The same gene editing technology that appeared in Issue 7 - where it was used to edit a patient's own blood stem cells to treat sickle cell disease - has now been turned in a different direction. A trial published in Nature Medicine this week used CRISPR to edit donor stem cells before transplantation, not to correct a genetic disease, but to make the transplanted cells invisible to a cancer-killing drug. The drug could then be used after transplant to attack residual leukaemia without destroying the graft. The idea is precise, the logic is clean, and the trial is early - all three of those things matter in equal measure.
THE RADAR
Recent developments in the field, explained without the press release.
Stem-cell-like T cells outperform standard CAR T cells in first human trial
A trial published in Cell on April 30 tested an enriched population of stem-cell-like T cells - called T stem cell memory cells, or Tscm - in patients with blood cancers that had either relapsed after a haematopoietic stem cell transplant or failed to respond to standard therapy. Of 11 patients who received the Tscm-enriched product, 5 achieved complete remission and 1 achieved partial remission. Of 10 patients who had previously received standard CAR T cell therapy as a historical comparison, 1 achieved complete remission. The trial was small, and the comparison group was historical rather than randomised - that is a real limitation on what the result can establish. The signal is strong enough to watch, and this is the first human trial to test that hypothesis directly.
CRISPR is being used on both sides of the transplant equation
This issue's lead story uses CRISPR to edit donor cells before transplant. A separate March 2026 study from UCSF, published in Nature, went further in the opposite direction: researchers used a dual-particle CRISPR-Cas9 system to reprogram T cells inside the body, without removing them first. In mouse models, a single injection eliminated aggressive leukaemia, multiple myeloma, and a solid tumour within two weeks. No human trials are underway. The approach, if it translates, removes one of the central logistical burdens of CAR T therapy: manufacturing cells outside the body. That is the step that makes current CAR T treatments expensive, slow, and inaccessible to most patients outside specialist centres.
CRISPR, a donor transplant, and the AML relapse problem
Acute myeloid leukaemia is one of the most treatment-resistant blood cancers. For patients with high-risk AML, a haematopoietic stem cell transplant from a matched donor is often the only potentially curative option. The donor's stem cells replace the patient's diseased blood-forming system entirely. When it works, it can produce long-term remission.
The problem is relapse. Even after a successful transplant, residual leukaemia cells survive in a significant proportion of patients and the disease returns. The most effective drugs for attacking those residual cells - including gemtuzumab ozogamicin, an antibody-drug conjugate that carries a cytotoxic payload directly to leukaemia cells - work by targeting a protein called CD33. CD33 is expressed on most AML cells. It is also expressed on the healthy donor myeloid cells that the transplant just established. A drug that targets CD33 cannot easily distinguish between them.
The trial published in Nature Medicine on May 12 began with a specific question: what if the donor cells did not have CD33 at all?
Researchers at Washington University School of Medicine in St Louis, working across 15 centres in the United States and Canada, used CRISPR-Cas9 gene editing to remove CD33 from donor haematopoietic stem cells before transplantation. The edited product is called trem-cel, manufactured by Vor Biopharma, which funded the trial. Several co-authors were Vor Biopharma employees at the time the work was conducted. That funding relationship is stated here because it is relevant to how the results should be read.
Thirty adult patients with AML or myelodysplastic syndrome at high risk of relapse received the CD33-deleted transplant, followed by cycles of gemtuzumab ozogamicin as post-transplant maintenance. The primary question the trial was designed to answer was safety: would CRISPR-edited donor stem cells engraft normally and produce healthy blood cells, or would the editing impair their function?
All 30 patients achieved neutrophil engraftment by day 28. Rates of graft-versus-host disease, transplant-related mortality, and time to engraftment were comparable to historical data from standard unmodified transplants, and the editing did not appear to impair the graft's ability to establish, produce blood, or tolerate the subsequent CD33-targeted drug.
What the trial did not establish, because it was not designed or powered to do so, is whether this approach reduces relapse rates compared to standard transplantation. That question requires a larger, randomised controlled trial. The principal investigator described the results as encouraging and consistent with the concept. That is the appropriate framing for a Phase 1/2 safety study.
The concept is worth holding separately from the trial results. CRISPR was applied here not to correct a disease but to create a therapeutic window: a gap between what the drug attacks and what it spares. Whether that window translates to longer survival, and whether the approach scales beyond specialist centres, are the questions the next phase must answer.
STUDY OF THE WEEK
Editing the graft to enable the drug
Phase 1/2 trial - trem-cel (tremtelectogene empogeditemcel), published in Nature Medicine, May 12, 2026
Researchers removed CD33 from donor haematopoietic stem cells using CRISPR-Cas9 before transplantation, then administered gemtuzumab ozogamicin as post-transplant maintenance therapy. The trial enrolled patients with AML or MDS at high risk of relapse across 15 US and Canadian centres.
Study type: Phase 1/2, open-label, multicenter; human patients, no randomised control group
Sample size: 30 patients
What they found: All 30 patients achieved neutrophil engraftment by day 28. Graft-versus-host disease incidence, transplant-related mortality, and engraftment timing were comparable to historical outcomes from standard unmodified transplants. CD33 deletion did not appear to impair donor stem cell function.
Most important caveat: This was a safety trial, not an efficacy trial. Whether CD33-deleted transplants reduce AML relapse rates compared to standard transplantation is unanswered. The trial was funded by Vor Biopharma, which manufactures the product. Historical comparisons are not a substitute for a randomised control group.
Why it matters anyway: This is the first human demonstration that CRISPR-edited donor stem cells can engraft safely and function normally while enabling a post-transplant drug that would otherwise destroy the graft. The concept is proven in humans. The clinical benefit question is next.
WHAT'S REAL / WHAT'S NOISE / WHAT TO WATCH
REAL
Haematopoietic stem cell transplantation for blood cancers — the established benchmark since Issue 3, now serving as the platform onto which CRISPR editing is being layered. The transplant itself is proven. The edited version is Phase 1/2, safety only. Those are two separate facts and both are real.
NOISE
Coverage describing the trem-cel trial as a proven treatment for AML. It is a safety trial with 30 patients and no randomised comparator. The primary finding is that gene-edited donor cells engraft. Whether they improve outcomes for AML patients is a question a Phase 3 trial must answer. The funding relationship between Vor Biopharma and the research team is a further reason to hold the result at arm's length until independent replication.
WATCH
The Tscm CAR T cell field following the Cell publication. If enriched stem-cell-like T cell products consistently outperform standard CAR T cells in larger trials, it changes the manufacturing and clinical logic of adoptive cell therapy across blood cancers. The comparison in the April trial was historical rather than randomised. The signal is early and the hypothesis is now in human testing, which is where Watch becomes something to track with closer attention.
THE RED FLAG REPORT
Industry-funded trials: what to ask, every time
The trem-cel trial was designed, manufactured, and funded by Vor Biopharma. Several co-authors were company employees when the work was conducted. This is stated in the lead story above and it is stated again here, because the pattern is worth naming directly.
Industry-funded trials are common in cell therapy, and many have produced reliable data - but funding by a manufacturer changes the questions a reader should ask: who chose the endpoints, was the trial designed to show safety or efficacy, and has the result been independently replicated?
A safety trial funded by the manufacturer of the product, that shows the product is safe, is not a finding of clinical benefit. That distinction holds even when the science is genuinely interesting.
READER LENS
What engraftment means - and why it was the right primary endpoint
Engraftment is the process by which transplanted donor stem cells establish themselves in the recipient's bone marrow and begin producing functional blood cells. Without it, the transplant has failed entirely: the patient has no functioning blood-forming system.
Neutrophil engraftment specifically, meaning adequate neutrophil counts appearing in the blood, is the signal that the graft has taken and the donor stem cells are producing functional immune cells.
Measuring neutrophil engraftment by day 28 was the appropriate primary endpoint for a trial whose central question was whether CRISPR editing impaired donor stem cell function. If the edited cells failed to engraft, nothing else about the trial would matter. They did not fail. That is what the primary endpoint established, and it is a meaningful result on its own terms - even before the efficacy question is asked.
AML kills most of the people it reaches, often within months of diagnosis. The drug that might prevent relapse destroys the transplant that makes survival possible. Someone looked at that problem and asked whether the transplant could be made to not care about the drug. Thirty patients later, the answer appears to be yes - and the next question is whether that matters for how long they live.
