Xenotransplantation: Engineering Pig Organs for Humans

Xenotransplantation — transplanting living organs from one species into another — stopped being a thought experiment in 2022 and became a clinical reality with a body count, a survival record, and three United States Food and Drug Administration-cleared trials. A man in New Hampshire lived 271 days with a gene-edited pig kidney before it failed, the longest anyone has survived on an animal organ. The reason this is suddenly working is not surgical. It is genetic. Roughly 100,000 Americans wait for a transplant while a pig’s anatomy is, conveniently, close enough to ours to function. The hard part is that a pig organ, dropped into a human bloodstream untouched, is destroyed within minutes. Closing that gap took dozens of precise CRISPR edits per donor animal, decades of immunology, and a manufacturing pipeline that looks more like semiconductor fabrication than farming.

What this covers: why the organ shortage forces the issue, the three waves of immune rejection and the sugar molecules that trigger them, the specific gene edits in today’s donor pigs, the recent human cases and trials, and the engineering pipeline that produces a transplantable organ.

Context and Background

The transplant waiting list is a slow-motion catastrophe. In the United States alone more than 100,000 people are listed at any moment, the large majority needing a kidney, and roughly 17 die each day before an organ arrives. Deceased-donor supply is structurally capped — it depends on a narrow set of deaths that leave organs viable — and living donation cannot scale to the gap. This is the demand that has kept xenotransplantation alive as a research program for over a century, through repeated, fatal failures with primate and pig organs.

The economics underline the urgency. End-stage renal disease is one of the most expensive chronic conditions in medicine: a patient on dialysis costs the health system on the order of tens of thousands of dollars per year and suffers a quality of life far below that of a transplant recipient, while a functioning kidney transplant is both cheaper over time and dramatically better for the patient. Every year a patient spends waiting is a year of accumulated cardiovascular damage from dialysis, declining transplant candidacy, and rising mortality risk. The waiting list, in other words, is not a queue that eventually clears — it is a population in which a meaningful fraction will die or become too sick to transplant before their number comes up. That structural shortfall, not scientific curiosity, is what justifies the risk of putting an experimental animal organ into a human being.

Why pigs, and not our closer relatives the baboons or chimpanzees? Three reasons. Pig organs are anatomically and physiologically close to human size and function. Pigs breed quickly with large litters, so a donor line can be propagated. And primates carry a high zoonotic risk and raise acute ethical objections, while pigs are already raised for food at industrial scale. The catch is immunological distance: that same evolutionary gap is exactly what makes a pig organ look unmistakably foreign to the human immune system.

The modern field rests on two enabling technologies. The first is CRISPR-Cas9, which made it cheap and precise to knock out pig genes and insert human ones — and, critically, to edit many sites at once. The second is the decades-long immunology that identified which molecules to edit. Companies such as Revivicor (a United Therapeutics subsidiary) and eGenesis turned that knowledge into engineered donor pig lines. For background on the editing tools themselves, see our explainer on prime editing. The U.S. National Institutes of Health summarizes the organ shortage data that drives the whole effort.

It is worth being honest about how long and how badly this failed before it worked. The first recorded animal-to-human transplants date to the early twentieth century, and the modern era includes Keith Reemtsma’s chimpanzee kidney transplants in the 1960s and the 1984 “Baby Fae” case, in which an infant received a baboon heart and survived 21 days. Every one of these attempts collapsed against the same wall: an immune system evolved to obliterate non-self tissue, with cross-species tissue the most non-self target of all. What changed was not the surgery — vascular anastomosis has been routine for decades — but the ability to change the donor. The gene-editing era reframed the entire problem from “how do we suppress the human response” to “how do we make the organ less of a target in the first place.” That reframing is the whole story of why 2022–2026 looks different from the preceding hundred years.

It is also worth naming why the pre-CRISPR era hit a ceiling. Earlier groups could knock out a single pig gene by homologous recombination in cell culture, but the process was slow, inefficient, and effectively limited to one or two changes at a time. You could make a GGTA1-knockout pig — and people did, in the early 2000s — but you could not realistically stack the antigen knockouts, human transgenes, and retrovirus inactivations that a clinical organ requires on top of one another. CRISPR collapsed that constraint. Suddenly a single multiplexed editing campaign could target many loci in one founder cell line, and the question stopped being “can we make this edit” and became “which edits do we actually need, and how do we validate that we made them cleanly.” That is the shift from a genetics curiosity to an engineering discipline.

How Xenotransplantation Works: The Engineering Problem

Xenotransplantation works by taking an organ from a pig whose genome has been edited to remove the molecular flags that mark it as foreign and to add human regulatory proteins that calm the recipient’s immune and clotting systems. The organ is implanted using ordinary transplant surgery; the engineering happens upstream, in the donor pig, plus an aggressive immunosuppression regimen in the patient.

Figure 1: The end-to-end xenotransplantation pipeline, from an engineered donor pig line to a transplanted human recipient under immunosuppression.

Figure 1 traces the path: a designated pathogen-free pig from a gene-edited founder line is raised in a clean facility, its organ is procured under sterile conditions, perfused and shipped, then transplanted into a recipient who receives both standard immunosuppressants and, in current protocols, an anti-CD40/CD154 costimulation blocker the standard kidney-rejection drugs do not include. Each stage is a control point with its own failure modes.

The upstream half of that pipeline is closer to pharmaceutical manufacturing than to agriculture. A donor pig is not a farm animal; it is a biological product. Founder animals are created by editing pig cells and producing pigs by cloning (somatic-cell nuclear transfer) or by editing embryos, then validated by whole-genome sequencing to confirm the intended edits landed and no catastrophic off-target changes occurred. From a validated founder line, animals are bred and raised inside designated pathogen-free (DPF) facilities — closed, biosecure barns with filtered air, controlled feed, caesarean-derived herds, and continuous screening for a defined list of viruses, bacteria, and parasites that could ride into an immunosuppressed recipient. United Therapeutics, for example, built a dedicated facility expressly for producing clinical-grade donor organs at scale. The organ procurement, preservation, and transport then have to meet the same sterility and ischemia-time constraints as any human transplant. The point is that the “factory” producing a transplantable pig organ is a regulated, monitored, GMP-adjacent operation — and any break in that chain, not just the genetics, can disqualify an organ.

Cloning the founder: somatic-cell nuclear transfer

The way you actually turn an edited cell into an edited pig is worth detailing, because it explains why the donor animal is a manufactured product rather than a bred one. The dominant route is somatic-cell nuclear transfer (SCNT) — the same cloning technique that produced Dolly the sheep. The workflow runs roughly like this: a population of pig fibroblasts (skin cells) is grown in culture and subjected to the full CRISPR editing campaign, knocking out the target antigen genes, knocking in the human transgenes, and inactivating the retroviral copies. Those edited cells are then screened and sequenced so that a single, fully validated clone — one cell line in which every intended edit is confirmed and nothing catastrophic happened elsewhere in the genome — can be chosen as the genetic template. The nucleus of a cell from that validated clone is transferred into a pig egg whose own nucleus has been removed, the reconstructed embryo is stimulated to begin dividing, and it is implanted into a surrogate sow. The piglets born are genetic copies of the validated founder cell, edits and all.

This matters for two reasons. First, it means the genome can be characterized exhaustively before any animal is born — you are not gambling on what an embryo turns into, you are copying a cell you have already sequenced. Second, it makes consistency tractable: once you have a validated founder, the breeding herd descends from a known, fixed genotype rather than from the random reassortment of sexual reproduction. That reproducibility is exactly what a regulator wants to see, because it means every organ coming off the line is, genetically, the same product.

The compatibility gap is mostly biochemical, not anatomical

A surgeon can plumb a pig kidney into a human renal artery, vein, and ureter without much novelty — the vessels are roughly the right caliber and the anastomoses are routine. The hard problem sits at the molecular interface between donor blood-vessel lining (endothelium) and recipient blood. The pig endothelium is studded with carbohydrate antigens and signaling proteins that the human immune and coagulation systems read as “destroy this.” No suture technique fixes that. Only changing what the cells display does.

The endothelium deserves emphasis because it is the actual battlefield. It is the single continuous layer of cells lining every blood vessel in the graft, and it is the only pig tissue in direct, constant contact with human blood. Everything the recipient’s antibodies, complement proteins, platelets, and immune cells encounter first, they encounter on that endothelial surface. This is why almost every gene edit in a donor pig is aimed at what the endothelium displays or secretes: the antigens it presents, the complement regulators it carries, the anticoagulant proteins it expresses. You are not really engineering “a pig kidney” so much as engineering the thin cellular lining where two species’ bloodstreams meet.

Editing moves the problem upstream into the genome

The strategic shift that made the field clinical was relocating the fight from the operating room to the embryo. Instead of trying to suppress a human immune response after the fact — which decades of failure showed is not enough on its own — you remove the triggers before the organ exists. That is why a single donor pig may now carry dozens of edits. The organ arrives “pre-cleared” of the worst antigens and pre-loaded with human brakes on rejection and clotting.

Two design philosophies, both in trials

The two leading programs embody different bets. United Therapeutics’ UKidney uses a pig with 10 gene edits — three xenoantigen knockouts, six human transgenes, and one growth-control edit — a minimalist set aimed at what is provably necessary. eGenesis’ candidate (EGEN-2784) uses 69 edits, layering many human regulatory transgenes on top of the antigen knockouts and adding dozens of porcine endogenous retrovirus inactivations. More edits may mean a more robust organ; it also means a more complex animal to validate. The trials will adjudicate which philosophy survives contact with patients.

There is a deeper engineering tension hiding in that edit-count gap. A multiplex CRISPR campaign is not free: each cut site is a chance for an off-target edit, a chromosomal rearrangement, or an unintended disruption of a nearby gene, and the more simultaneous double-strand breaks you induce, the higher the risk of translocations between cut sites. eGenesis’ approach of inactivating tens of PERV copies in one go is genuinely impressive multiplex engineering, but it also means the resulting animal needs far more whole-genome validation to prove nothing else broke. United Therapeutics’ bet is that a smaller, well-characterized edit set is easier to manufacture reproducibly and easier to defend to a regulator. Neither bet is obviously right. The clinical endpoints — graft survival, cause of failure, and any genomic instability detected in the donor line — are what will settle it, and that is precisely why running both philosophies through trials in parallel is good for the field.

It is worth being specific about why the translocation risk scales the way it does. CRISPR-Cas9 works by introducing a double-strand break at a targeted site, after which the cell’s own repair machinery stitches the DNA back together — most commonly through non-homologous end joining, which is what produces the small disruptive insertions and deletions used to knock a gene out. When you make one cut, the cell repairs one break. When you make dozens of cuts simultaneously — as you must to inactivate tens of PERV copies in a single campaign — the cell is presented with many free DNA ends at once, and the repair machinery can mistakenly join the wrong ends together. That mis-joining is a chromosomal translocation: a piece of one chromosome welded onto another. Translocations are exactly the kind of lesion that can disrupt a distant gene or destabilize the genome, and they are precisely why a heavily multiplexed founder line cannot be trusted on the basis of “we confirmed the target edits.” You also have to confirm, by whole-genome sequencing, that no large-scale rearrangement crept in between the cut sites — and that verification burden grows with every additional simultaneous break.

The Immunology of Rejection: Three Waves and Three Sugars

Rejection of a pig organ is not one event but a sequence, each phase driven by a different arm of the immune system and operating on a different timescale. Understanding the three waves explains exactly why each gene edit exists.

Figure 2: The three temporal waves of cross-species transplant rejection — hyperacute (minutes), acute (days to weeks), and chronic (months and beyond) — each with a distinct immune mechanism.

Hyperacute rejection and the alpha-Gal catastrophe

The first wave is hyperacute rejection, and it is fast — minutes to hours. Humans, apes, and Old World monkeys lost a functioning GGTA1 gene millions of years ago, so we do not make the sugar galactose-alpha-1,3-galactose, known as alpha-Gal. Pigs do, decorating their endothelium with it. Because alpha-Gal also appears on gut bacteria, every human is awash in pre-formed antibodies against it from birth.

When a pig organ hits human blood, those antibodies bind alpha-Gal instantly and trigger the complement cascade — a chain of plasma proteins that punches holes in cell membranes and recruits clotting. The mechanism is worth spelling out because it explains the speed. Antibody binding activates the classical complement pathway; complement proteins assemble into the membrane attack complex that lyses the pig endothelial cells lining the graft’s blood vessels; the exposed, dying endothelium becomes intensely pro-coagulant; platelets and fibrin pile in; and the organ’s microvasculature clots off from the inside. The organ turns black and thromboses within minutes to a couple of hours. Knocking out GGTA1 removes alpha-Gal and abolishes the single largest cause of hyperacute rejection. This is the non-negotiable first edit; without it, nothing else matters. The very first GGTA1-knockout pigs were produced around 2003, well before CRISPR, but doing it cleanly and combining it with other edits is what CRISPR made practical.

It helps to walk the classical complement cascade one step further, because the structure of it is what tells you where to intervene. The pre-formed anti-Gal antibody, once bound to alpha-Gal on the pig endothelium, recruits the complement protein C1, which kicks off a proteolytic chain. C1 cleaves C4 and C2, and their fragments assemble into the C3 convertase, an enzyme that splits C3 — the central amplification node of the entire system. A single C3 convertase can cleave many C3 molecules, so the signal multiplies explosively. Some of those C3 fragments tag the cell surface for destruction by phagocytes, while others combine with the convertase to form the C5 convertase, which cleaves C5 and launches the terminal pathway. The terminal components C5b, C6, C7, C8, and multiple copies of C9 then assemble into the membrane attack complex (MAC) — a literal pore drilled through the endothelial cell membrane that lets ions and water rush in until the cell bursts. Understanding this chain explains why the human complement-regulator transgenes (discussed below) are positioned where they are: they act at the convertase step, before the cascade amplifies, which is the only efficient place to shut a runaway enzymatic loop down.

The other two sugars: Neu5Gc and Sd(a)

Removing alpha-Gal exposes two more carbohydrate antigens that humans also react to. The enzyme CMAH produces Neu5Gc (N-glycolylneuraminic acid), a sialic acid humans cannot synthesize — we, too, carry a broken CMAH gene — so many people have anti-Neu5Gc antibodies. The enzyme B4GALNT2 produces the Sd(a) antigen, another human-reactive glycan. Knocking out GGTA1, CMAH, and B4GALNT2 together — the “triple knockout” or 3KO pig — eliminates all three major xenoantigens and dramatically reduces human antibody binding. The 3KO genotype is the carbohydrate foundation of every clinical donor pig today.

There is a clinical subtlety in the Neu5Gc story worth flagging. Because humans cannot make Neu5Gc but do ingest it in red meat, the antibody response to it varies between individuals — some recipients carry high titers of anti-Neu5Gc antibody, others much less. That variability means the CMAH knockout matters more for some patients than others, and it complicates pre-transplant cross-matching: the standard panel-reactive antibody tests developed for human-to-human transplants do not directly measure a recipient’s reactivity to pig glycans. Building reliable cross-match assays for the xeno setting — tests that predict whether a given patient’s serum will attack a given donor pig — is an active, underappreciated part of making these trials safe.

Acute and chronic rejection: the role of complement and coagulation

The second wave, acute humoral and cellular rejection, unfolds over days to weeks. Even with the sugars gone, residual and newly formed antibodies and the recipient’s T cells attack the graft. Two distinct problems compound here. The humoral arm involves “non-Gal” antibodies — those directed at pig proteins and at the remaining minor glycans — which is exactly why the complement-regulator transgenes matter even after the antigens are knocked out. The cellular arm is driven by T cells, and pig and human cells present antigen to each other through both direct and indirect pathways; this is the part that conventional immunosuppressants and the newer anti-CD40/CD154 costimulation blockade are fighting. A pig’s own complement regulatory proteins do not efficiently restrain human complement, so without added human regulators the cascade runs comparatively unchecked on pig tissue.

The costimulation-blockade point deserves unpacking, because it is the part of the immunosuppression regimen that distinguishes xeno protocols from ordinary kidney transplants. A T cell needs two signals to fully activate against a graft: the first is recognition of the foreign antigen itself, and the second is a “costimulatory” handshake, most importantly the CD40–CD154 (CD40 ligand) interaction between antigen-presenting cells and T cells. Block that second signal and the T cell, instead of mounting a full attack, can be steered toward anergy or tolerance — it sees the antigen but does not receive the go-ahead to escalate. The anti-CD40/CD154 agents in current xeno protocols target precisely this pathway, and they exist in the regimen because the cross-species T-cell response is fiercer than the within-species one; the conventional drug cocktail that suffices for a human kidney is not enough to hold a pig graft. The cost of that extra suppression is extra vulnerability to infection, which is one of the central trade-offs the trials are watching.

The coagulation mismatch: a clotting system that cannot talk to itself

The third wave, chronic rejection, plays out over months — a grinding combination of low-grade immune injury, thrombotic microangiopathy (microscopic clots in the graft’s small vessels caused by a pig-human coagulation mismatch), and fibrosis that slowly scars the organ. The coagulation mismatch deserves its own treatment because it is, mechanistically, one of the most stubborn problems in the whole field, and it is not an immune problem at all — it is a protein-compatibility problem.

In a healthy human vessel, the endothelium actively prevents clotting through a feedback loop centered on thrombomodulin and protein C. When the clotting cascade starts to generate thrombin, thrombin binds to thrombomodulin on the endothelial surface; that thrombomodulin–thrombin complex then activates protein C (greatly accelerated by the endothelial protein C receptor, EPCR), and activated protein C in turn shuts down further thrombin generation. It is an elegant self-limiting brake: the very signal that says “start clotting” is harnessed to say “now stop.” The problem is that pig thrombomodulin does not activate human protein C efficiently. The molecular fit between the pig protein and the human substrate is poor, so the brake barely engages. On a pig graft sitting in human blood, thrombin generation is therefore poorly restrained, and the natural anticoagulant feedback that keeps a human’s own vessels patent is effectively missing.

The downstream consequence is thrombotic microangiopathy — a storm of microscopic clots forming in the graft’s smallest vessels, consuming platelets, shearing red blood cells, and progressively choking off the organ’s blood supply. This can throttle a graft even when the immune response is reasonably well controlled, which is what makes it so insidious: you can win the antigen and complement battles and still lose the organ to clotting. This is precisely why human thrombomodulin (THBD) and EPCR are standard transgenes in clinical donor pigs — they reinstall a human-compatible anticoagulant brake directly onto the pig endothelium so that the recipient’s thrombin can engage a protein C system that actually works. Human proteins added to the pig are aimed squarely at these later waves, as the next section details.

The Gene Edits: From Three Knockouts to the 63-Edit Pig

The donor pig genome is engineered along three axes: remove antigens, add human regulators, and silence retroviruses. Figure 3 groups the edits by purpose.

Figure 3: The three categories of edits in a gene-edited pig organ — xenoantigen knockouts, human protective transgenes, and porcine endogenous retrovirus inactivation.

Knockouts: deleting what the human immune system targets

The knockouts remove triggers. The carbohydrate trio — GGTA1, CMAH, B4GALNT2 — kills the antigen problem described above. A fourth common knockout disables the pig growth hormone receptor (GHR), so the organ does not keep growing to pig size inside a human; an untouched pig kidney would outgrow the abdominal cavity. United Therapeutics’ UKidney describes exactly this set: three antigen knockouts plus one growth-control edit.

The growth-control edit is a nice illustration of how cross-species biology forces edits that have nothing to do with rejection. Pigs reach a body mass many times that of a human, and their organs are programmed to scale accordingly under the influence of growth hormone. Drop an unmodified pig kidney into a human recipient and it will, over months, keep responding to growth signals and enlarge toward pig proportions — a mechanical problem in a human abdomen quite apart from any immune issue. Disabling the growth hormone receptor uncouples the organ from that signal, so it settles at a size appropriate to its new host. It is a reminder that “engineering an organ for a human” is not only about making it invisible to the immune system; it is also about making its physiology behave inside a body it was never evolved for.

Knock-ins: human transgenes that brake rejection and clotting

The knock-ins add human control proteins so the recipient’s regulatory machinery actually recognizes the graft. They cluster by function:

• Complement regulators — human CD46 (membrane cofactor protein) and CD55 (decay-accelerating factor, DAF) sit on the pig endothelium and shut down the human complement cascade locally, blunting acute humoral rejection. Mechanistically, CD55 accelerates the decay of the C3 and C5 convertases — it pries apart the very enzymes that amplify the cascade — while CD46 acts as a cofactor that helps a plasma protease chop up deposited C3b and C4b fragments, marking them for inactivation. Both work at the convertase step, which is why they are so effective: they choke the cascade at its amplification node rather than trying to mop up the membrane attack complex after it has already formed.
• Coagulation regulators — human thrombomodulin (THBD/TBM) and endothelial protein C receptor (EPCR) correct the pig-human clotting mismatch that causes thrombotic microangiopathy, reinstalling the human-compatible protein C brake described in the previous section.
• Anti-inflammatory and anti-phagocytic signals — transgenes such as CD47 (a “don’t eat me” signal that engages the human macrophage SIRPα receptor and tells phagocytes to leave the cell alone), HO-1 (heme oxygenase-1, a cytoprotective enzyme that dampens oxidative stress and inflammation), and A20 (a regulator that restrains inflammatory NF-κB signaling and protects cells from programmed death) dampen inflammation and cell death.

United Therapeutics’ 10-edit pig adds six human transgenes; eGenesis’ 69-edit design layers in many more, which is most of the difference between the two edit counts.

It helps to do the arithmetic on the eGenesis number, because “69 edits” sounds mysterious until you decompose it. Start with the three carbohydrate-antigen knockouts. Add the growth-control and a handful of human regulatory transgenes spanning complement, coagulation, and inflammation control. Then add on the order of 59 PERV inactivations — each integrated retroviral copy disabled at its catalytic site. Three plus the transgene set plus roughly 59 retroviral edits lands you in the high-60s. The headline number, in other words, is dominated by retrovirus silencing, not by an exotic immune strategy; the immunological edits are broadly similar in kind to the 10-edit pig’s, just somewhat more numerous. Understanding that decomposition is the difference between treating the edit count as marketing and reading it as an engineering bill of materials.

PERV inactivation: the retrovirus in the genome

The third axis is the most distinctive. Pig genomes carry dozens of copies of porcine endogenous retroviruses (PERVs) woven into their DNA — relics of ancient retroviral infections that became permanently integrated into the germline. They cannot simply be bred out, because they are inherited like any other gene, present in every cell of every pig. Three PERV subtypes exist; PERV-A and PERV-B are present in all pigs and can infect human cells in culture, while PERV-C is pig-tropic but can recombine with PERV-A into a more infectious form. In the lab, PERVs can infect human cells, raising the spectre of a new cross-species infection seeded by the transplant itself — a risk that is unusual because it is not just a risk to the patient but, in the worst case, a public-health risk if a novel retrovirus established itself in a human population. In 2015 the Church lab at Harvard used CRISPR to inactivate all 62 PERV copies in pig cells at once — at the time the most extensive multiplex edit ever reported — and in 2017, with eGenesis, produced live pigs with PERVs inactivated. eGenesis’ clinical pig carries on the order of 59 PERV inactivations; combined with its antigen knockouts and human transgenes, that is how the edit count reaches the high-60s. United Therapeutics’ program treats PERV risk primarily through screening and a designated pathogen-free herd rather than inactivating every copy — a genuine strategic divergence in the field. The U.S. FDA’s guidance on xenotransplantation source animals makes pathogen control a regulatory requirement, not just a precaution.

The way the Church-lab campaign actually inactivated all those copies is instructive, because it exploited a structural feature of the retroviral genome. Every PERV copy, regardless of subtype, shares a highly conserved sequence in its pol gene, which encodes the reverse transcriptase the virus needs to replicate. By designing CRISPR guide RNAs against that conserved pol sequence, a single pair of guides could target dozens of integrated copies simultaneously, disabling the catalytic machinery in all of them at once. That is what made the feat possible: you are not designing 62 separate edits, you are designing one edit that hits 62 places. The downside — and it brings us back to the multiplex risk discussed earlier — is that inducing dozens of simultaneous double-strand breaks scattered across the genome is exactly the scenario most likely to produce translocations and to stress the cell, which is why the early high-copy edits showed reduced cell viability and why producing healthy live animals from the most heavily edited lines took additional years of refinement. The biological reason these viruses cannot just be screened away rather than edited is that they are vertically inherited: unlike an infection that a clean herd can exclude, an endogenous retrovirus is written into the germline of every pig, so the only way to remove the genetic risk entirely is to disable the sequences themselves.

AI protein design is starting to touch the problem

The edit lists above were assembled by classical genetics — find the offending gene, knock it out or swap in the human ortholog. The frontier is computational: using AI structure-prediction and generative protein design to engineer transgene variants that suppress human complement or coagulation more efficiently than the wild-type human protein, or to design tolerance-inducing molecules from scratch. This connects xenotransplantation to the broader wave of AI-driven protein design reshaping biotech. It is early, but the long-run trajectory points at donor pigs whose human transgenes are themselves optimized in silico.

Trade-offs, Gotchas, and What Goes Wrong

The clinical record so far is sobering, and honesty about it matters more than hype. The hard limits are biological and they are not fully solved.

Rejection still wins eventually. Every recipient to date has either died or had the organ removed. David Bennett received the first gene-edited pig heart in January 2022 at the University of Maryland and lived about two months; the heart was later found to carry porcine cytomegalovirus (pCMV), an infection that may have contributed to its failure. Lawrence Faucette received a second pig heart in September 2023 and lived about 40 days before rejection set in. On the kidney side, Tim Andrews’ 271-day survival at Mass General Brigham is the record — and it still ended in rejection and a return to dialysis. These are real, hard-won data points, not failures of nerve.

The Bennett case in particular became a teaching case for the whole field, and it is worth dwelling on what it taught. The donor heart carried porcine cytomegalovirus at a level the pre-transplant screening assay failed to detect — the test used at the time looked for active viral replication and could miss a latent infection sitting quietly in the tissue. Under the profound immunosuppression Bennett required, that latent virus reactivated, and the resulting viral load is thought to have driven endothelial damage and contributed to the graft’s decline. The direct, concrete lesson was that screening must be sensitive enough to catch latent as well as active infection, and the assays used in subsequent cases were upgraded accordingly. It is a clean example of how this field learns: not from theory, but from autopsy.

Latent pig viruses are a recurring threat. The pCMV episode shows that PERV is not the only zoonotic concern. Designated pathogen-free (DPF) herds and highly sensitive screening assays are now mandatory, but a virus that evades the assay can ride into an immunosuppressed patient. The screening test that missed pCMV in Bennett’s donor was upgraded before later cases — a direct lesson learned.

The immunosuppression is heavier than for human grafts. Current protocols use anti-CD40/CD154 costimulation blockade on top of conventional drugs, leaving patients more exposed to infection and the side effects of profound immune suppression. The graft you save can be undone by the pneumonia you cause.

More edits is not obviously better. Each additional edit is another way to introduce off-target effects, another genotype to validate, and another variable confounding why a given organ failed. The 10-edit versus 69-edit divergence is an open empirical question, not a settled one. And growth regulation, coagulation matching, and chronic injury remain only partially addressed by the current transgene sets. The validation burden compounds the problem: when a heavily edited organ fails, untangling whether the cause was rejection, infection, clotting, or some unintended consequence of an edit is genuinely hard, and a more complex genome means more candidate explanations to rule out.

Physiology, not just immunology, can fail. Even a perfectly accepted pig kidney has to do a pig kidney’s job inside a human body, and the two species differ in blood pressure, in the hormones that regulate salt and water, and in the renin-angiotensin signaling that links kidney to circulation. A pig organ may handle human electrolyte and fluid balance imperfectly, and pig proteins it secretes — erythropoietin, for instance — may not interoperate cleanly with human receptors. These cross-species physiological mismatches are subtler than rejection but real, and they are harder to fix with gene edits because they involve whole signaling systems rather than single antigens.

The ethics and consent bar is high. Every recipient so far has been critically ill and ineligible for a conventional transplant, which is the only setting in which the risk is justifiable — but it also means the early data come from the sickest possible patients, biasing outcomes downward and making it hard to separate xenograft failure from the patient’s underlying disease. Informed consent for a first-in-human experimental organ, with an unknown infectious-risk profile that could in principle extend to contacts, is genuinely difficult to obtain well. There is a structural tension here that no protocol fully resolves: the patients sick enough to ethically justify the risk are also the patients least likely to survive long enough to show what the organ can really do, which means the early survival numbers systematically understate the technology’s ceiling even as they accurately report its current floor.

Practical Recommendations: What to Watch

For anyone tracking this field — clinically, scientifically, or as an investor — the signal is in the trial data, not the press releases. The 2025–2026 FDA clearances moved xenotransplantation from heroic one-off compassionate-use cases into structured trials with endpoints, which is where real evidence will accumulate. Watch graft survival at 6 and 12 months, the cause of every graft loss (rejection versus infection versus thrombosis), and whether any PERV or pCMV transmission is ever detected in a recipient.

Be precise about clinical status. As of mid-2026, no pig organ is an approved therapy; every case is investigational. Treat any claim of a “cure for the organ shortage” as premature.

A practical watch-list:

• United Therapeutics EXPAND trial — UKidney (10-edit Revivicor pig); first transplant November 2025 at NYU Langone, initial cohort of six end-stage renal disease patients, expandable toward 50.
• eGenesis EGEN-2784 — 69-edit pig; Phase 1/2/3 design evaluating safety and efficacy at 24 weeks in 33 end-stage kidney disease patients.
• Graft-loss cause analysis — whether failures are immune, infectious, or coagulation-driven, since each points to a different next edit.
• Zoonosis surveillance — any signal of PERV or pCMV transmission, the field’s existential risk.
• Heart program progress — UHeart (10-edit) following the kidney trials, given the harder logistics of cardiac xenografts.
• AI-designed transgenes — the first donor pig carrying a computationally optimized human regulatory protein.

Figure 4 ties the strands together, mapping each category of edit to the rejection wave it is designed to defeat — the conceptual core of why a modern donor pig carries the genome it does.

Figure 4: How each class of edit maps to the rejection mechanism it counters — antigen knockouts against hyperacute rejection, complement and coagulation transgenes against acute and chronic injury, and PERV inactivation against zoonotic risk.

The most useful mental model for an outside observer is that of a layered defense, each layer addressing a different timescale of attack. The antigen knockouts buy you the first minutes by removing the targets of pre-formed antibody. The complement transgenes hold the line over the following days by throttling the cascade that residual and newly formed antibodies still trigger. The coagulation transgenes contest the months-long war of attrition by reinstalling the anticoagulant brake the pig endothelium cannot provide on its own. The costimulation blockade and conventional immunosuppression fight the T-cell response across all of those windows. And the PERV strategy — whether by editing or by screening — addresses a risk that operates on no fixed timescale at all but whose worst case extends beyond the patient. Read the watch-list above through that lens and each data point tells you something specific: a hyperacute loss implicates the antigen layer, a thrombotic loss implicates the coagulation layer, an infectious loss implicates the herd and the assays, and a slow chronic decline implicates the parts of the problem no current edit has fully solved.

Frequently Asked Questions

Why use pig organs instead of organs from monkeys or apes?

Pigs win on three counts. Their organs are close to human size and physiology, they breed fast in large litters so a donor line can be scaled, and they carry lower zoonotic and ethical burdens than primates. Crucially, that same evolutionary distance is why pigs lack — or can be edited to lack — the antigens humans attack, whereas primate organs trigger their own severe rejection. Primates also pose a higher risk of transmitting dangerous viruses to humans, which effectively rules them out as a practical organ source.

What is the alpha-Gal antigen and why does it matter?

Alpha-Gal (galactose-alpha-1,3-galactose) is a sugar that pigs put on their cell surfaces but humans cannot make, because we lost the GGTA1 gene millions of years ago. Since alpha-Gal also appears on common gut bacteria, every human carries pre-formed antibodies against it. Those antibodies bind a pig organ within minutes and trigger hyperacute rejection, destroying it almost immediately. Knocking out GGTA1 removes alpha-Gal and is the single most important gene edit in xenotransplantation.

What are PERVs and why are they knocked out?

PERVs are porcine endogenous retroviruses — viral sequences permanently embedded in the pig genome and inherited like ordinary genes, so they cannot be bred out. In laboratory studies PERVs can infect human cells, raising the risk that a transplant could seed a new cross-species infection. Using CRISPR, researchers have inactivated dozens of PERV copies at once in donor pigs by targeting a conserved sequence shared across all copies. Some programs inactivate them genetically; others rely on screening and pathogen-free herds. Either way, PERV control is a regulatory requirement.

How long have humans lived with a gene-edited pig organ?

The current record is 271 days, set by Tim Andrews, who received a gene-edited pig kidney at Mass General Brigham before it was removed in late 2025 and he returned to dialysis. On the heart side, David Bennett lived about two months in 2022 and Lawrence Faucette about 40 days in 2023. Every case so far has ended in graft loss, which is why these remain investigational trials rather than approved therapies as of 2026.

Is xenotransplantation an approved treatment in 2026?

No. As of mid-2026 every pig-organ transplant is investigational — performed under FDA-cleared clinical trials or earlier compassionate-use authorizations, not as approved care. United Therapeutics’ EXPAND trial and eGenesis’ EGEN-2784 trial are enrolling end-stage kidney disease patients, but approval depends on those trials demonstrating durable graft survival and safety. Anyone describing it as a routine cure is overstating where the science stands.

What is the difference between the 10-edit and 69-edit pigs?

They reflect two design philosophies. United Therapeutics’ 10-edit UKidney pig uses a minimal set — three xenoantigen knockouts, six human transgenes, and one growth-control edit — betting on what is provably necessary. eGenesis’ 69-edit pig layers in many more human regulatory transgenes and adds roughly 59 PERV inactivations, betting that a more thoroughly engineered organ will last longer. The bulk of that 69 is retrovirus silencing rather than novel immune strategy. Whether the extra edits improve outcomes or just add complexity is exactly what the trials will reveal.

Why is a pig kidney the first organ in human trials and not a heart?

Kidneys are the logical entry point for several reasons. The kidney waiting list is by far the largest, so the unmet need is greatest there. A kidney failure is also recoverable in a way a heart failure is not: if a transplanted pig kidney is rejected, the patient can return to dialysis, as Tim Andrews did after 271 days, whereas a failing transplanted heart is immediately life-threatening with no fallback. That safety margin makes the kidney the ethically defensible first organ. Heart programs such as United Therapeutics’ UHeart are advancing in parallel, but they follow the kidney trials precisely because the stakes of a graft failure are so much higher.

Why does the pig-human clotting mismatch matter so much?

Because it can destroy a graft even when the immune response is controlled. In human vessels, the endothelial protein thrombomodulin teams up with protein C and EPCR to switch off excess clotting — a self-limiting brake. Pig thrombomodulin does not activate human protein C efficiently, so on a pig graft that brake barely works, and microscopic clots build up in the small vessels (thrombotic microangiopathy) and slowly strangle the organ. That is why human thrombomodulin and EPCR are standard transgenes: they reinstall a clotting brake the recipient’s blood can actually engage.

Further Reading

• Prime editing explained: the next generation of CRISPR — the precision gene-editing tools underpinning donor-pig engineering.
• AI protein design and RFdiffusion — how generative models are starting to design the human transgenes used in xenografts.
• How CAR-T cell therapy works — another frontier where engineering the immune system is the whole game.
• External: U.S. FDA, Xenotransplantation regulatory guidance.
• External: U.S. Health Resources and Services Administration, organ donation and transplant statistics.

By Riju — about