[Frontiers in Bioscience 1, d248-265, September 1, 1996]


D.K.C. Cooper, MA, PhD, MD, MS, FRCS, FACC, FACS

Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma, USA

Received 07/16/96; Accepted 08/12/96; On-line 09/01/96


Most attention has to-date been paid towards overcoming HAR for, if this cannot be achieved, then there seems little purpose in directing efforts to overcoming the delayed xenograft rejection (and cellular rejection) that occurs subsequently. Work in this field can loosely be divided into 4 main approaches (Table 3).

  • (i) Depletion in the recipient of anti-pig antibodies or inhibition of their attachment to graft antigens.
  • (ii) Depletion or inhibition of complement in the recipient.
  • (iii) The development of immunological tolerance in the recipient to donor tissues by the creation of mixed chimerism.
  • (iv) Genetic engineering of a donor pig whose organs are protected from the antibody-mediated complement activation that would follow transplantation into a human.
Table 3: Methods of prevention of hyperacute rejection in discordant xenotransplantation.

Directed at the Recipient
  • Depletion or inhibition of anti-pig antibody
  • Depletion or inhibition of complement
  • Development of immunological tolerance to donor tissues (mixed chimerism)
Directed at the Donor*
  • Replacement of alphaGal epitopes by expression of another carbohydrate structure
  • Expression of human complement inhibiting proteins

*By genetic engineering techniques or possibly by gene therapy.

The currently available pharmacologic immunosuppressive agents are totally ineffective in preventing HAR, but have been shown to play a role in reducing the rapidity of delayed xenograft rejection (17,18). There is no evidence to-date, however, that they can totally prevent delayed vascular rejection. Clearly, until this hurdle has been overcome, their role in the prevention of other cellular responses that are likely to follow remains uncertain.

5.1. Anti-pig antibody depletion or inhibition

The potential recipient can be depleted of all antibodies by plasma exchange (56), or of some antibodies by immunoadsorption techniques using immunoaffinity columns consisting of, for example, staphyloccocal protein A (17,57,58). However, these techniques deplete the patient of antibodies that may be important in protecting against infection. A preferable technique is to utilize highly specific extracorporeal immunoaffinity columns where only those anti- pig antibodies that are detrimental to the transplant will be depleted (22-24). This can be achieved by utilizing an extracorporeal immunoaffinity column of an alphaGal oligosaccharide (59-61, and Taniguchi, S., et al., submitted for publication).

Such a technique has been shown to be successful clinically with regard to depletion of anti-A or anti-B histo- blood group antibodies in patients receiving ABO-incompatible organs (62) or bone marrow (63,64) allografts, where the pattern of HAR that can occur in the unmodified recipient is almost identical to that seen in discordant xenotransplantation. There is increasing evidence that it will also be successful with regard to depletion of anti-alphaGal antibodies.

Anti-alphaGal or anti-A or anti-B antibodies clearly return once the course of extracorporeal immunoadsorption is discontinued, although concomitant pharmacologic immunosuppressive therapy (and possibly splenectomy) prevents significant antibody rebound and maintains a low level of antibody.

In the case of organ allografting across the ABO barrier, the return of antibody directed against target epitopes on the donor organ does not result in rejection of the organ. The mechanism by which this phenomenon (known as accommodation (65,66)) occurs remains unknown. It is not yet certain that accommodation will take place after discordant xenotransplantation in temporarily antibody-depleted recipients.

An alternative approach would be to carry out what has been termed "specific intravenous carbohydrate therapy," in which synthetic or natural alphaGal oligosaccharides are infused continuously into the recipient circulation (23,67-69). The oligosaccharides are bound by the anti-alphaGal antibodies in the blood, causing "neutralization" of the antibodies so that they are no longer free to attack the pig organ when it is transplanted.

Once again, this has been demonstrated to be a successful experimental approach with regard to inhibition of anti-A or anti-B antibodies. There is now a considerable amount of in vitro (Figure 4) (59) and a little in vivo (Figure 5) (60,61, and Taniguchi, S., et al.,submitted for publication) evidence that this approach may also be successful with regard to anti-alphaGal antibodies.

Figure 4: Reduction in cytotoxicity of (A) human and (B) baboon serum on PK15 cells after incubation of serum with increasing concentrations of alpha-galactosyl oligosaccharides (Dextra, Reading, UK) B-disaccharide = Gal alpha1-3Gal; B- trisaccharide = Gal alpha1-3Galß1-4Gal; B-tetrasaccharide = Gal alpha1-3Galß1-4Gal alpha1-3Gal. (From Neethling, F.A., et al. (59))

Figure 5: Serum cytotoxicity to PK15 cells and anti-alphaGal antibody levels (measured by mouse laminin ELISA and alphaGal disaccharide ELISA) in a baboon that underwent a total of 4 extracorporeal immunoadsorptions using an immunoaffinity column of alphaGal disaccharide. Cytotoxicity to the serum was immediately depleted after the first EIA and remained at extremely low or insignificant levels for approximately 10 days. (From Taniguchi, S., et al., submitted for publication)

The major limiting factor of the above approaches at the present time has been the difficulty and expense of synthesizing these oligosaccharides in the large quantities required, particularly if they are to be infused i.v., when very large quantities are required. This problem is likely to be overcome in the near future by the use of enzymatic methods to produce the relevant oligosaccharides in kilogram quantities. The alphaGal oligosaccharide must be of the Gal alpha1-3Gal configuration at its reducing end, and can be a di-, tri-, tetra-, or a pentasaccharide. There is evidence that the structure of the non-reducing end also plays a role (though less significant) in the efficiency of the oligosaccharide to inhibit antibody (Table 4) (70).

Table 4: Oligosaccharide concentrations

Oligosaccharide*Concentration (µm) in serum
1. *Fuc alpha1-2Galß1-R>10000>10000
2. Galß1-R>10000>10000
3. Gal alpha1-2Galß1-R'7000>10000
4. Gal alpha1-3Gal386(±149)a301(±44)e
5. Gal alpha1-3Galß1-4Gal163(±73)b141(±60)f
6. Gal alpha1-3Galß1-4Gal alpha1-3Gal54(±31)c119(±30)g
7. Gal alpha1-3Galß1-4GlcNAc27(±11)d31(±4)h

*Underlined type indicates structural differences of the oligosaccharide from the major pig vascular endothelium glycolipid Gal alpha1-3Galß1-4GlcNAcß1-3Galß1-4Glcß1-Cer. R represents 1-3 or 1-4 linkages to Gal or to GlcNAc; R' is -O(CH2)3NHCOCH3.
The results of the strong inhibitors (a-h) are expressed as mean ± SD (n=3). Statistical significance: a vs c (t=3.1); a vs d (t=3.9); b vs d (t=3.1); e vs h (t=9.7); f vs h (t=3.1); g vs h (t=5.4), all with P<0.02. The other comparisons did not reach the P=0.05 level of significance.(From Neethling, F.A., et al. (70))

A search for a cheap source of alphaGal oligosaccharides has revealed that pig stomach mucin, which is readily available commercially, contains a subfraction that is highly efficient in inhibiting human and baboon anti-alphaGal antibody lysis of pig cells in culture (71,72).

An alternative to the use of alphaGal oligosaccharides, either in immunoaffinity columns or as an i.v. infusate, is the anti-idiotypic antibody. Koren et al. (73, 74) have produced anti-idiotypic antibodies in mice by the injection into the mouse of human anti-pig antibody (eluted from pig organs after repeated perfusion with human plasma). Several of these anti-idiotypic antibodies, when incubated with human serum, have been demonstrated to have a major inhibitory effect on serum cytotoxicity towards pig PK15 cells in vitro. Furthermore, when infused i.v. in combinations of two into baboons, serum cytotoxicity has again been markedly reduced (from 100% to approximately 10%).

5.2. Complement depletion or inhibition

Purified cobra venom factor (CVF) has been shown to be extremely effective in depleting complement and can clearly protect a discordant organ from HAR (17,18,58). However, even when the complement level is unmeasurable by standard laboratory tests, histopathological features of delayed xenograft rejection begin to develop within 2-3 days and lead to graft failure within a relatively short period of time (<1 week) (18). The addition of concomitant pharmacologic immunosuppressive therapy, presumably by suppressing both B and T cell activity, delays rejection further, but mixed vascular and cellular rejection is seen within days with the longest survival of a pig organ in a nonhuman primate to-date being 27 days (18).

Soluble complement receptor type I (sCR1) has also had success in prolonging discordant xenograft function (75-78). Human complement receptor 1 is a single-chain cell-surface glycoprotein found on erythrocytes, some T lymphocytes, all mature B lymphocytes, neutrophils, eosinophils, basophils, monocytes/macrophages, and certain other cells (79). It is also found circulating as a soluble form in plasma at low concentrations. The interaction of complement receptor 1 with some fractions of the complement cascade regulates complement activation through its convertase decay accelerating activity and its factor 1 cofactor activity (79-81). Fearon and colleagues constructed a soluble form of complement receptor 1 which lacked the transmembrane and cytoplasmic protein domains (81). This sCR1 retains all the known activities of the native cell surface receptor, and has been demonstrated to be a potent and selective inhibitor of both the classical and alternative complement pathways. Discordant xenografts have survived for over three weeks when protected by sCR1 (82).

It would seem, however, that complement depletion or inhibition alone, although valuable therapeutic approaches to assist in overcoming HAR, will not be sufficient to prolong discordant xenograft survival indefinitely.

5.3. The development of tolerance to donor species by mixed chimerism

Donor species-specific tolerance would clearly be desirable and may indeed prove essential if late rejection of a discordant xenograft proves to be significantly more severe than that of an allograft. Important studies have been carried out over a number of years in experimental animals by two groups, namely those headed by Myburgh at the University of the Witwatersrand in Johannesburg (83-85) and by Sachs, formerly at the National Institutes of Health in Bethesda and more recently at Harvard Medical School (86-93).

The induction of donor specific tolerance would clearly eliminate the development of acute or chronic rejection. The elimination of chronic rejection (e.g. graft atherosclerosis or bronchiolitis obliterans) is possibly even more important than that of acute rejection as there is no effective treatment for chronic rejection even in allografts. If tolerance could be achieved, pharmacologic immunosuppressive therapy would not be necessary and therefore the accompanying risks of opportunistic infection, malignancy, and drug toxicity would be avoided.

Sykes and Sachs (93) have pointed out that the tolerance approach may be well suited for xenotransplantation since animal donors are available electively (and not under emergency conditions as are cadaveric human donors) allowing for the timing of tolerance induction and transplantation to be elective. Tolerance-inducing cell populations (e.g. bone marrow) can be obtained from the donor, the recipient can undergo the procedure to induce a tolerant immune system, and the organ graft from the same donor can be inserted at the optimum time. In addition, the potential for generating fully inbred xenograft donors (e.g. miniature swine) provides the possibility of using an unlimited source of genetically homogeneous tissue whenever it is required for maintenance of the tolerant state. Xenogeneic donors could be modified using genetic engineering or gene therapy techniques to facilitate induction of tolerance to xenoantigens.

Two approaches are being investigated by the Harvard group (93), namely (i) the use of xenogeneic hematopoietic cell transplantation to induce a state of mixed chimerism, and (ii) thymectomy followed by replacement with a xenogeneic donor thymus after depletion of the preexisting peripheral T cell repertoire. In this brief review, only the mixed chimerism approach will be discussed.

After much preliminary work in rodents, the Sachs group has investigated the development of tolerance to allografts and xenografts in nonhuman primates. The basic protocol (Figure 6) consists of the nonhuman primate receiving 3.0 Gy of whole body irradiation (WBI), 7.0 Gy of thymic irradiation, and horse anti-human ATG preoperatively. Bilateral nephrectomy, splenectomy, orthotopic kidney transplantation, and donor bone marrow administration are all performed on day 0. In order to supplement suppression of mature T cells by ATG, treatment with cyclosporine intramuscularly is begun on day 1 and continued for 4 weeks, but then no further immunosuppression is administered.

Figure 6: Schematic representation of protocol for non-myeloablative preparative regimen to induce tolerance between full MHC haplotype-mismatched cynomolgus monkeys. In attempting to induce transplantation tolerance across the discordant xenogeneic barrier pig-to-cynomolgus monkey through establishment of mixed chimerism, extracorporeal immunoadsorption of monkey blood through a pig liver or immunoaffinity column of a Gal alpha1-3Gal oligosaccharide is performed prior to the pig kidney transplant in an effort to deplete anti-pig antibody. Additional pharmacologic immunosuppressive therapy, e.g. with 15-deoxyspergualin, has been used to inhibit B cell activity and therefore reduce the rate of return of antibody. (From Sykes, M. and Sachs, D.H. (93))

Clear evidence for chimerism amongst lymphoid, myeloid, and monocytic subpopulations, was generally detected first on about day 8, persisting until about day 30. Thereafter the levels of detectable chimerism decreased progressively. However, in recipient animals given allografts and concordant xenografts, transplantation tolerance was induced, as assessed by MLR assays, by monitoring of kidney transplant function, and in one case by acceptance of a full thickness graft of frozen skin from the kidney donor.

More recent studies have attempted to extend this non-myeloablative regimen for production of mixed chimerism in the discordant pig-to-primate combination (92,93). The major addition to previous protocols is the need to remove natural antibodies from the recipient circulation in order to avoid HAR. This has been attempted by extracorporeal perfusion of the monkey's blood either through an isolated pig liver or through specific synthetic oligosaccharide columns and has been carried out immediately prior to kidney transplantation.

Using this regimen, pig kidney grafts have functioned normally for <15 days in cynomolgus monkeys, but have failed from a vascular form of rejection. In addition, there has been only transient evidence for pig cell chimerism, with a low level of pig cells (1-5%) in the peripheral blood.

5.4 The genetically engineered pig

5.4.1. Expression of human complement-inhibiting proteins

Most advances in this field have come from efforts to genetically engineer a pig that expresses one or more of the human complement-inhibiting proteins. Complement-inhibiting proteins are believed, under most circumstances, to block autologous complement but not that of other species (94,95). For example, pig organs express complement-inhibiting proteins that block pig complement, but do not adequately block human complement. It is therefore believed that the development of a pig that expresses human complement-inhibiting proteins on its vascular endothelium will be successful in blocking human complement.

The human complement-inhibiting proteins include CD46 (membrane cofactor protein, MCP), CD55 (decay accelerating factor, DAF), and CD59 (homologous restriction factor). Pigs have been bred that express one or more of these proteins (96-101), but few data are yet available on their efficacy at preventing HAR. Survival of pig hearts expressing MCP and DAF was extended from a few minutes to 30 hours in one baboon. The most encouraging results achieved to-date have been by the Cambridge, UK, group of White and his colleagues (101), who have reported heterotopic pig heart survival for <60 days in one cynomolgus monkey. This model of pig-to-cynomolgus monkey, however, is unusual in that some of the control (non-transgenic) pig hearts survived several days, suggesting that HAR is not uniform in this combination. Investigations in this field are progressing rapidly.

5.4.2. alphaGal 'Knockout'

The second approach with regard to a genetically engineered pig would be to produce a pig that is deficient in alphaGal epitopes, thus leaving no target for human anti-alphaGal antibodies (102). In the pig, Gal alpha1-3Gal is produced by the enzyme alpha1,3galactosyltransferase (alpha1,3GT), which is encoded by a single gene (29). If this gene could be "knocked out" by a technique such as homologous recombination, then an alphaGal-deficient pig would be produced. The only hitherto discovered difference between pigs and humans with regard to the oligosaccharides expressed on the vascular endothelium is the presence of alphaGal in the pig where ABH oligosaccharide is expressed in the human (25) (Table 2). Whether an alphaGal-depleted pig would be a fully viable, healthy pig remains uncertain, but the fact that there are some human subjects who are depleted of ABH antigen (the so-called "Bombay" histo-blood type) who appear to be clinically well in all respects, would suggest that alphaGal-depleted pigs will similarly be healthy.

The "knockout" technique, which requires the manipulation of stem cells, is not yet possible in the pig. Mice, however, have been bred which do not express alphaGal epitopes (103,104). One strain of these mice have certain physical defects in the form of the early development of cataracts (103). In vitro and in vivo studies, however, suggest that the absence of alphaGal exposes the presence of underlying "cryptic" oligosaccharide epitopes against which humans also have antibodies. This approach would not appear so promising as originally hoped.

5.4.3. Competitive glycosylation One alternative approach would be to genetically engineer a pig with an abundance of another oligosaccharide epitope that would "mask" the alphaGal epitope. Suggested candidates have been sialic acid or the H histo-blood group antigen (Table 5) (26,102). This method, involving the microinjection of a gene to express the required oligosaccharide, is possible in the pig. What percentage of alphaGal expression needs to be "masked" before HAR is prevented remains uncertain, but it seems likely that it will be virtually 100%.

Table 5: Fucosylation or sialylation of N-acetyl-lactosamine (ßGal1-4ßGlcNAc-R) can impair the formation of terminal non-reducing alphaGal epitopes

Chemical StructureEpitope Name

*R are glycolipid or glycoprotein carrier molecules anchored in the cell membrane. The presence of fucose or sialic acid represents a steric hindrance for the enzyme activity of the pig alpha(1,3) galactosyltransferase. Added fucose or sialic acid are underlined. (From DKC Cooper, et al. (26))

Good progress in this field has been made by Sandrin et al. (105,106) who have demonstrated in vitro that competition between alpha1,2 fucosyltransferase (H transferase) and alpha1,3GT takes place for the substrate N-acetyllactosamine (Figure 7). H transferase is significantly more successful and the H epitope predominates, reducing the presence of alphaGal to approximately 5% of its original expression.

Figure 7: Biosynthetic pathway for synthesis of Gal alpha1-3Gal. The alpha1,3 galactosyltransferase enzyme adds galactose to N-acetyllactosamine (Galß1-4GlcNAc) to generate Gal alpha1-3Gal. The same substrate can be utilized by transgenically-introduced alpha1,2 fucosyltransferase to produce the H histo-blood group epitope. Gal alpha1-3Gal can also be eliminated by the introduction of alpha-galactosidase, which enables the N-acetyllactosamine substrate to be available again for further fucosylation. (Modified from Sandrin, M.S., et al. (106))

One interesting point is that pigs do, in fact, have the gene for H transferase and express H oligosaccharide epitopes, not on vascular endothelium but in certain other tissues (25). It is therefore essential to ensure that the H transferase produced as a result of the introduction of H cDNA functions at the correct site, and this may prove to be less easy than is immediately obvious.

Unless H epitopes replace the alphaGal epitopes completely, the number of alphaGal epitopes remaining on the vascular endothelium would still make such a pig organ susceptible to HAR. The ultimate solution, therefore, may be to combine expression of H transferase with that of agalactosidase (107) (Figure 7). The remaining epitopes expressing alphaGal will be depleted of the alphaGal by alpha-galactosidase, rendering the N-acetyllactosamine again available for the H transferase.

At the present time, this approach of modifying the donor organ by deletion of alphaGal expression, together with the expression of complement-inhibiting proteins (perhaps in addition to providing some therapy to the recipient) would appear to be the most promising method of overcoming HAR and of being able to apply discordant xenotransplantation clinically.

[Table of Contents ] [Next Section] [Previous Section]