium iodide exclusion was noted. From these data, we conclude that TGN expression of the galactosidase fusion protein was able to partially suppressα-1,3Gal, causing a shift in surface carbohydrate expression. More importantly, expression of the coffee bean α-1,3-Galactosidase transgene resulted in an increased resistance to complement mediated lysis not seen in previous transgenic models.

Humoral Responses to Pig-to-Baboon Cardiac Transplantation: Implications for the Pathogenesis and Treatment of Acute Vascular Rejection and for Accommodation

ABSTRACT: Organs transplanted between phylogenetically-disparate species, such as from the pig into the primate, are subject to hyperacute and acute vascular rejection. Hyperacute rejection of a porcine organ by a primate is thought to be initiated by the binding of xenoreactive natural antibodies to Galα1-3Gal expressed on the endothelial lining of blood vessels in the xenograft. The factor(s) which initiates acute vascular rejection is uncertain; however, there is some evidence implicating xenoreactive antibodies. The nature of the humoral response which might contribute to acute vascular rejection of a porcine organ was investigated in baboons which received a porcine cardiac xenograft plus immunosuppression with methylprednisolone, azathioprine, and cyclosporine. Following rejection and surgical removal of the xenografts, the serum concentration of xenoreactive antibodies increased in untreated animals but in immunosuppressed animals was similar to the concentration in pre-immune serum. The antibodies in the sensitized recipients were specific for Galα1-3Gal (70–95%) and other determinants (5–30%). However, cross-blocking studies showed that, following xenotransplantation, the immunosuppressed baboons had no detectable IgM or IgG directed against “new” endothelial antigens. These results indicate that antibodies made by immunosuppressed individuals in response to xenotransplantation are much like xenoreactive natural antibodies and suggest that acute vascular rejection might in some cases be addressed by therapeutic strategies aimed at those antibodies.

Comparative study of target antigens for primate xenoreactive natural antibodies in pig and rat endothelial cells.

A rat-to-primate cardiac xenograft model has been proposed as an alternative to the clinically relevant but more cumbersome pig-to-primate model for assessing the efficacy of strategies aimed at preventing xenograft hyperacute rejection. As in pig xenografts, the rejection of rat hearts was mediated by the binding of xenoreactive natural antibodies (XNA) and complement activation. The present study was conducted to identify target antigens recognized by cynomolgus and rhesus monkey IgM XNA on rat tissues and cells in comparison with pig cells. The reactivity of rhesus or cynomolgus serum on pig and rat endothelial cells (ECs) was studied by flow cytometry, ELISA, and complement-dependent cytotoxicity, after removal of primate XNA by perfusion of pig livers, immunoadsorption on a Gal alpha(1,3)Gal affinity column, and enzymatic removal of alpha-galactosyl epitopes from the cell surface. Rat and pig EC extracts were also immunoprecipitated with primate serum and resolved in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The expression of the Gal alpha(1,3)Gal epitope was analyzed on rat tissues and ECs by immunohistochemistry, flow cytometry, and Western blot, using the isolectin B4 from Griffonia simplicifolia. Removal of primate XNA or of alphaGal epitopes resulted in a decrease in XNA binding to pig and rat cells, leaving a similar degree of residual reactivity in the two species. At least five proteins of 260, 210, 110, 56, and 50 kDa were immunoprecipitated on rat ECs, with molecular weight similar to several proteins identified on pig ECs. These results suggest that primate XNA recognize similar antigens on rat and pig ECs. Rat cells expressed lower levels of the Gal alpha(1,3)Gal epitope than pig cells. A large proportion, but not all, of primate XNA react with this epitope on pig and rat ECs. This study suggests that the rat is a valuable species for the evaluation of genetic engineering strategies on the vascular endothelium aimed at preventing hyperacute xenograft rejection.

Expression of α-1,3-Galactose and Other Type 2 Oligosaccharide Structures in a Porcine Endothelial Cell Line Transfected with Human α-1,2-Fucosyltransferase cDNA

The binding of xenoreactive natural antibodies to the Galalpha1-3Galbeta1-4GlcNAc (alpha-galactose) oligosaccharide epitope on pig cells activates the recipient's complement system in pig to primate xenotransplantation. Expression of human alpha-1, 2-fucosyltransferase in pigs has been proposed as a strategy for reducing the expression level of the alpha-galactose epitope, thereby rendering the pig organs more suitable for transplantation into humans. The aim of this study was to examine how the cell surface expression of alpha-galactose, H, and related fucosylated and sialylated structures on a pig liver endothelial cell line is affected by transfection of human alpha-1,2-fucosyltransferase cDNA. Nontransfected and mock-transfected cells expressed alpha-galactose, alpha-2,3-sialylated, and alpha-2,6-sialylated epitopes strongly, with low level expression of type 2 H and LewisX. By contrast, expression of the H epitope was increased 5-8-fold in transfected cells with a 40% reduction in the expression of alpha-galactose epitope and a 50% decrease in sialylation, as measured by binding of Maackia amurensis and Sambuccus nigra agglutinins. LewisX expression was reduced to background levels, while the LewisY neoepitope was induced in human alpha-1,2-fucosyltransferase-expressing pig cells. The activities of endogenous alpha-1,3-galactosyltransferase, alpha-1,3-fucosyltransferases, and alpha-2,3- and alpha-2, 6-sialyltransferases acting on lactosamine were unaffected. Our results show that a reduction in alpha-galactose epitope expression in porcine endothelial cells transfected with human alpha-1, 2-fucosyltransferase cDNA may be achieved but at the expense of considerable distortion of the overall cell surface glycosylation profile, including the appearance of carbohydrate epitopes that are absent from the parent cells.

Human natural killer lymphocytes directly recognize evolutionarily conserved oligosaccharide ligands expressed by xenogeneic tissues.

In discordant xenogeneic species combinations, vascularized transplants are hyperacutely rejected, due to binding of xenoreactive natural antibodies (XNA) to selected tissues of the graft, followed by activation of the complement and coagulation cascades. A major epitope recognized by human XNA is the terminal disaccharide Gal alpha(1,3)Gal. Poorly defined, early cell-mediated events also contribute to recognition and rejection of discordant xenografts, and we have suggested a role of natural killer (NK) lymphocytes in this process. Human NK cells were used as effectors in functional assays of adhesion to and lysis of xenogeneic discordant endothelial cells in vitro. Adhesion and lysis inhibition experiments were performed using a large panel of carbohydrates, as well as F(ab')2 fragments of human XNA. COS cells transduced with the porcine alpha-galactosyltransferase were also used as targets for NK cell adhesion. We demonstrate that XNA-reactive carbohydrate epitopes expressed by xenogeneic cells, including Gal alpha(1,3)Gal, are also directly recognized by human NK cells. First, selected carbohydrates in solution displace with comparable efficiency both XNA and NK cell binding to xenogeneic endothelium; second, XNA F(ab')2 fragments selectively inhibit human NK cell adhesion to porcine endothelium, but not to human endothelium; third, unstimulated NK lymphocytes adhere selectively to COS-7 cells expressing the porcine glycosyltransferase that encodes the Gal alpha(1,3)Gal epitope. Collectively, our findings suggest that humoral and cellular components of the natural immune response against heterologous species independently evolved recognition patterns directed against overlapping carbohydrate determinants.

Intact pig pancreatic islet function in the presence of human xenoreactive natural antibody binding and complement activation.

The expression of xenogeneic epitopes and the activation of human complement by adult pig islets after prolonged culture have hitherto not been described. Freshly isolated and cultured islets were analyzed by fluorescence-activated cell sorter analysis, fluorescence microscopy, and immunohistology for expression of Gal(alpha1,3)Gal epitopes, binding of human xenoreactive natural antibodies (XNA), and complement deposition. Freshly isolated and cultured islets showed detectable Gal(alpha1,3)Gal expression and human XNA binding limited to intraislet capillary endothelial cells. No significant modification in Gal(alpha1,3)Gal expression and human XNA binding levels was detected in adult pig islets cultured for up to 4 days compared with freshly isolated islets. Incubation of pig islets with human serum demonstrated the deposition of C3, C4, and membrane attack complex, but not factor B with a similar pattern to XNA. However C3 and C4 showed a more widespread deposition. Despite complement activation, no cytotoxic effect on islets was detected after 4 hr of incubation with human serum capable of killing porcine endothelial cells. Even after 4 days of culture in 50% intact human serum, pig islets retained both their normal morphology and a normal insulin response to glucose stimulation. Neither islet cell lysis nor, more importantly, any alteration in beta cell function occurred, which suggests that adult pig islets may not be directly damaged by serum after xenotransplantation in humans. Nevertheless, complement activation in vivo could trigger rapid cellular rejection mechanisms through islet cell opsonization and release of bioactive fragments.

Human peripheral blood leukocyte-reconstituted severe combined immunodeficient mouse: analysis of the human immune response against porcine islet transplantation.

Human peripheral blood leukocyte (PBL)-reconstituted severe combined immunodeficient (SCID) mice (Hu-PBL-SCID) were used as a model to study xenograft rejection in humans. SCID mice were reconstituted with human PBL using a protocol that included a booster injection with anti-human CD3 antibody-primed cells. This protocol enhanced chimera establishment in SCID mice and resulted in the detection of higher levels of human Ig when compared with SCID mice receiving unprimed PBL alone. Human xenoreactive natural antibodies (XNA), both IgM and IgG subtypes, which recognized porcine islets (PI), were detected in sera of Hu-PBL-SCID by cytofluorometric analysis. Pretreatment of porcine cells with GS-IB4 lectin inhibited the XNA binding, demonstrating the specificity of the XNA from Hu-PBL-SCID. Western blot analysis showed that XNA from normal human serum and Hu-PBL-SCID serum recognized similar xenoantigens on PI, indicating that Hu-PBL-SCID contained a XNA repertoire representative of normal human serum. Immunofluorescent staining of the tissue sections revealed that both human IgG and IgM bound in vivo to the PI engrafted beneath the kidney capsule of Hu-PBL-SCID. In addition, mouse complement (C3) was detected on xenografted PI. The function of xenografted islets were monitored by measuring porcine insulin concentration using a radioimmunoassay. Porcine insulin concentration in the sera of both Hu-PBL-SCID and plain SCID xenografted with PI was similar for up to 14 days after transplantation, after which the insulin levels in Hu-PBL-SCID decreased, thereby indicating rejection. Therefore, PI transplanted into the Hu-PBL-SCID should be a useful model for the study of cellular as well as acquired humoral immune response against xenoislets.

Tumor necrosis factor-induced expression of porcine glycoproteins gp65 and gp100 recognized by human xenoreactive natural antibodies.

In the pig-to-primate model of xenotransplantation, graft rejection is initiated by binding of the recipient's xenoreactive natural antibodies (XNA), mainly of the IgM type, to antigens constitutively expressed on donor endothelial cells (EC). As a consequence of XNA binding and complement fixation, the EC become activated, which is considered to be a major mechanism promoting hyperacute as well as later phases of graft rejection. It is not clear whether binding of XNA to activated EC also contributes to delayed rejection. We asked whether EC activation by cytokines results in the expression of other novel surface antigens recognized by XNA which might become relevant in progressive stages of graft rejection. We activated porcine aortic EC and smooth muscle cells with tumor necrosis factor (TNF), interleukin 1, or lipopolysaccharide and studied expression of new XNA-binding antigens. Expression of two glycoproteins, gp65 and gp100, was strongly induced by recombinant human TNF in EC but not in smooth muscle cells. Notably, gp100 expression was specific to TNF activation, whereas gp65 could also be induced by interleukin 1 or lipopolysaccharide. Cell surface labeling indicated that gp65 is expressed on the plasma membrane. Recognition of XNA-binding antigens on resting EC occurs via alpha-galactosyl epitopes. In contrast, gp65 and gp100 were recognized independently of this epitope. Our data show that gp65 and gp100 represent selective cytokine-induced markers on EC that may have importance in a porcine-to-primate model of xenotransplantation. Conceivable functions of gp65 and gp100 are discussed.

Reduction in the level of Gal(alpha1,3)Gal in transgenic mice and pigs by the expression of an alpha(1,2)fucosyltransferase.

Hyperacute rejection of a porcine organ by higher primates is initiated by the binding of xenoreactive natural antibodies of the recipient to blood vessels in the graft leading to complement activation. The majority of these antibodies recognize the carbohydrate structure Gal(alphal,3)Gal (gal epitope) present on cells of pigs. It is possible that the removal or lowering of the number of gal epitopes on the graft endothelium could prevent hyperacute rejection. The Gal(alpha1,3) Gal structure is formed by the enzyme Galbeta1,4GlcNAc3-alpha-D-galactosyltransferase [alpha(1,3)GT; EC 2.4.1.51], which transfers a galactose molecule to terminal N-acetyllactosamine (N-lac) present on various glycoproteins and glycolipids. The N-lac structure might be utilized as an acceptor by other glycosyltransferases such as Galbeta1,4GlcNAc 6-alpha-D-sialyltransferase [alpha(2,6)ST], Galbeta1,4GlcNAc 3-alpha-D-Sialyltransferase [alpha(2,3)ST], or Galbeta 2-alpha-L-fucosyltransferase [alpha(1,2)FT; EC 2.4.1.691, etc. In this report we describe the competition between alpha(1,2)FT and alpha(1,3)GT in cells in culture and the generation of transgenic mice and transgenic pigs that express alpha(1,2)Fr leading to synthesis of Fucalpha,2Galbeta- (H antigen) and a concomitant decrease in the level of Gal(alpha1,3)Gal. As predicted, this resulted in reduced binding of xenoreactive natural antibodies to endothelial cells of transgenic mice and protection from complement mediated lysis.

Expanding the donor pool

Expanding the donor pool

Xenoreactive natural antibodies in the world of natural antibodies: typical or unique?

Xenoreactive natural antibodies are thought to be responsible for initiating the hyperacute rejection of porcine organs transplanted into primates. Progress has been made in recently years in the characterization of the specificity and functions of these antibodies. Xenoreactive natural antibodies recognize Galα1-3Gal, a carbohydrate related to the blood group A and blood group B antigens. The presence of Galα1-3Gal may not be sufficient to allow the binding of xenoreactive natural antibodies under physiological conditions; rather, the clustering of Galα3-3Gal determinants may dictate the extent to which xenoreactive natural antibodies attach to surfaces on which Galα1-3Gal is expressed. The predominant role of xenoreactive natural antibodies in the pathogenesis of hyperacute rejection in porcine-to-primate xenotransplantation involves the activation of complement. Complement activation is mediated by IgM, not IgG xenoreactive natural antibodies. Based on functional avidity, thermal binding optima, thermal liability and homogeneity of binding interactions, xenoreactive natural IgM appear to be members of a family of natural antibodies which includes isohaemagglutinins. The classification of xenoreactive antibodies and isochaemagglutinins into a family of antibodies may provide further insight into the nature and physiological function of xenoreactive natural antibodies.

VARIATION IN THE LEVEL OF XENOANTIGEN EXPRESSION IN PORCINE ORGANS

Hyperacute rejection of vascularized porcine to primate xenografts is initiated by the binding of xenoreactive natural antibodies to donor endothelium. We tested the hypothesis that the level of xenoantigen expression varies in the population of potential porcine donors and may determine the amount of binding of xenoreactive natural antibodies to a porcine organ perfused by xenogeneic blood. Two hundred ninety pigs were studied using an inhibition ELISA that quantitated the xenoantigen level on porcine platelets. Based on this assay, the levels of xenoantigen expression in the population adhered to a normal distribution. Kidneys from pigs found to express high antigen levels and kidneys from pigs found to express low antigen levels were perfused with baboon blood using an extracorporeal circuit. In multiple experiments, a significant difference was observed in the amount of xenoreactive natural antibody adsorbed by high antigen versus low antigen organs. Normalizing for the weight of the perfused organs and for levels of natural antibody in individual baboons, high antigen organs adsorbed 3.6 +/- 1.3 U of xenoreactive natural antibody/g and low antigen organs adsorbed -0.8 +/- 1.0 U of xenoreactive natural antibody/g (P < 0.002). Immunopathology of tissues from the perfused organs demonstrated more deposition of IgM and C4 in high than in low xenoantigen organs. The quantitative relationship between binding of xenoreactive natural antibodies to platelets and to whole organs suggests that platelets are a valid representation of endothelial cell antigen expression in vivo. Despite the probable importance of Gal alpha(1-3)Gal as an epitope recognized by xenoreactive natural antibodies, differences in the binding to platelets or to organs of the GS-I-B4 lectin that recognizes that sugar had no correlation with the differences in binding of IgM to these tissues. Variation in expression of xenoantigen may be exploited to selectively breed donors for xenotransplantation that are less susceptible to attack by xenoreactive natural antibodies.