Organs on the Hoof: Xenotransplantation Trades Its Miracle Era for a Manufacturing Line

2026-08-19 · Fortnightly · Issue № 4 · Scientific Frontier Collider

Lead selection note: this edition's other finalist was the AI-designed bacteriophage genome paper (Science, 6 Aug). It is the stronger "topology" story, but its public-awareness wave has already broken — NYT, ABC, BBC, IEEE Spectrum, Nature. The Collider's job is to arrive before the crowd, so the phages appear below in The Sleeper, via the rebuttal that will matter longer than the headline.

The Collision

For sixty years, xenotransplantation was medicine's longest-running punchline — Norman Shumway's quip that it "is the future, and always will be" dates to an era when surgeons were bolting chimpanzee kidneys onto dialysis patients and hoping for the best. The punchline has quietly expired. The field has stopped performing miracles and started running trials.

The convergence is a genuinely multi-field pileup: CRISPR multiplex genome editing (which made a 10-edit pig a routine breeding objective rather than a decade-long project), somatic-cell cloning, transplant immunology, zoonotic-disease surveillance, and — the piece nobody writes headlines about — regulatory science. In February 2025 the FDA cleared United Therapeutics' EXPAND trial (NCT06878560), the first bona fide clinical trial of a xenokidney. The first patient was transplanted at NYU Langone on 3 November 2025 with the UKidney: a pig kidney carrying six human gene knock-ins to calm the human immune system and four porcine knockouts to blunt rejection and organ overgrowth. eGenesis followed in September 2025 with FDA clearance of EGEN-2784, its own CRISPR'd pig kidney, into a Phase 1/2/3 trial. China is running a parallel program: ClonOrgan's gene-edited kidney functioned for roughly eight and a half months in a 69-year-old woman at Xi'jing Hospital before removal in November. And in May, Liao and colleagues reported in Med the first combined transplant of a pig liver and two pig kidneys into a (clinically deceased) human, with organ function sustained for nearly five days and no hyperacute rejection.

The point is not any single patient. The point is the phase transition: one-off compassionate-use heroics have become a regulated, multi-company, multi-continent clinical program with endpoints, data-safety boards, and — United Therapeutics has said so plainly — a Biologics License Application at the end of the road. An organ is starting to look less like a gift and more like a product.

The Constraint It Breaks

The constraint under attack is one of the oldest in medicine: organ supply is a fixed natural constant set by how many strangers die generously. In the United States alone, roughly 90,000 people are waiting for a kidney, 10,000 for a liver, 3,500 for a heart. Dialysis keeps kidney-failure patients alive the way a spare tire keeps a car moving — technically, miserably, and not for long. If xenografts work at scale, "donor organ" stops being a lottery won at someone else's funeral and becomes a scheduled manufacturing run. The deeper, quieter constraint it breaks is immunological: the species barrier itself, first cracked when PPL Therapeutics knocked out the alpha-gal sugar in the 1990s (killing hyperacute rejection), now being dismantled edit by edit.

The Evidence Ladder

Here is where humility earns its keep. Graded on the Collider ladder:

• Decedent demonstrations — NYU's brain-dead-body protocols; Liao et al.'s multi-organ study (~5 days, no hyperacute rejection). Solid rung 3: laboratory-grade human-biology evidence, not therapy.
• Living compassionate-use case series — a handful of patients, and the record book is honest about it. Towana Looney: 130 days off dialysis, a record at the time, then acute rejection after her immunosuppression was lowered to treat an unrelated infection; graft removed April 2025. Tim Andrews (MGH, eGenesis kidney): the current record, on the order of eight and a half to nine months, set this past October. ClonOrgan's patient in China: ~8.5 months, removed ten days short of Andrews' mark. Rung 5: early human evidence, n ≈ countable on one hand, and — the number that should temper every headline — no pig organ has yet crossed twelve months in a living person.
• Formal trials — EXPAND and EGEN-2784 are rungs being climbed in real time: protocolized, multi-site-intended, with 24-week primary endpoints (graft and patient survival, acute rejection) and expansion to ~50 patients pending safety review. No results yet. That is the story: the field is mid-climb from "brave anecdotes" to "replicated early human evidence."
• Not on the ladder at all: durability past a year, any randomized comparison against human allografts, any approved product. Years away.

The Next Gate

David Cooper — who has been transplanting pig organs into baboons since 1985, which is its own kind of credential — frames it precisely: "We are at a tipping point now. What we need now is one or two patients with a kidney transplant who do well for several months or a year — or longer." The concrete gates: (1) clean 24-week endpoint readouts from EXPAND's first cohort; (2) a graft surviving past twelve months; (3) an immunosuppression regimen that doesn't force the Looney trade-off — infection versus rejection; (4) zoonotic surveillance that scales past heroic case management (porcine endogenous retroviruses are quiet so far; "so far" is doing heavy lifting); (5) designated-pathogen-free herd scale-up — turning a Virginia farm into a supply chain; (6) eGenesis enrollment data and China's regulatory posture.

Applications & Misuse Pathways

Applications: kidneys first (largest waitlist, dialysis as fallback safety net); hearts via the Maryland/Revivicor baboon program; livers as bridge therapy; eventually multi-organ protocols and tolerance-induction strategies that could shrink immunosuppression itself. The downstream fantasy — organs on order — would restructure transplant medicine, dialysis economics, and arguably the actuarial assumptions of end-stage renal disease.

Misuse pathways: transplant tourism to jurisdictions with thinner oversight once any positive trial readout lands; premature marketing to desperate patients (the consent process for people who will die without the organ is inherently coercive-shaped); a two-tier ethics problem in which the wealthy get human organs and everyone else gets the pig; industrial animal-welfare questions that scale with the herds; and the slow-burn biosafety risk of a zoonotic event in an era of porous global surveillance. None of these are reasons to stop. All of them are reasons to read press releases with the same skepticism one brings to a Voronoff gonad graft.

Who to Watch

• Robert Montgomery (NYU Langone Transplant Institute) — EXPAND principal investigator; the field's public face.
• Tatsuo Kawai & David Cooper (Massachusetts General Hospital) — the eGenesis surgical axis and its longest institutional memory.
• David Sachs (MGH/Columbia) — the house skeptic-immunologist; his "a lot of open questions" is the correct prior.
• David Ayares (Revivicor / United Therapeutics) — keeper of the Blacksburg herd and the 10-edit architecture; Martine Rothblatt's 2011 rescue of Revivicor is the reason this field exists at industrial scale.
• eGenesis (Cambridge) — EGEN-2784 enrollment pace is the competitive tell.
• Dengke Pan (ClonOrgan) and Kefeng Dou (Xi'jing Hospital) — China's program is months behind on edits but not on ambition.
• Also watch: FDA's evolving xeno guidance and EXPAND's data-safety board, which holds the ~50-patient expansion keys.

Platform Tech Watch

The enabling layer of this entire issue is multiplex genome editing as infrastructure: the same CRISPR stack that makes a 10-edit pig routine is the stack behind genome-language-model phage design and the $27.7M ARPA-H award (July) to an IGI-led consortium for pediatric in vivo editing. Editing has stopped being the result and become the substrate. Track the substrate, and the applications announce themselves.

The Sleeper

The apparently minor result that could become foundational: Oliver Crook's independent reanalysis of the AI-designed phage genomes (Oxford; bioRxiv, 10.64898/2026.06.12.731871), surfaced by Elie Dolgin's IEEE Spectrum piece on 12 August. Crook found the 16 viable AI phages average ~97% sequence identity to their ΦX174 template and sit inside natural phage diversity — "brothers and sisters of the original virus," not a new branch on the tree. It reads like a pedantic footnote to August's splashiest paper. It is actually the prototype of a genre the field desperately needs: independent novelty accounting for generative biology. Every "AI invented X" headline from here forward needs a Crook check — ANI distributions, phylogenetic placement, functional counterfactuals. And the genre already has its calibration case: King's rejoinder (the truncated DNA-packaging protein borrowed from a distant phage, viable where conventional engineering had failed) is exactly the kind of functional-novelty evidence an audit should demand. If a standard audit practice forms, it becomes the load-bearing wall between genuine capability expansion and expensive training-set parroting. Watch this genre, not just this paper.

Sources

• United Therapeutics press release, first EXPAND transplant (2025-11-03): ir.unither.com
• NYU Langone Health: "First Gene-Edited Pig Kidney Transplant Clinical Trial Begins" (EXPAND / NCT06878560)
• PMC: "The First Clinical Renal Xenotransplantation Study (EXPAND): Evaluating Safety, Function, and Zoonotic Risks"
• Cohen, J., Science news feature: "Can gene-edited pigs solve the organ transplant shortage?"
• CRISPR Medicine News (2025-09-10): eGenesis FDA clearance, EGEN-2784
• Liao, J. et al., Med 7, 101148 (2026); Nature news (2026-05-29): "First pig liver and kidneys transplanted into a person"
• Science news update (2025-12-10): ClonOrgan kidney removed after ~8.5 months
• Dolgin, E., IEEE Spectrum (2026-08-12): "AI Can Now Design Functional Viruses. Should We Worry?"
• Crook, O. et al., bioRxiv 10.64898/2026.06.12.731871
• King, S.H., Hie, B.L. et al., Science (2026-08-06), DOI 10.1126/science.aec2657
• Science Media Centre expert reaction roundup (2026-08-06); Palmer et al., bioRxiv (2026-08-04), input screening for protein design tools