Main

Circe offered Odysseus's men a potion mixed with cheese, honey and wine. And when they had emptied their bowls, they grew pigs' heads and bristles, and they grunted like pigs; but their minds were as human as ever. Credit: AKG LONDON

Homer's story of what Circe achieved with an early formulation like cyclosporine has inspired the name of a biotechnology company that provides pig hepatocytes for ex vivo treatment of fulminant liver failure, and we named a pig retrovirus in her honour. Down the ages there have been many tales of men being transformed into beasts, and princes into frogs. Marie Darrieussecq's Truismes, now available in English translation1, is a modern parable on the porcine fate of a young woman in a ‘massage’ parlour who gradually turns plumper and pinker and who begins to grunt and squeal in a way that initially delights her clients. Billed as a feminist version of Kafka's Metamorphosis, it probably owes as much in its undertones of fascism to Ionesco's Rhinoceros. Conversely, anthropomorphism in animals is an equally popular genre for cautionary tales, ranging from Aesop's Fables to the telling climax in Orwell's Animal Farm when the pigs totter around on their hind legs. It is this second aspect, the humanization of pigs, that I wish to address in this commentary on xenotransplantation — in which tissue is transferred from one species to another — and infection.

Preventing hyperacute rejection

Monkeys are unlikely to provide an adequate supply of clean tissues for xenotransplantation, so most investment is being made in the use of pigs2. Although there has been limited success in transplanting pig islet and neural cells, a major barrier to the ‘discordant’ xenotransplantation of whole organs from pigs into primates is a phenomenon known as hyperacute rejection. This involves the destruction of the vascular endothelium of the donor organ within minutes of exposure to human or monkey blood.

The lysis of endothelial cells occurs through a complement-dependent mechanism following the binding of naturally occurring antibodies to carbohydrate antigens on the glycolipids and glycoproteins of the donor cells. It is akin to the destruction of red blood cells in ABO-mismatched transfusion. The main xeno-antigen recognized on porcine endothelial cells is a galactose α(1-3)-galactose terminal sugar residue (αGal)3. Pigs and most mammals express an α1-3 galactosyltransferase, which places αGal on glycoconjugates. Along with other old-world primates, humans lack this enzyme; we appear to be natural genetic knockouts for the gene encoding galactosyltransferase. Because we are exposed to αGal antigen (for example, on bacteria in the gut) we respond by making antibodies to αGal. Indeed, more than 2 per cent of total human IgM and IgG in the circulation represents αGal antibody4, and this triggers the complement-mediated lysis of xenogeneic cells bearing αGal antigens.

Several strategies are being developed by biotechnology companies to block hyperacute rejection and so help porcine tissues to survive in humans2,5. One approach involves circulating human blood plasma over αGal-immunoadsorbent columns to remove anti-αGal antibodies; unfortunately, anti-αGal levels quickly return to normal in the blood. Another method is to inhibit triggering of the complement cascade by treatment with antibodies to complement components or with excess amounts of a soluble form of complement receptor. A third strategy is to generate transgenic pigs that overexpress a fucosyltransferase that competes with the substrate for αGal attachment so that a fucose group, representing a natural human blood-group antigen, is substituted for αGal. A fourth idea is to knock out the pig gene that encodes α1-3 galactosyltransferase. This would eliminate the main cause of hyperacute rejection in the first place, but the technology for porcine knockouts is not yet available. The fifth strategy is to generate transgenic pigs that express human complement-regulatory proteins. These proteins, CD55 (also known as decay accelerating factor, DAF), CD46 (membrane cofactor protein, MCP-1) and CD59 (protectin) do not prevent anti-αGal antibody binding to endothelial cells to trigger the classical pathway of complement activation; rather, they inhibit downstream steps in the complement cascade so that cell lysis may be prevented. Several transgenic pig herds have been developed that express one or more of the human genes encoding CD55, CD46 and CD59.

Virolysis

Many animal viruses with lipid envelopes are sensitive to inactivation by human complement. Virolysis was first observed with retroviruses6. It was thought to explain in part why we seldom if ever become infected with the retroviruses frequently shed by our animal neighbours, such as cats and mice, even though some of these viruses readily replicate in human cells in culture. At first it seemed that lysis of animal retroviruses by complement occurred without specific antibody, because all human sera showed this activity. But we now know that lysis of retroviruses is triggered by the binding of anti-αGal antibodies in human sera to αGal residues expressed on the viral envelope7,8.

Virus grown in non-primate cells is sensitive to inactivation by fresh human serum, whereas the same virus propagated in human cells is not. Other enveloped viruses grown in animal cells are also sensitive to lysis by human complement, including arenavirus, paramyxovirus, alphavirus and the rhabdoviral pig pathogen, vesicular stomatitis virus8,9. If αGal is on the host cell then the viral envelope becomes sensitive to rapid lysis by human serum. In other words, virus inactivation occurs by precisely the same mechanism as hyperacute rejection of xenografts.

It follows that modifications to make porcine xenografts resistant to hyperacute rejection may also make any enveloped viruses of pigs similarly resistant to lysis. This must be so for the removal of anti-αGal antibodies from the patient's plasma, or for the removal of αGal from the surface of pig cells by competing fucosyltransferase or by knocking out galactosyltransferase function. It might also be the case where pigs are transgenic for human complement-regulatory genes, provided that enough of the human proteins are incorporated into the viral envelope.

Proteins CD46, CD55 and CD59 are present in the HIV envelope10,11,12 and can protect HIV from complement-mediated lysis10,11. Whether the presence of human complement regulatory factors in pig cells changes the serum sensitivity of viruses budding from them still needs to be rigorously tested. But pig endogenous retrovirus released from pig cells is highly sensitive to human complement-mediated lysis, whereas the same virus passaged in human cells lacking αGal immediately becomes resistant13.

The extent to which anti-αGal and complement protect us from infection by animal viruses is unknown. Individuals with deficiencies in the complement pathway tend to be more susceptible to bacterial than viral infections. But the possibility remains that viral zoonoses are more likely to occur in xenotransplantation of transgenic rather than unmodified pig tissue.

Human receptors in transgenic pigs

Another, neglected facet of porcine transgenesis aimed at blocking hyperacute rejection potentially adds considerably to the risk of cross-species infection. Two of the three human complement regulatory proteins are also receptors for human viral pathogens: CD46 is the cell-surface receptor for measles virus14, and CD55 can serve as a binding receptor for Echo and Coxsackie B picornaviruses15,16. Coxsackie B virus causes myocarditis and might endanger the pig heart in an immunosuppressed recipient of a xenograft. Clearly, valuable transgenic herds should not be husbanded by infected staff in case the pigs become infected by measles virus or picornaviruses. But the converse situation worries me more.

Transgenic pigs may provide an opportunity for animal viruses to adapt to a human host range. Human Coxsackie B virus, for example, can be adapted to grow in mice, and in some human cell cultures it increases its infectivity a million-fold by adopting the CD55 receptor16. If pigs were to harbour picornaviruses that use the porcine equivalent of CD55, such viruses might readily adapt to recognize human CD55 in transgenic pigs that express both the native porcine and the human forms of the receptor. These viruses would then be preadapted for transmission to the xenograft recipient and for human-to-human transmission. There is already concern that mice transgenic for the human poliovirus receptor should not escape and become a non-human reservoir for a human pathogen.

Animal morbilliviruses (measles-related viruses such as canine distemper virus and rinderpest virus), too, might become pre-adapted for human transmission in CD46-transgenic pigs. Morbilliviruses are known to jump host species to wreak havoc, as in the recent epidemic in seals and dolphins. In Australia a veterinarian and a stable-hand died after an autopsy of a horse with a new type of morbillivirus infection, which in turn was probably acquired from fruit bats17. However remote, we should be mindful of the additional public health risk of genetically modified pigs that express human virus receptors.

Balancing benefit and risk

It is sometimes said that because pigs and humans have lived together for thousands of years there cannot be new microbes to cross between these species. Notwithstanding pigs as a source of new influenza pandemics, xenotransplantation will potentiate the risk of pig-to-human transfer of viruses not transmitted by the respiratory route. First, the physical barrier is broken by transplanting living porcine tissues or organs into humans. Second, the immunosuppression required to prevent xenograft rejection may allow zoonotic viruses to adapt to human infection. Third, as discussed above, the genetic modification of pigs may allow preadaptation of animal viruses for human infection.

None of these risks is easily quantifiable. We know that human tumour tissue xenotransplanted into immunodeficient mice frequently becomes infected by endogenous, xenotropic mouse retrovirus. Thus pig-to-human viral transfer could also occur by xenografts, as we have found that two of three distinct porcine retroviruses can infect some human cells in culture13,18. Moreover, because these viruses are endogenous (that is, their genomes are carried as Mendelian traits in the pig chromosomes), raising virus-free pigs will not be easy. Horizontally transmitted viruses may be more readily excluded through screening, provided that we know of their existence. But we may not know; only a few months ago, a new porcine virus related to human hepatitis virus E was reported19. How many more potential pathogens remain to be discovered?

To the individual xenograft recipient, the benefit of a successful transplant will almost certainly outweigh the risk of any subsequent untoward effects of infection by a pig virus. The transplant surgeon can therefore sleep with a clear conscience that he is helping his patient. We microbiologists, on the other hand, wish to alert society to the more remote but possible risk of setting off a new human epidemic. After all, HIV-1 probably began by cross-species transfer (HIV-2 certainly did). I would not expect the pig endogenous retroviruses to present such a problem, because they grow to only low titres in human cells and so are unlikely to be pathogenic. But we cannot dismiss virus adaptation or recombination with other retroviruses in the new host.

We need a Hippocratic ethic for community and public health, at least to do no harm. Fred Murphy20 remarked that the acceptability of xenotransplantation at the level of social risk has followed a rollercoaster course. Personally, I am in favour of moving forward — cautiously — neither prohibiting human xenotransplantation nor rushing headlong like the Gadarene swine over this precipice of new technology.

We need to set in place the appropriate system to monitor possible virus transmission in human xenotransplant trials. Before proceeding, however, we should find out as much as possible about whether xenogeneic infection has occurred in individuals already exposed to live porcine tissue. No one wishes to promote an unforeseen Odyssey of pig viruses journeying through a xenograft recipient to his or her contacts and onwards to the population at large.